import pandas as pd
import torch
import numpy as np
import random
import time, timeit

# @title
"""network2.py
~~~~~~~~~~~~~~

An improved version of network.py, implementing the stochastic
gradient descent learning algorithm for a feedforward neural network.
Improvements include the addition of the cross-entropy cost function,
regularization, and better initialization of network weights.  Note
that I have focused on making the code simple, easily readable, and
easily modifiable.  It is not optimized, and omits many desirable
features.

"""

#### Libraries
# Standard library
import json
import random
import sys

# Third-party libraries
import numpy as np


#### Define the quadratic and cross-entropy cost functions

class QuadraticCost(object):

    @staticmethod
    def fn(a, y):
        """Return the cost associated with an output ``a`` and desired output
        ``y``.

        """
        return 0.5*np.linalg.norm(a-y)**2

    @staticmethod
    def delta(z, a, y):
        """Return the error delta from the output layer."""
        return (a-y) * sigmoid_prime(z)


class CrossEntropyCost(object):

    @staticmethod
    def fn(a, y):
        """Return the cost associated with an output ``a`` and desired output
        ``y``.  Note that np.nan_to_num is used to ensure numerical
        stability.  In particular, if both ``a`` and ``y`` have a 1.0
        in the same slot, then the expression (1-y)*np.log(1-a)
        returns nan.  The np.nan_to_num ensures that that is converted
        to the correct value (0.0).

        """
        return np.sum(np.nan_to_num(-y*np.log(a)-(1-y)*np.log(1-a)))

    @staticmethod
    def delta(z, a, y):
        """Return the error delta from the output layer.  Note that the
        parameter ``z`` is not used by the method.  It is included in
        the method's parameters in order to make the interface
        consistent with the delta method for other cost classes.

        """
        return (a-y)


#### Main Network class
class Network(object):

    def __init__(self, sizes, cost=CrossEntropyCost):
        """The list ``sizes`` contains the number of neurons in the respective
        layers of the network.  For example, if the list was [2, 3, 1]
        then it would be a three-layer network, with the first layer
        containing 2 neurons, the second layer 3 neurons, and the
        third layer 1 neuron.  The biases and weights for the network
        are initialized randomly, using
        ``self.default_weight_initializer`` (see docstring for that
        method).

        """
        self.num_layers = len(sizes)
        self.sizes = sizes
        self.default_weight_initializer()
        self.cost=cost

    def default_weight_initializer(self):
        """Initialize each weight using a Gaussian distribution with mean 0
        and standard deviation 1 over the square root of the number of
        weights connecting to the same neuron.  Initialize the biases
        using a Gaussian distribution with mean 0 and standard
        deviation 1.

        Note that the first layer is assumed to be an input layer, and
        by convention we won't set any biases for those neurons, since
        biases are only ever used in computing the outputs from later
        layers.

        """
        self.biases = [np.random.randn(y, 1) for y in self.sizes[1:]]
        self.weights = [np.random.randn(y, x)/np.sqrt(x)
                        for x, y in zip(self.sizes[:-1], self.sizes[1:])]

    def large_weight_initializer(self):
        """Initialize the weights using a Gaussian distribution with mean 0
        and standard deviation 1.  Initialize the biases using a
        Gaussian distribution with mean 0 and standard deviation 1.

        Note that the first layer is assumed to be an input layer, and
        by convention we won't set any biases for those neurons, since
        biases are only ever used in computing the outputs from later
        layers.

        This weight and bias initializer uses the same approach as in
        Chapter 1, and is included for purposes of comparison.  It
        will usually be better to use the default weight initializer
        instead.

        """
        self.biases = [np.random.randn(y, 1) for y in self.sizes[1:]]
        self.weights = [np.random.randn(y, x)
                        for x, y in zip(self.sizes[:-1], self.sizes[1:])]

    def feedforward(self, a):
        """Return the output of the network if ``a`` is input."""
        for b, w in zip(self.biases, self.weights):
            a = sigmoid(np.dot(w, a)+b)
        return a

    def SGD(self, training_data, epochs, mini_batch_size, eta,
            lmbda = 0.0,
            evaluation_data=None,
            monitor_evaluation_cost=False,
            monitor_evaluation_accuracy=False,
            monitor_training_cost=False,
            monitor_training_accuracy=False):
        """Train the neural network using mini-batch stochastic gradient
        descent.  The ``training_data`` is a list of tuples ``(x, y)``
        representing the training inputs and the desired outputs.  The
        other non-optional parameters are self-explanatory, as is the
        regularization parameter ``lmbda``.  The method also accepts
        ``evaluation_data``, usually either the validation or test
        data.  We can monitor the cost and accuracy on either the
        evaluation data or the training data, by setting the
        appropriate flags.  The method returns a tuple containing four
        lists: the (per-epoch) costs on the evaluation data, the
        accuracies on the evaluation data, the costs on the training
        data, and the accuracies on the training data.  All values are
        evaluated at the end of each training epoch.  So, for example,
        if we train for 30 epochs, then the first element of the tuple
        will be a 30-element list containing the cost on the
        evaluation data at the end of each epoch. Note that the lists
        are empty if the corresponding flag is not set.

        """
        if evaluation_data: n_data = len(list(evaluation_data))
        n = len(list(training_data))
        evaluation_cost, evaluation_accuracy = [], []
        training_cost, training_accuracy = [], []
        for j in range(epochs):
            random.shuffle(training_data)
            mini_batches = [
                training_data[k:k+mini_batch_size]
                for k in range(0, n, mini_batch_size)]
            for mini_batch in mini_batches:
                self.update_mini_batch(
                    mini_batch, eta, lmbda, len(training_data))
            print("Epoch %s training complete" % j)
            if monitor_training_cost:
                cost = self.total_cost(training_data, lmbda)
                training_cost.append(cost)
                print("Cost on training data: {}".format(cost))
            if monitor_training_accuracy:
                accuracy = self.accuracy(training_data, convert=True)
                training_accuracy.append(accuracy)
                print("Accuracy on training data: {} / {}".format(
                    accuracy, n))
            if monitor_evaluation_cost:
                cost = self.total_cost(evaluation_data, lmbda, convert=True)
                evaluation_cost.append(cost)
                print("Cost on evaluation data: {}".format(cost))
            if monitor_evaluation_accuracy:
                accuracy = self.accuracy(evaluation_data)
                evaluation_accuracy.append(accuracy)
                print("Accuracy on evaluation data: {} / {}".format(
                    self.accuracy(evaluation_data), n_data))
        return evaluation_cost, evaluation_accuracy, \
            training_cost, training_accuracy

    def update_mini_batch(self, mini_batch, eta, lmbda, n):
        """Update the network's weights and biases by applying gradient
        descent using backpropagation to a single mini batch.  The
        ``mini_batch`` is a list of tuples ``(x, y)``, ``eta`` is the
        learning rate, ``lmbda`` is the regularization parameter, and
        ``n`` is the total size of the training data set.

        """
        nabla_b = [np.zeros(b.shape) for b in self.biases]
        nabla_w = [np.zeros(w.shape) for w in self.weights]
        for x, y in mini_batch:
            delta_nabla_b, delta_nabla_w = self.backprop(x, y)
            nabla_b = [nb+dnb for nb, dnb in zip(nabla_b, delta_nabla_b)]
            nabla_w = [nw+dnw for nw, dnw in zip(nabla_w, delta_nabla_w)]
        self.weights = [(1-eta*(lmbda/n))*w-(eta/len(mini_batch))*nw
                        for w, nw in zip(self.weights, nabla_w)]
        self.biases = [b-(eta/len(mini_batch))*nb
                       for b, nb in zip(self.biases, nabla_b)]

    def backprop(self, x, y):
        """Return a tuple ``(nabla_b, nabla_w)`` representing the
        gradient for the cost function C_x.  ``nabla_b`` and
        ``nabla_w`` are layer-by-layer lists of numpy arrays, similar
        to ``self.biases`` and ``self.weights``."""
        nabla_b = [np.zeros(b.shape) for b in self.biases]
        nabla_w = [np.zeros(w.shape) for w in self.weights]
        # feedforward
        activation = x
        activations = [x] # list to store all the activations, layer by layer
        zs = [] # list to store all the z vectors, layer by layer
        for b, w in zip(self.biases, self.weights):
            z = np.dot(w, activation)+b
            zs.append(z)
            activation = sigmoid(z)
            activations.append(activation)
        # backward pass
        delta = (self.cost).delta(zs[-1], activations[-1], y)
        nabla_b[-1] = delta
        nabla_w[-1] = np.dot(delta, activations[-2].transpose())
        # Note that the variable l in the loop below is used a little
        # differently to the notation in Chapter 2 of the book.  Here,
        # l = 1 means the last layer of neurons, l = 2 is the
        # second-last layer, and so on.  It's a renumbering of the
        # scheme in the book, used here to take advantage of the fact
        # that Python can use negative indices in lists.
        for l in range(2, self.num_layers):
            z = zs[-l]
            sp = sigmoid_prime(z)
            delta = np.dot(self.weights[-l+1].transpose(), delta) * sp
            nabla_b[-l] = delta
            nabla_w[-l] = np.dot(delta, activations[-l-1].transpose())
        return (nabla_b, nabla_w)

    def accuracy(self, data, convert=False):
        """Return the number of inputs in ``data`` for which the neural
        network outputs the correct result. The neural network's
        output is assumed to be the index of whichever neuron in the
        final layer has the highest activation.

        The flag ``convert`` should be set to False if the data set is
        validation or test data (the usual case), and to True if the
        data set is the training data. The need for this flag arises
        due to differences in the way the results ``y`` are
        represented in the different data sets.  In particular, it
        flags whether we need to convert between the different
        representations.  It may seem strange to use different
        representations for the different data sets.  Why not use the
        same representation for all three data sets?  It's done for
        efficiency reasons -- the program usually evaluates the cost
        on the training data and the accuracy on other data sets.
        These are different types of computations, and using different
        representations speeds things up.  More details on the
        representations can be found in
        mnist_loader.load_data_wrapper.

        """
        if convert:
            results = [(np.argmax(self.feedforward(x)), np.argmax(y))
                       for (x, y) in data]
        else:
            results = [(np.argmax(self.feedforward(x)), y)
                        for (x, y) in data]
        return sum(int(x == y) for (x, y) in results)

    def total_cost(self, data, lmbda, convert=False):
        """Return the total cost for the data set ``data``.  The flag
        ``convert`` should be set to False if the data set is the
        training data (the usual case), and to True if the data set is
        the validation or test data.  See comments on the similar (but
        reversed) convention for the ``accuracy`` method, above.
        """
        cost = 0.0
        for x, y in data:
            a = self.feedforward(x)
            if convert: y = vectorized_result(y)
            cost += self.cost.fn(a, y)/len(data)
        cost += 0.5*(lmbda/len(data))*sum(
            np.linalg.norm(w)**2 for w in self.weights)
        return cost

    def save(self, filename):
        """Save the neural network to the file ``filename``."""
        data = {"sizes": self.sizes,
                "weights": [w.tolist() for w in self.weights],
                "biases": [b.tolist() for b in self.biases],
                "cost": str(self.cost.__name__)}
        f = open(filename, "w")
        json.dump(data, f)
        f.close()

# @title
#### Loading a Network
def load(filename):
    """Load a neural network from the file ``filename``.  Returns an
    instance of Network.

    """
    f = open(filename, "r")
    data = json.load(f)
    f.close()
    cost = getattr(sys.modules[__name__], data["cost"])
    net = Network(data["sizes"], cost=cost)
    net.weights = [np.array(w) for w in data["weights"]]
    net.biases = [np.array(b) for b in data["biases"]]
    return net

#### Miscellaneous functions
def vectorized_result(j):
    """Return a 10-dimensional unit vector with a 1.0 in the j'th position
    and zeroes elsewhere.  This is used to convert a digit (0...9)
    into a corresponding desired output from the neural network.

    """
    e = np.zeros((10, 1))
    e[j] = 1.0
    return e

def sigmoid(z):
    """The sigmoid function."""
    return 1.0/(1.0+np.exp(-z))

def sigmoid_prime(z):
    """Derivative of the sigmoid function."""
    return sigmoid(z)*(1-sigmoid(z))

# @title
"""network2mat.py
~~~~~~~~~~~~~~

An improved version of network2.py, to perform fully matrix-based
backpropagation. This version avoids the loop over each training
example by using matrix operations on the entire mini-batch.

"""

#### Main Network class
class Network2mat(object):

    def __init__(self, sizes, cost=CrossEntropyCost, reg='l2'):
        self.num_layers = len(sizes)
        self.sizes = sizes
        self.default_weight_initializer()
        self.cost = cost
        self.reg = reg

    def default_weight_initializer(self):
        self.biases = [np.random.randn(y, 1) for y in self.sizes[1:]]
        self.weights = [np.random.randn(y, x)/np.sqrt(x)
                        for x, y in zip(self.sizes[:-1], self.sizes[1:])]

    def large_weight_initializer(self):
        self.biases = [np.random.randn(y, 1) for y in self.sizes[1:]]
        self.weights = [np.random.randn(y, x)
                        for x, y in zip(self.sizes[:-1], self.sizes[1:])]

    def feedforward(self, a):
        """Return the output of the network if ``a`` is input."""
        for b, w in zip(self.biases, self.weights):
            a = sigmoid(np.dot(w, a)+b)
        return a

    def SGD(self, training_data, epochs, mini_batch_size, eta,
            lmbda = 0.0,
            evaluation_data=None,
            monitor_evaluation_cost=False,
            monitor_evaluation_accuracy=False,
            monitor_training_cost=False,
            monitor_training_accuracy=False):

        if evaluation_data: n_data = len(list(evaluation_data))
        n = len(list(training_data))
        evaluation_cost, evaluation_accuracy = [], []
        training_cost, training_accuracy = [], []
        for j in range(epochs):
            random.shuffle(training_data)
            mini_batches = [
                training_data[k:k+mini_batch_size]
                for k in range(0, n, mini_batch_size)]
            for mini_batch in mini_batches:
                self.update_mini_batch(
                    mini_batch, eta, lmbda, len(training_data))
            print("Epoch %s training complete" % j)
            if monitor_training_cost:
                cost = self.total_cost(training_data, lmbda)
                training_cost.append(cost)
                print("Cost on training data: {}".format(cost))
            if monitor_training_accuracy:
                accuracy = self.accuracy(training_data, convert=True)
                training_accuracy.append(accuracy)
                print("Accuracy on training data: {} / {}".format(
                    accuracy, n))
            if monitor_evaluation_cost:
                cost = self.total_cost(evaluation_data, lmbda, convert=True)
                evaluation_cost.append(cost)
                print("Cost on evaluation data: {}".format(cost))
            if monitor_evaluation_accuracy:
                accuracy = self.accuracy(evaluation_data)
                evaluation_accuracy.append(accuracy)
                print("Accuracy on evaluation data: {} / {}".format(
                    self.accuracy(evaluation_data), n_data))
        return evaluation_cost, evaluation_accuracy, \
            training_cost, training_accuracy

    def update_mini_batch(self, mini_batch, eta, lmbda, n):
        """Update the network's weights and biases by applying gradient descent
        using backpropagation to a single mini-batch."""

        # Convert the mini-batch into matrix form
        X = np.column_stack([x for x, y in mini_batch])
        Y = np.column_stack([y for x, y in mini_batch])

        # Perform forward and backward passes
        nabla_b, nabla_w = self.backprop_matrix(X, Y)

        # Update weights and biases using gradient descent and L1 or L2 regularization
        if self.reg == 'l2':
          self.weights = [(1 - eta * (lmbda/n)) * w - (eta/len(mini_batch)) * nw
                          for w, nw in zip(self.weights, nabla_w)]
        elif self.reg == 'l1':
          self.weights = [w - (eta * (lmbda/n)) * np.sign(w) - (eta / len(mini_batch)) * nw
                          for w, nw in zip(self.weights, nabla_w)]
        self.biases = [b - (eta / len(mini_batch)) * nb
                      for b, nb in zip(self.biases, nabla_b)]


    def backprop_matrix(self, X, Y):
        """Matrix-based backpropagation for a mini-batch."""

        # Initialize gradients
        nabla_b = [np.zeros(b.shape) for b in self.biases]
        nabla_w = [np.zeros(w.shape) for w in self.weights]

        # Step 1: Forward pass
        activation = X
        activations = [X]  # Store all activations layer by layer
        zs = []  # Store all z vectors layer by layer

        # Feedforward step through all layers
        for b, w in zip(self.biases, self.weights):
            z = np.dot(w, activation) + b
            zs.append(z)
            activation = sigmoid(z)
            activations.append(activation)

        # Step 2: Backward pass

        # Compute delta at the output layer
        delta = self.cost.delta(zs[-1], activations[-1], Y)

        # Gradients for the output layer
        nabla_b[-1] = np.sum(delta, axis=1, keepdims=True)
        nabla_w[-1] = np.dot(delta, activations[-2].T)

        # Backpropagate through previous layers
        for l in range(2, self.num_layers):
            z = zs[-l]
            sp = sigmoid_prime(z)
            delta = np.dot(self.weights[-l+1].T, delta) * sp
            nabla_b[-l] = np.sum(delta, axis=1, keepdims=True)
            nabla_w[-l] = np.dot(delta, activations[-l-1].T)

        return (nabla_b, nabla_w)

    def accuracy(self, data, convert=False):
        if convert:
            results = [(np.argmax(self.feedforward(x)), np.argmax(y))
                       for (x, y) in data]
        else:
            results = [(np.argmax(self.feedforward(x)), y)
                        for (x, y) in data]
        return sum(int(x == y) for (x, y) in results)

    def total_cost(self, data, lmbda, convert=False):
        cost = 0.0
        for x, y in data:
            a = self.feedforward(x)
            if convert: y = vectorized_result(y)
            cost += self.cost.fn(a, y)/len(data)
        if self.reg == 'l2':
          cost += 0.5*(lmbda/len(data))*sum(np.linalg.norm(w)**2 for w in self.weights)
        elif self.reg == 'l1':
          cost += (lmbda/len(data))*sum(np.linalg.norm(w, ord=1) for w in self.weights)
        return cost

    def save(self, filename):
        """Save the neural network to the file ``filename``."""
        data = {"sizes": self.sizes,
                "weights": [w.tolist() for w in self.weights],
                "biases": [b.tolist() for b in self.biases],
                "cost": str(self.cost.__name__)}
        f = open(filename, "w")
        json.dump(data, f)
        f.close()

# !pip install mnist

import mnist_loader

# Load data
training_data, validation_data, test_data = mnist_loader.load_data_wrapper()

# Convert data to lists
training_data = list(training_data)
test_data = list(test_data)

# Determine device
device = 'cuda' if torch.cuda.is_available() else 'cpu'
print(f"Using {device} device")

Using cpu device

%%time
# Initialize and train network
net = Network2mat([784, 30, 10], cost=CrossEntropyCost)
net.large_weight_initializer()
net.SGD(training_data, 30, 10, 0.5, evaluation_data=test_data, monitor_evaluation_accuracy=True)

Epoch 0 training complete
Accuracy on evaluation data: 9183 / 10000
Epoch 1 training complete
Accuracy on evaluation data: 9265 / 10000
Epoch 2 training complete
Accuracy on evaluation data: 9364 / 10000
Epoch 3 training complete
Accuracy on evaluation data: 9413 / 10000
Epoch 4 training complete
Accuracy on evaluation data: 9470 / 10000
Epoch 5 training complete
Accuracy on evaluation data: 9458 / 10000
Epoch 6 training complete
Accuracy on evaluation data: 9498 / 10000
Epoch 7 training complete
Accuracy on evaluation data: 9476 / 10000
Epoch 8 training complete
Accuracy on evaluation data: 9513 / 10000
Epoch 9 training complete
Accuracy on evaluation data: 9496 / 10000
Epoch 10 training complete
Accuracy on evaluation data: 9500 / 10000
Epoch 11 training complete
Accuracy on evaluation data: 9517 / 10000
Epoch 12 training complete
Accuracy on evaluation data: 9482 / 10000
Epoch 13 training complete
Accuracy on evaluation data: 9503 / 10000
Epoch 14 training complete
Accuracy on evaluation data: 9532 / 10000
Epoch 15 training complete
Accuracy on evaluation data: 9525 / 10000
Epoch 16 training complete
Accuracy on evaluation data: 9560 / 10000
Epoch 17 training complete
Accuracy on evaluation data: 9541 / 10000
Epoch 18 training complete
Accuracy on evaluation data: 9515 / 10000
Epoch 19 training complete
Accuracy on evaluation data: 9540 / 10000
Epoch 20 training complete
Accuracy on evaluation data: 9536 / 10000
Epoch 21 training complete
Accuracy on evaluation data: 9547 / 10000
Epoch 22 training complete
Accuracy on evaluation data: 9507 / 10000
Epoch 23 training complete
Accuracy on evaluation data: 9530 / 10000
Epoch 24 training complete
Accuracy on evaluation data: 9546 / 10000
Epoch 25 training complete
Accuracy on evaluation data: 9538 / 10000
Epoch 26 training complete
Accuracy on evaluation data: 9551 / 10000
Epoch 27 training complete
Accuracy on evaluation data: 9528 / 10000
Epoch 28 training complete
Accuracy on evaluation data: 9544 / 10000
Epoch 29 training complete
Accuracy on evaluation data: 9550 / 10000
CPU times: user 1min 23s, sys: 23.9 s, total: 1min 46s
Wall time: 1min 30s

([],
 [9183,
  9265,
  9364,
  9413,
  9470,
  9458,
  9498,
  9476,
  9513,
  9496,
  9500,
  9517,
  9482,
  9503,
  9532,
  9525,
  9560,
  9541,
  9515,
  9540,
  9536,
  9547,
  9507,
  9530,
  9546,
  9538,
  9551,
  9528,
  9544,
  9550],
 [],
 [])

# %%time
# net = Network([784, 100, 10], cost=CrossEntropyCost)
# net.large_weight_initializer()
# net.SGD(training_data, 30, 10, 0.5, evaluation_data=test_data, monitor_evaluation_accuracy=True)

%%time
net = Network2mat([784, 100, 10], cost=CrossEntropyCost)
net.large_weight_initializer()
net.SGD(training_data, 30, 10, 0.5, evaluation_data=test_data, monitor_evaluation_accuracy=True)

Epoch 0 training complete
Accuracy on evaluation data: 9179 / 10000
Epoch 1 training complete
Accuracy on evaluation data: 9435 / 10000
Epoch 2 training complete
Accuracy on evaluation data: 9474 / 10000
Epoch 3 training complete
Accuracy on evaluation data: 9555 / 10000
Epoch 4 training complete
Accuracy on evaluation data: 9543 / 10000
Epoch 5 training complete
Accuracy on evaluation data: 9589 / 10000
Epoch 6 training complete
Accuracy on evaluation data: 9588 / 10000
Epoch 7 training complete
Accuracy on evaluation data: 9614 / 10000
Epoch 8 training complete
Accuracy on evaluation data: 9627 / 10000
Epoch 9 training complete
Accuracy on evaluation data: 9642 / 10000
Epoch 10 training complete
Accuracy on evaluation data: 9633 / 10000
Epoch 11 training complete
Accuracy on evaluation data: 9655 / 10000
Epoch 12 training complete
Accuracy on evaluation data: 9654 / 10000
Epoch 13 training complete
Accuracy on evaluation data: 9652 / 10000
Epoch 14 training complete
Accuracy on evaluation data: 9626 / 10000
Epoch 15 training complete
Accuracy on evaluation data: 9650 / 10000
Epoch 16 training complete
Accuracy on evaluation data: 9655 / 10000
Epoch 17 training complete
Accuracy on evaluation data: 9654 / 10000
Epoch 18 training complete
Accuracy on evaluation data: 9681 / 10000
Epoch 19 training complete
Accuracy on evaluation data: 9666 / 10000
Epoch 20 training complete
Accuracy on evaluation data: 9645 / 10000
Epoch 21 training complete
Accuracy on evaluation data: 9661 / 10000
Epoch 22 training complete
Accuracy on evaluation data: 9665 / 10000
Epoch 23 training complete
Accuracy on evaluation data: 9678 / 10000
Epoch 24 training complete
Accuracy on evaluation data: 9667 / 10000
Epoch 25 training complete
Accuracy on evaluation data: 9675 / 10000
Epoch 26 training complete
Accuracy on evaluation data: 9669 / 10000
Epoch 27 training complete
Accuracy on evaluation data: 9674 / 10000
Epoch 28 training complete
Accuracy on evaluation data: 9674 / 10000
Epoch 29 training complete
Accuracy on evaluation data: 9679 / 10000
CPU times: user 3min 34s, sys: 2min 4s, total: 5min 38s
Wall time: 3min 33s

([],
 [9179,
  9435,
  9474,
  9555,
  9543,
  9589,
  9588,
  9614,
  9627,
  9642,
  9633,
  9655,
  9654,
  9652,
  9626,
  9650,
  9655,
  9654,
  9681,
  9666,
  9645,
  9661,
  9665,
  9678,
  9667,
  9675,
  9669,
  9674,
  9674,
  9679],
 [],
 [])

"""
Author: Piotr A. Zolnierczuk (zolnierczukp at ornl dot gov)
URL: https://www.johndcook.com/blog/2011/06/21/how-to-fit-an-elephant/

Based on a paper by:
Drawing an elephant with four complex parameters
Jurgen Mayer, Khaled Khairy, and Jonathon Howard,
Am. J. Phys. 78, 648 (2010), DOI:10.1119/1.3254017
"""
import numpy as np
import pylab

# elephant parameters
p1, p2, p3, p4 = (50 - 30j, 18 +  8j, 12 - 10j, -14 - 60j )
p5 = 40 + 20j # eyepiece

def fourier(t, C):
    f = np.zeros(t.shape)
    A, B = C.real, C.imag
    for k in range(len(C)):
        f = f + A[k]*np.cos(k*t) + B[k]*np.sin(k*t)
    return f

def elephant(t, p1, p2, p3, p4, p5):
    npar = 6
    Cx = np.zeros((npar,), dtype='complex')
    Cy = np.zeros((npar,), dtype='complex')

    Cx[1] = p1.real*1j
    Cx[2] = p2.real*1j
    Cx[3] = p3.real
    Cx[5] = p4.real

    Cy[1] = p4.imag + p1.imag*1j
    Cy[2] = p2.imag*1j
    Cy[3] = p3.imag*1j

    x = np.append(fourier(t,Cx), [-p5.imag])
    y = np.append(fourier(t,Cy), [p5.imag])

    return x,y

x, y = elephant(np.linspace(0,2*np.pi,1000), p1, p2, p3, p4, p5)
pylab.plot(y,-x,'.')
pylab.show()

%%time
net = Network2mat([784, 30, 10], cost=CrossEntropyCost)
net.large_weight_initializer()
net.SGD(training_data[:1000], 400, 10, 0.5, evaluation_data=test_data, monitor_evaluation_accuracy=True, monitor_training_cost=True)

Epoch 0 training complete
Cost on training data: 2.1334873081011976
Accuracy on evaluation data: 5362 / 10000
Epoch 1 training complete
Cost on training data: 1.6117154896081094
Accuracy on evaluation data: 6508 / 10000
Epoch 2 training complete
Cost on training data: 1.3435858874341642
Accuracy on evaluation data: 6996 / 10000
Epoch 3 training complete
Cost on training data: 1.118955170706648
Accuracy on evaluation data: 7367 / 10000
Epoch 4 training complete
Cost on training data: 0.9901307112582952
Accuracy on evaluation data: 7592 / 10000
Epoch 5 training complete
Cost on training data: 0.8519259734137855
Accuracy on evaluation data: 7708 / 10000
Epoch 6 training complete
Cost on training data: 0.7960956201359782
Accuracy on evaluation data: 7752 / 10000
Epoch 7 training complete
Cost on training data: 0.6720702851986642
Accuracy on evaluation data: 7931 / 10000
Epoch 8 training complete
Cost on training data: 0.5998755333945173
Accuracy on evaluation data: 8037 / 10000
Epoch 9 training complete
Cost on training data: 0.5532262347451836
Accuracy on evaluation data: 8080 / 10000
Epoch 10 training complete
Cost on training data: 0.48301226958062626
Accuracy on evaluation data: 8136 / 10000
Epoch 11 training complete
Cost on training data: 0.43713906935077373
Accuracy on evaluation data: 8149 / 10000
Epoch 12 training complete
Cost on training data: 0.4054099079755518
Accuracy on evaluation data: 8160 / 10000
Epoch 13 training complete
Cost on training data: 0.37177860867561674
Accuracy on evaluation data: 8255 / 10000
Epoch 14 training complete
Cost on training data: 0.3347236482029461
Accuracy on evaluation data: 8258 / 10000
Epoch 15 training complete
Cost on training data: 0.3071926589220765
Accuracy on evaluation data: 8267 / 10000
Epoch 16 training complete
Cost on training data: 0.27895724476902123
Accuracy on evaluation data: 8277 / 10000
Epoch 17 training complete
Cost on training data: 0.2652053196232567
Accuracy on evaluation data: 8281 / 10000
Epoch 18 training complete
Cost on training data: 0.24971868688894489
Accuracy on evaluation data: 8276 / 10000
Epoch 19 training complete
Cost on training data: 0.22578091792710658
Accuracy on evaluation data: 8293 / 10000
Epoch 20 training complete
Cost on training data: 0.21674447452601472
Accuracy on evaluation data: 8319 / 10000
Epoch 21 training complete
Cost on training data: 0.19696616273484055
Accuracy on evaluation data: 8324 / 10000
Epoch 22 training complete
Cost on training data: 0.18367647407713247
Accuracy on evaluation data: 8324 / 10000
Epoch 23 training complete
Cost on training data: 0.1697081997249623
Accuracy on evaluation data: 8370 / 10000
Epoch 24 training complete
Cost on training data: 0.15694410574943818
Accuracy on evaluation data: 8351 / 10000
Epoch 25 training complete
Cost on training data: 0.14931148314777995
Accuracy on evaluation data: 8344 / 10000
Epoch 26 training complete
Cost on training data: 0.1376326060327698
Accuracy on evaluation data: 8345 / 10000
Epoch 27 training complete
Cost on training data: 0.12822834529691673
Accuracy on evaluation data: 8352 / 10000
Epoch 28 training complete
Cost on training data: 0.12334225702609936
Accuracy on evaluation data: 8348 / 10000
Epoch 29 training complete
Cost on training data: 0.11344356383141305
Accuracy on evaluation data: 8374 / 10000
Epoch 30 training complete
Cost on training data: 0.10874642376376649
Accuracy on evaluation data: 8386 / 10000
Epoch 31 training complete
Cost on training data: 0.10232741359823394
Accuracy on evaluation data: 8360 / 10000
Epoch 32 training complete
Cost on training data: 0.09800310442716333
Accuracy on evaluation data: 8361 / 10000
Epoch 33 training complete
Cost on training data: 0.09268095912167088
Accuracy on evaluation data: 8381 / 10000
Epoch 34 training complete
Cost on training data: 0.08985211317214391
Accuracy on evaluation data: 8365 / 10000
Epoch 35 training complete
Cost on training data: 0.08529139016670571
Accuracy on evaluation data: 8376 / 10000
Epoch 36 training complete
Cost on training data: 0.08354889001668293
Accuracy on evaluation data: 8390 / 10000
Epoch 37 training complete
Cost on training data: 0.08003507351151332
Accuracy on evaluation data: 8399 / 10000
Epoch 38 training complete
Cost on training data: 0.07650548159073085
Accuracy on evaluation data: 8389 / 10000
Epoch 39 training complete
Cost on training data: 0.0735963249949398
Accuracy on evaluation data: 8398 / 10000
Epoch 40 training complete
Cost on training data: 0.07054144890418132
Accuracy on evaluation data: 8382 / 10000
Epoch 41 training complete
Cost on training data: 0.06887572297685532
Accuracy on evaluation data: 8400 / 10000
Epoch 42 training complete
Cost on training data: 0.0663065350160309
Accuracy on evaluation data: 8394 / 10000
Epoch 43 training complete
Cost on training data: 0.06408076105973032
Accuracy on evaluation data: 8398 / 10000
Epoch 44 training complete
Cost on training data: 0.061858574556072424
Accuracy on evaluation data: 8388 / 10000
Epoch 45 training complete
Cost on training data: 0.06058942537298044
Accuracy on evaluation data: 8392 / 10000
Epoch 46 training complete
Cost on training data: 0.05873446242587899
Accuracy on evaluation data: 8400 / 10000
Epoch 47 training complete
Cost on training data: 0.05593898642402481
Accuracy on evaluation data: 8396 / 10000
Epoch 48 training complete
Cost on training data: 0.055205914338667846
Accuracy on evaluation data: 8410 / 10000
Epoch 49 training complete
Cost on training data: 0.05313535099205961
Accuracy on evaluation data: 8402 / 10000
Epoch 50 training complete
Cost on training data: 0.051116209408344354
Accuracy on evaluation data: 8398 / 10000
Epoch 51 training complete
Cost on training data: 0.049505308014994136
Accuracy on evaluation data: 8405 / 10000
Epoch 52 training complete
Cost on training data: 0.04786572144766827
Accuracy on evaluation data: 8403 / 10000
Epoch 53 training complete
Cost on training data: 0.04640639733502174
Accuracy on evaluation data: 8399 / 10000
Epoch 54 training complete
Cost on training data: 0.045380854077187775
Accuracy on evaluation data: 8410 / 10000
Epoch 55 training complete
Cost on training data: 0.04412889015222005
Accuracy on evaluation data: 8401 / 10000
Epoch 56 training complete
Cost on training data: 0.04290577626118409
Accuracy on evaluation data: 8403 / 10000
Epoch 57 training complete
Cost on training data: 0.04194561198588216
Accuracy on evaluation data: 8405 / 10000
Epoch 58 training complete
Cost on training data: 0.04079484197387679
Accuracy on evaluation data: 8410 / 10000
Epoch 59 training complete
Cost on training data: 0.039718523191149574
Accuracy on evaluation data: 8416 / 10000
Epoch 60 training complete
Cost on training data: 0.038891299256091943
Accuracy on evaluation data: 8408 / 10000
Epoch 61 training complete
Cost on training data: 0.03801167501946484
Accuracy on evaluation data: 8411 / 10000
Epoch 62 training complete
Cost on training data: 0.03700659233294767
Accuracy on evaluation data: 8410 / 10000
Epoch 63 training complete
Cost on training data: 0.03619008401446996
Accuracy on evaluation data: 8413 / 10000
Epoch 64 training complete
Cost on training data: 0.0354101234112211
Accuracy on evaluation data: 8404 / 10000
Epoch 65 training complete
Cost on training data: 0.03446877877563418
Accuracy on evaluation data: 8417 / 10000
Epoch 66 training complete
Cost on training data: 0.03371575942835778
Accuracy on evaluation data: 8418 / 10000
Epoch 67 training complete
Cost on training data: 0.032671918060232265
Accuracy on evaluation data: 8417 / 10000
Epoch 68 training complete
Cost on training data: 0.03191790922420547
Accuracy on evaluation data: 8420 / 10000
Epoch 69 training complete
Cost on training data: 0.031155109011162536
Accuracy on evaluation data: 8412 / 10000
Epoch 70 training complete
Cost on training data: 0.030581435138702377
Accuracy on evaluation data: 8411 / 10000
Epoch 71 training complete
Cost on training data: 0.029850450319083684
Accuracy on evaluation data: 8420 / 10000
Epoch 72 training complete
Cost on training data: 0.02927639825367843
Accuracy on evaluation data: 8410 / 10000
Epoch 73 training complete
Cost on training data: 0.028877550454782647
Accuracy on evaluation data: 8421 / 10000
Epoch 74 training complete
Cost on training data: 0.028148984621107827
Accuracy on evaluation data: 8420 / 10000
Epoch 75 training complete
Cost on training data: 0.027645446138674794
Accuracy on evaluation data: 8426 / 10000
Epoch 76 training complete
Cost on training data: 0.027157308884780768
Accuracy on evaluation data: 8410 / 10000
Epoch 77 training complete
Cost on training data: 0.02660532840891866
Accuracy on evaluation data: 8423 / 10000
Epoch 78 training complete
Cost on training data: 0.02630193671251622
Accuracy on evaluation data: 8431 / 10000
Epoch 79 training complete
Cost on training data: 0.025685865680059764
Accuracy on evaluation data: 8421 / 10000
Epoch 80 training complete
Cost on training data: 0.025230066226201827
Accuracy on evaluation data: 8424 / 10000
Epoch 81 training complete
Cost on training data: 0.024826864455365017
Accuracy on evaluation data: 8434 / 10000
Epoch 82 training complete
Cost on training data: 0.0243866839572904
Accuracy on evaluation data: 8433 / 10000
Epoch 83 training complete
Cost on training data: 0.02399530254399858
Accuracy on evaluation data: 8437 / 10000
Epoch 84 training complete
Cost on training data: 0.023730621576772862
Accuracy on evaluation data: 8419 / 10000
Epoch 85 training complete
Cost on training data: 0.02325836603295703
Accuracy on evaluation data: 8435 / 10000
Epoch 86 training complete
Cost on training data: 0.02289041133860511
Accuracy on evaluation data: 8435 / 10000
Epoch 87 training complete
Cost on training data: 0.022542631111450415
Accuracy on evaluation data: 8424 / 10000
Epoch 88 training complete
Cost on training data: 0.02219589347435611
Accuracy on evaluation data: 8435 / 10000
Epoch 89 training complete
Cost on training data: 0.021936901953618273
Accuracy on evaluation data: 8437 / 10000
Epoch 90 training complete
Cost on training data: 0.02157206047399831
Accuracy on evaluation data: 8441 / 10000
Epoch 91 training complete
Cost on training data: 0.021223795818202454
Accuracy on evaluation data: 8441 / 10000
Epoch 92 training complete
Cost on training data: 0.02096752615216147
Accuracy on evaluation data: 8430 / 10000
Epoch 93 training complete
Cost on training data: 0.02063715475299584
Accuracy on evaluation data: 8443 / 10000
Epoch 94 training complete
Cost on training data: 0.020356029193518026
Accuracy on evaluation data: 8448 / 10000
Epoch 95 training complete
Cost on training data: 0.020072665857854867
Accuracy on evaluation data: 8441 / 10000
Epoch 96 training complete
Cost on training data: 0.019834001926711618
Accuracy on evaluation data: 8441 / 10000
Epoch 97 training complete
Cost on training data: 0.01954121858171723
Accuracy on evaluation data: 8437 / 10000
Epoch 98 training complete
Cost on training data: 0.019286776431325092
Accuracy on evaluation data: 8448 / 10000
Epoch 99 training complete
Cost on training data: 0.01905240925628857
Accuracy on evaluation data: 8444 / 10000
Epoch 100 training complete
Cost on training data: 0.01880649215622303
Accuracy on evaluation data: 8447 / 10000
Epoch 101 training complete
Cost on training data: 0.01854682840467356
Accuracy on evaluation data: 8448 / 10000
Epoch 102 training complete
Cost on training data: 0.018399883172119994
Accuracy on evaluation data: 8445 / 10000
Epoch 103 training complete
Cost on training data: 0.018090713620616473
Accuracy on evaluation data: 8443 / 10000
Epoch 104 training complete
Cost on training data: 0.01789813825489991
Accuracy on evaluation data: 8454 / 10000
Epoch 105 training complete
Cost on training data: 0.01765118746282579
Accuracy on evaluation data: 8446 / 10000
Epoch 106 training complete
Cost on training data: 0.017473085175173292
Accuracy on evaluation data: 8446 / 10000
Epoch 107 training complete
Cost on training data: 0.017244494927694504
Accuracy on evaluation data: 8444 / 10000
Epoch 108 training complete
Cost on training data: 0.01703486435487612
Accuracy on evaluation data: 8441 / 10000
Epoch 109 training complete
Cost on training data: 0.016859129978839392
Accuracy on evaluation data: 8446 / 10000
Epoch 110 training complete
Cost on training data: 0.01664474198599423
Accuracy on evaluation data: 8447 / 10000
Epoch 111 training complete
Cost on training data: 0.016492894164797507
Accuracy on evaluation data: 8444 / 10000
Epoch 112 training complete
Cost on training data: 0.016277872550429357
Accuracy on evaluation data: 8449 / 10000
Epoch 113 training complete
Cost on training data: 0.01609761470488398
Accuracy on evaluation data: 8446 / 10000
Epoch 114 training complete
Cost on training data: 0.01591808120014808
Accuracy on evaluation data: 8442 / 10000
Epoch 115 training complete
Cost on training data: 0.01574090632673071
Accuracy on evaluation data: 8443 / 10000
Epoch 116 training complete
Cost on training data: 0.015569273751146593
Accuracy on evaluation data: 8446 / 10000
Epoch 117 training complete
Cost on training data: 0.015401941782147353
Accuracy on evaluation data: 8444 / 10000
Epoch 118 training complete
Cost on training data: 0.015238524446743831
Accuracy on evaluation data: 8447 / 10000
Epoch 119 training complete
Cost on training data: 0.015108287488949757
Accuracy on evaluation data: 8439 / 10000
Epoch 120 training complete
Cost on training data: 0.014915727919302634
Accuracy on evaluation data: 8439 / 10000
Epoch 121 training complete
Cost on training data: 0.014765483792022284
Accuracy on evaluation data: 8447 / 10000
Epoch 122 training complete
Cost on training data: 0.014608352572538171
Accuracy on evaluation data: 8444 / 10000
Epoch 123 training complete
Cost on training data: 0.014466773855429526
Accuracy on evaluation data: 8444 / 10000
Epoch 124 training complete
Cost on training data: 0.014318795347529108
Accuracy on evaluation data: 8443 / 10000
Epoch 125 training complete
Cost on training data: 0.014166768716930093
Accuracy on evaluation data: 8441 / 10000
Epoch 126 training complete
Cost on training data: 0.01403021765009755
Accuracy on evaluation data: 8447 / 10000
Epoch 127 training complete
Cost on training data: 0.013889891280038329
Accuracy on evaluation data: 8451 / 10000
Epoch 128 training complete
Cost on training data: 0.013756143033653877
Accuracy on evaluation data: 8444 / 10000
Epoch 129 training complete
Cost on training data: 0.013620655462519938
Accuracy on evaluation data: 8447 / 10000
Epoch 130 training complete
Cost on training data: 0.013493108647379412
Accuracy on evaluation data: 8449 / 10000
Epoch 131 training complete
Cost on training data: 0.013364012980409488
Accuracy on evaluation data: 8447 / 10000
Epoch 132 training complete
Cost on training data: 0.013238234613230631
Accuracy on evaluation data: 8448 / 10000
Epoch 133 training complete
Cost on training data: 0.013117270271950057
Accuracy on evaluation data: 8448 / 10000
Epoch 134 training complete
Cost on training data: 0.01299690578155512
Accuracy on evaluation data: 8445 / 10000
Epoch 135 training complete
Cost on training data: 0.012882657709243744
Accuracy on evaluation data: 8447 / 10000
Epoch 136 training complete
Cost on training data: 0.012758697885839247
Accuracy on evaluation data: 8447 / 10000
Epoch 137 training complete
Cost on training data: 0.012649429525397628
Accuracy on evaluation data: 8448 / 10000
Epoch 138 training complete
Cost on training data: 0.012538584374843613
Accuracy on evaluation data: 8448 / 10000
Epoch 139 training complete
Cost on training data: 0.012422775590650628
Accuracy on evaluation data: 8449 / 10000
Epoch 140 training complete
Cost on training data: 0.012319840609501332
Accuracy on evaluation data: 8448 / 10000
Epoch 141 training complete
Cost on training data: 0.012209649003281678
Accuracy on evaluation data: 8447 / 10000
Epoch 142 training complete
Cost on training data: 0.012102365427467381
Accuracy on evaluation data: 8448 / 10000
Epoch 143 training complete
Cost on training data: 0.012000446070765454
Accuracy on evaluation data: 8448 / 10000
Epoch 144 training complete
Cost on training data: 0.011904631671549513
Accuracy on evaluation data: 8452 / 10000
Epoch 145 training complete
Cost on training data: 0.011805864322997753
Accuracy on evaluation data: 8447 / 10000
Epoch 146 training complete
Cost on training data: 0.011698740239416737
Accuracy on evaluation data: 8447 / 10000
Epoch 147 training complete
Cost on training data: 0.011606371241906156
Accuracy on evaluation data: 8447 / 10000
Epoch 148 training complete
Cost on training data: 0.011514374611308427
Accuracy on evaluation data: 8445 / 10000
Epoch 149 training complete
Cost on training data: 0.011416940179436775
Accuracy on evaluation data: 8450 / 10000
Epoch 150 training complete
Cost on training data: 0.011324325638086047
Accuracy on evaluation data: 8452 / 10000
Epoch 151 training complete
Cost on training data: 0.011232413314906249
Accuracy on evaluation data: 8451 / 10000
Epoch 152 training complete
Cost on training data: 0.011143023721712417
Accuracy on evaluation data: 8444 / 10000
Epoch 153 training complete
Cost on training data: 0.011057575745872053
Accuracy on evaluation data: 8442 / 10000
Epoch 154 training complete
Cost on training data: 0.010968259739096101
Accuracy on evaluation data: 8448 / 10000
Epoch 155 training complete
Cost on training data: 0.01088128644534935
Accuracy on evaluation data: 8446 / 10000
Epoch 156 training complete
Cost on training data: 0.010799772166372307
Accuracy on evaluation data: 8446 / 10000
Epoch 157 training complete
Cost on training data: 0.010717817480123754
Accuracy on evaluation data: 8444 / 10000
Epoch 158 training complete
Cost on training data: 0.010638889630533489
Accuracy on evaluation data: 8445 / 10000
Epoch 159 training complete
Cost on training data: 0.010554400509895827
Accuracy on evaluation data: 8447 / 10000
Epoch 160 training complete
Cost on training data: 0.010475138359528497
Accuracy on evaluation data: 8448 / 10000
Epoch 161 training complete
Cost on training data: 0.010397616721467969
Accuracy on evaluation data: 8450 / 10000
Epoch 162 training complete
Cost on training data: 0.010321260115164373
Accuracy on evaluation data: 8445 / 10000
Epoch 163 training complete
Cost on training data: 0.010244649482948473
Accuracy on evaluation data: 8445 / 10000
Epoch 164 training complete
Cost on training data: 0.010170491415277361
Accuracy on evaluation data: 8443 / 10000
Epoch 165 training complete
Cost on training data: 0.01009958618454486
Accuracy on evaluation data: 8445 / 10000
Epoch 166 training complete
Cost on training data: 0.010024345542467384
Accuracy on evaluation data: 8445 / 10000
Epoch 167 training complete
Cost on training data: 0.009953348266001014
Accuracy on evaluation data: 8445 / 10000
Epoch 168 training complete
Cost on training data: 0.009883756280116205
Accuracy on evaluation data: 8448 / 10000
Epoch 169 training complete
Cost on training data: 0.009811861956010477
Accuracy on evaluation data: 8447 / 10000
Epoch 170 training complete
Cost on training data: 0.009749156561601519
Accuracy on evaluation data: 8449 / 10000
Epoch 171 training complete
Cost on training data: 0.009679778377319816
Accuracy on evaluation data: 8444 / 10000
Epoch 172 training complete
Cost on training data: 0.009609249102981717
Accuracy on evaluation data: 8443 / 10000
Epoch 173 training complete
Cost on training data: 0.009544288870302616
Accuracy on evaluation data: 8444 / 10000
Epoch 174 training complete
Cost on training data: 0.009480519623452777
Accuracy on evaluation data: 8452 / 10000
Epoch 175 training complete
Cost on training data: 0.009416154781809356
Accuracy on evaluation data: 8449 / 10000
Epoch 176 training complete
Cost on training data: 0.009354569506204469
Accuracy on evaluation data: 8446 / 10000
Epoch 177 training complete
Cost on training data: 0.00929067436973555
Accuracy on evaluation data: 8449 / 10000
Epoch 178 training complete
Cost on training data: 0.009229670289513256
Accuracy on evaluation data: 8445 / 10000
Epoch 179 training complete
Cost on training data: 0.009169356854431477
Accuracy on evaluation data: 8444 / 10000
Epoch 180 training complete
Cost on training data: 0.009111816213489487
Accuracy on evaluation data: 8442 / 10000
Epoch 181 training complete
Cost on training data: 0.0090543641894766
Accuracy on evaluation data: 8446 / 10000
Epoch 182 training complete
Cost on training data: 0.008993560686706871
Accuracy on evaluation data: 8447 / 10000
Epoch 183 training complete
Cost on training data: 0.008938581789090341
Accuracy on evaluation data: 8447 / 10000
Epoch 184 training complete
Cost on training data: 0.008880962419123395
Accuracy on evaluation data: 8446 / 10000
Epoch 185 training complete
Cost on training data: 0.008826301190224393
Accuracy on evaluation data: 8444 / 10000
Epoch 186 training complete
Cost on training data: 0.008769672548831493
Accuracy on evaluation data: 8446 / 10000
Epoch 187 training complete
Cost on training data: 0.008716483735853068
Accuracy on evaluation data: 8445 / 10000
Epoch 188 training complete
Cost on training data: 0.00866460369160469
Accuracy on evaluation data: 8449 / 10000
Epoch 189 training complete
Cost on training data: 0.00861036903221576
Accuracy on evaluation data: 8447 / 10000
Epoch 190 training complete
Cost on training data: 0.008557577345388928
Accuracy on evaluation data: 8443 / 10000
Epoch 191 training complete
Cost on training data: 0.00850933448370844
Accuracy on evaluation data: 8443 / 10000
Epoch 192 training complete
Cost on training data: 0.008456169763878356
Accuracy on evaluation data: 8442 / 10000
Epoch 193 training complete
Cost on training data: 0.008404662640716853
Accuracy on evaluation data: 8447 / 10000
Epoch 194 training complete
Cost on training data: 0.008354622029987917
Accuracy on evaluation data: 8448 / 10000
Epoch 195 training complete
Cost on training data: 0.00830697908651364
Accuracy on evaluation data: 8442 / 10000
Epoch 196 training complete
Cost on training data: 0.008257180529236265
Accuracy on evaluation data: 8448 / 10000
Epoch 197 training complete
Cost on training data: 0.008210014573194865
Accuracy on evaluation data: 8446 / 10000
Epoch 198 training complete
Cost on training data: 0.008162098131249284
Accuracy on evaluation data: 8449 / 10000
Epoch 199 training complete
Cost on training data: 0.00811539010760783
Accuracy on evaluation data: 8446 / 10000
Epoch 200 training complete
Cost on training data: 0.008072431304794821
Accuracy on evaluation data: 8444 / 10000
Epoch 201 training complete
Cost on training data: 0.00802376202280493
Accuracy on evaluation data: 8448 / 10000
Epoch 202 training complete
Cost on training data: 0.007979104863057019
Accuracy on evaluation data: 8447 / 10000
Epoch 203 training complete
Cost on training data: 0.007937625159070238
Accuracy on evaluation data: 8445 / 10000
Epoch 204 training complete
Cost on training data: 0.007890554712392644
Accuracy on evaluation data: 8444 / 10000
Epoch 205 training complete
Cost on training data: 0.00784790016459241
Accuracy on evaluation data: 8447 / 10000
Epoch 206 training complete
Cost on training data: 0.007802929222624051
Accuracy on evaluation data: 8443 / 10000
Epoch 207 training complete
Cost on training data: 0.007760549692860105
Accuracy on evaluation data: 8446 / 10000
Epoch 208 training complete
Cost on training data: 0.007718367124741382
Accuracy on evaluation data: 8447 / 10000
Epoch 209 training complete
Cost on training data: 0.007676850015661369
Accuracy on evaluation data: 8446 / 10000
Epoch 210 training complete
Cost on training data: 0.007636902480344225
Accuracy on evaluation data: 8448 / 10000
Epoch 211 training complete
Cost on training data: 0.007594654515272643
Accuracy on evaluation data: 8450 / 10000
Epoch 212 training complete
Cost on training data: 0.007557266065446237
Accuracy on evaluation data: 8446 / 10000
Epoch 213 training complete
Cost on training data: 0.007517549897936283
Accuracy on evaluation data: 8447 / 10000
Epoch 214 training complete
Cost on training data: 0.007475327712505067
Accuracy on evaluation data: 8447 / 10000
Epoch 215 training complete
Cost on training data: 0.007435933824650646
Accuracy on evaluation data: 8449 / 10000
Epoch 216 training complete
Cost on training data: 0.007397819828337148
Accuracy on evaluation data: 8448 / 10000
Epoch 217 training complete
Cost on training data: 0.007359477880672005
Accuracy on evaluation data: 8449 / 10000
Epoch 218 training complete
Cost on training data: 0.0073208837720296705
Accuracy on evaluation data: 8449 / 10000
Epoch 219 training complete
Cost on training data: 0.007283805820236689
Accuracy on evaluation data: 8449 / 10000
Epoch 220 training complete
Cost on training data: 0.007247153976131099
Accuracy on evaluation data: 8451 / 10000
Epoch 221 training complete
Cost on training data: 0.0072106084576830706
Accuracy on evaluation data: 8448 / 10000
Epoch 222 training complete
Cost on training data: 0.0071737567683653304
Accuracy on evaluation data: 8445 / 10000
Epoch 223 training complete
Cost on training data: 0.007137721603632793
Accuracy on evaluation data: 8446 / 10000
Epoch 224 training complete
Cost on training data: 0.0071018808134562916
Accuracy on evaluation data: 8446 / 10000
Epoch 225 training complete
Cost on training data: 0.007066331622788004
Accuracy on evaluation data: 8450 / 10000
Epoch 226 training complete
Cost on training data: 0.007031389918120413
Accuracy on evaluation data: 8450 / 10000
Epoch 227 training complete
Cost on training data: 0.006996545294324694
Accuracy on evaluation data: 8450 / 10000
Epoch 228 training complete
Cost on training data: 0.00696280808553848
Accuracy on evaluation data: 8451 / 10000
Epoch 229 training complete
Cost on training data: 0.0069292201735108635
Accuracy on evaluation data: 8451 / 10000
Epoch 230 training complete
Cost on training data: 0.0068955718106028935
Accuracy on evaluation data: 8446 / 10000
Epoch 231 training complete
Cost on training data: 0.006861764530310163
Accuracy on evaluation data: 8449 / 10000
Epoch 232 training complete
Cost on training data: 0.006828683516656813
Accuracy on evaluation data: 8451 / 10000
Epoch 233 training complete
Cost on training data: 0.006796285258697305
Accuracy on evaluation data: 8452 / 10000
Epoch 234 training complete
Cost on training data: 0.006764736003988712
Accuracy on evaluation data: 8453 / 10000
Epoch 235 training complete
Cost on training data: 0.006731546513205398
Accuracy on evaluation data: 8450 / 10000
Epoch 236 training complete
Cost on training data: 0.006700374780671107
Accuracy on evaluation data: 8451 / 10000
Epoch 237 training complete
Cost on training data: 0.006669692110701232
Accuracy on evaluation data: 8454 / 10000
Epoch 238 training complete
Cost on training data: 0.0066375050596078275
Accuracy on evaluation data: 8451 / 10000
Epoch 239 training complete
Cost on training data: 0.006607049930111248
Accuracy on evaluation data: 8453 / 10000
Epoch 240 training complete
Cost on training data: 0.006575578731710099
Accuracy on evaluation data: 8450 / 10000
Epoch 241 training complete
Cost on training data: 0.006545454671152349
Accuracy on evaluation data: 8451 / 10000
Epoch 242 training complete
Cost on training data: 0.0065158945218681775
Accuracy on evaluation data: 8455 / 10000
Epoch 243 training complete
Cost on training data: 0.006485383413394594
Accuracy on evaluation data: 8449 / 10000
Epoch 244 training complete
Cost on training data: 0.006456175961449934
Accuracy on evaluation data: 8450 / 10000
Epoch 245 training complete
Cost on training data: 0.006426848105220396
Accuracy on evaluation data: 8452 / 10000
Epoch 246 training complete
Cost on training data: 0.006398061937999
Accuracy on evaluation data: 8450 / 10000
Epoch 247 training complete
Cost on training data: 0.0063690181176364775
Accuracy on evaluation data: 8450 / 10000
Epoch 248 training complete
Cost on training data: 0.0063400142312427694
Accuracy on evaluation data: 8449 / 10000
Epoch 249 training complete
Cost on training data: 0.00631159377241229
Accuracy on evaluation data: 8449 / 10000
Epoch 250 training complete
Cost on training data: 0.006283580444020859
Accuracy on evaluation data: 8450 / 10000
Epoch 251 training complete
Cost on training data: 0.006256424756791315
Accuracy on evaluation data: 8450 / 10000
Epoch 252 training complete
Cost on training data: 0.006228133859257206
Accuracy on evaluation data: 8448 / 10000
Epoch 253 training complete
Cost on training data: 0.0062012656237021
Accuracy on evaluation data: 8448 / 10000
Epoch 254 training complete
Cost on training data: 0.006173386086059297
Accuracy on evaluation data: 8452 / 10000
Epoch 255 training complete
Cost on training data: 0.006146747634716472
Accuracy on evaluation data: 8448 / 10000
Epoch 256 training complete
Cost on training data: 0.006119690002133509
Accuracy on evaluation data: 8450 / 10000
Epoch 257 training complete
Cost on training data: 0.006093457222384939
Accuracy on evaluation data: 8448 / 10000
Epoch 258 training complete
Cost on training data: 0.0060670225519100715
Accuracy on evaluation data: 8449 / 10000
Epoch 259 training complete
Cost on training data: 0.006040967453545693
Accuracy on evaluation data: 8450 / 10000
Epoch 260 training complete
Cost on training data: 0.006016303393940634
Accuracy on evaluation data: 8449 / 10000
Epoch 261 training complete
Cost on training data: 0.005989868820592518
Accuracy on evaluation data: 8452 / 10000
Epoch 262 training complete
Cost on training data: 0.005965394043346562
Accuracy on evaluation data: 8449 / 10000
Epoch 263 training complete
Cost on training data: 0.005938647466701287
Accuracy on evaluation data: 8449 / 10000
Epoch 264 training complete
Cost on training data: 0.005913173471753071
Accuracy on evaluation data: 8451 / 10000
Epoch 265 training complete
Cost on training data: 0.00588804629659548
Accuracy on evaluation data: 8450 / 10000
Epoch 266 training complete
Cost on training data: 0.005863373702903854
Accuracy on evaluation data: 8451 / 10000
Epoch 267 training complete
Cost on training data: 0.005838688139919397
Accuracy on evaluation data: 8452 / 10000
Epoch 268 training complete
Cost on training data: 0.005814238787509755
Accuracy on evaluation data: 8450 / 10000
Epoch 269 training complete
Cost on training data: 0.005789700551838541
Accuracy on evaluation data: 8451 / 10000
Epoch 270 training complete
Cost on training data: 0.0057654899095211255
Accuracy on evaluation data: 8450 / 10000
Epoch 271 training complete
Cost on training data: 0.005741943688407352
Accuracy on evaluation data: 8452 / 10000
Epoch 272 training complete
Cost on training data: 0.005717660700910018
Accuracy on evaluation data: 8450 / 10000
Epoch 273 training complete
Cost on training data: 0.005694161758288953
Accuracy on evaluation data: 8451 / 10000
Epoch 274 training complete
Cost on training data: 0.005670904531158865
Accuracy on evaluation data: 8450 / 10000
Epoch 275 training complete
Cost on training data: 0.005647668912346308
Accuracy on evaluation data: 8447 / 10000
Epoch 276 training complete
Cost on training data: 0.005624116044239148
Accuracy on evaluation data: 8447 / 10000
Epoch 277 training complete
Cost on training data: 0.005601062368304801
Accuracy on evaluation data: 8448 / 10000
Epoch 278 training complete
Cost on training data: 0.00557800370514958
Accuracy on evaluation data: 8450 / 10000
Epoch 279 training complete
Cost on training data: 0.005555208869651533
Accuracy on evaluation data: 8450 / 10000
Epoch 280 training complete
Cost on training data: 0.0055328375244215085
Accuracy on evaluation data: 8454 / 10000
Epoch 281 training complete
Cost on training data: 0.005510002368060477
Accuracy on evaluation data: 8456 / 10000
Epoch 282 training complete
Cost on training data: 0.0054877717401872585
Accuracy on evaluation data: 8457 / 10000
Epoch 283 training complete
Cost on training data: 0.0054656957066184534
Accuracy on evaluation data: 8453 / 10000
Epoch 284 training complete
Cost on training data: 0.0054436599852498045
Accuracy on evaluation data: 8452 / 10000
Epoch 285 training complete
Cost on training data: 0.005421555530871548
Accuracy on evaluation data: 8453 / 10000
Epoch 286 training complete
Cost on training data: 0.005399528537032674
Accuracy on evaluation data: 8452 / 10000
Epoch 287 training complete
Cost on training data: 0.005377943555842746
Accuracy on evaluation data: 8451 / 10000
Epoch 288 training complete
Cost on training data: 0.005356354186241844
Accuracy on evaluation data: 8451 / 10000
Epoch 289 training complete
Cost on training data: 0.005334673757338759
Accuracy on evaluation data: 8451 / 10000
Epoch 290 training complete
Cost on training data: 0.005313056392103426
Accuracy on evaluation data: 8452 / 10000
Epoch 291 training complete
Cost on training data: 0.005291963549073837
Accuracy on evaluation data: 8452 / 10000
Epoch 292 training complete
Cost on training data: 0.005270880039216974
Accuracy on evaluation data: 8456 / 10000
Epoch 293 training complete
Cost on training data: 0.005249934586441436
Accuracy on evaluation data: 8450 / 10000
Epoch 294 training complete
Cost on training data: 0.005229237625108339
Accuracy on evaluation data: 8452 / 10000
Epoch 295 training complete
Cost on training data: 0.005208364767316311
Accuracy on evaluation data: 8452 / 10000
Epoch 296 training complete
Cost on training data: 0.005187662045800864
Accuracy on evaluation data: 8452 / 10000
Epoch 297 training complete
Cost on training data: 0.005167176146519333
Accuracy on evaluation data: 8452 / 10000
Epoch 298 training complete
Cost on training data: 0.005147164009746564
Accuracy on evaluation data: 8451 / 10000
Epoch 299 training complete
Cost on training data: 0.005126728309057134
Accuracy on evaluation data: 8450 / 10000
Epoch 300 training complete
Cost on training data: 0.005106603389814559
Accuracy on evaluation data: 8451 / 10000
Epoch 301 training complete
Cost on training data: 0.005086805388927815
Accuracy on evaluation data: 8453 / 10000
Epoch 302 training complete
Cost on training data: 0.005067189496995626
Accuracy on evaluation data: 8453 / 10000
Epoch 303 training complete
Cost on training data: 0.005047382562905665
Accuracy on evaluation data: 8454 / 10000
Epoch 304 training complete
Cost on training data: 0.005028034574598833
Accuracy on evaluation data: 8452 / 10000
Epoch 305 training complete
Cost on training data: 0.0050084702392899
Accuracy on evaluation data: 8453 / 10000
Epoch 306 training complete
Cost on training data: 0.004989487783001165
Accuracy on evaluation data: 8453 / 10000
Epoch 307 training complete
Cost on training data: 0.004970364946417611
Accuracy on evaluation data: 8454 / 10000
Epoch 308 training complete
Cost on training data: 0.0049517185197455625
Accuracy on evaluation data: 8454 / 10000
Epoch 309 training complete
Cost on training data: 0.004932909742147225
Accuracy on evaluation data: 8454 / 10000
Epoch 310 training complete
Cost on training data: 0.004914067483989186
Accuracy on evaluation data: 8456 / 10000
Epoch 311 training complete
Cost on training data: 0.004895392136366004
Accuracy on evaluation data: 8457 / 10000
Epoch 312 training complete
Cost on training data: 0.004877447404631756
Accuracy on evaluation data: 8456 / 10000
Epoch 313 training complete
Cost on training data: 0.004858917942767792
Accuracy on evaluation data: 8458 / 10000
Epoch 314 training complete
Cost on training data: 0.00484093082756921
Accuracy on evaluation data: 8456 / 10000
Epoch 315 training complete
Cost on training data: 0.004822977903949536
Accuracy on evaluation data: 8456 / 10000
Epoch 316 training complete
Cost on training data: 0.0048051048027296156
Accuracy on evaluation data: 8456 / 10000
Epoch 317 training complete
Cost on training data: 0.004787525172086458
Accuracy on evaluation data: 8456 / 10000
Epoch 318 training complete
Cost on training data: 0.004769962311108191
Accuracy on evaluation data: 8457 / 10000
Epoch 319 training complete
Cost on training data: 0.004752465103495572
Accuracy on evaluation data: 8456 / 10000
Epoch 320 training complete
Cost on training data: 0.004735117711139855
Accuracy on evaluation data: 8457 / 10000
Epoch 321 training complete
Cost on training data: 0.004718069015654681
Accuracy on evaluation data: 8460 / 10000
Epoch 322 training complete
Cost on training data: 0.004700954567374447
Accuracy on evaluation data: 8456 / 10000
Epoch 323 training complete
Cost on training data: 0.004684125026718987
Accuracy on evaluation data: 8457 / 10000
Epoch 324 training complete
Cost on training data: 0.004667369207845105
Accuracy on evaluation data: 8459 / 10000
Epoch 325 training complete
Cost on training data: 0.004650662161445141
Accuracy on evaluation data: 8456 / 10000
Epoch 326 training complete
Cost on training data: 0.004634210332208765
Accuracy on evaluation data: 8457 / 10000
Epoch 327 training complete
Cost on training data: 0.0046179946139440755
Accuracy on evaluation data: 8456 / 10000
Epoch 328 training complete
Cost on training data: 0.0046015235949743655
Accuracy on evaluation data: 8456 / 10000
Epoch 329 training complete
Cost on training data: 0.004585486665367072
Accuracy on evaluation data: 8458 / 10000
Epoch 330 training complete
Cost on training data: 0.004569374762516196
Accuracy on evaluation data: 8459 / 10000
Epoch 331 training complete
Cost on training data: 0.004553359198725704
Accuracy on evaluation data: 8457 / 10000
Epoch 332 training complete
Cost on training data: 0.004537789650018289
Accuracy on evaluation data: 8455 / 10000
Epoch 333 training complete
Cost on training data: 0.004521863397284219
Accuracy on evaluation data: 8456 / 10000
Epoch 334 training complete
Cost on training data: 0.0045063059547557245
Accuracy on evaluation data: 8454 / 10000
Epoch 335 training complete
Cost on training data: 0.004490830953853623
Accuracy on evaluation data: 8454 / 10000
Epoch 336 training complete
Cost on training data: 0.004475399270124156
Accuracy on evaluation data: 8454 / 10000
Epoch 337 training complete
Cost on training data: 0.0044601104423474735
Accuracy on evaluation data: 8454 / 10000
Epoch 338 training complete
Cost on training data: 0.004444892712063275
Accuracy on evaluation data: 8454 / 10000
Epoch 339 training complete
Cost on training data: 0.0044299373366120266
Accuracy on evaluation data: 8455 / 10000
Epoch 340 training complete
Cost on training data: 0.00441495550594844
Accuracy on evaluation data: 8454 / 10000
Epoch 341 training complete
Cost on training data: 0.00439999573863545
Accuracy on evaluation data: 8455 / 10000
Epoch 342 training complete
Cost on training data: 0.0043851796595118855
Accuracy on evaluation data: 8454 / 10000
Epoch 343 training complete
Cost on training data: 0.004370509967274126
Accuracy on evaluation data: 8453 / 10000
Epoch 344 training complete
Cost on training data: 0.004356001437446666
Accuracy on evaluation data: 8454 / 10000
Epoch 345 training complete
Cost on training data: 0.004341476075756548
Accuracy on evaluation data: 8454 / 10000
Epoch 346 training complete
Cost on training data: 0.004327375344027786
Accuracy on evaluation data: 8458 / 10000
Epoch 347 training complete
Cost on training data: 0.004312816833407688
Accuracy on evaluation data: 8455 / 10000
Epoch 348 training complete
Cost on training data: 0.004298550589725273
Accuracy on evaluation data: 8453 / 10000
Epoch 349 training complete
Cost on training data: 0.004284437102809583
Accuracy on evaluation data: 8453 / 10000
Epoch 350 training complete
Cost on training data: 0.004270396041901047
Accuracy on evaluation data: 8455 / 10000
Epoch 351 training complete
Cost on training data: 0.0042564485888378
Accuracy on evaluation data: 8454 / 10000
Epoch 352 training complete
Cost on training data: 0.004242700773955589
Accuracy on evaluation data: 8457 / 10000
Epoch 353 training complete
Cost on training data: 0.004228973608756927
Accuracy on evaluation data: 8456 / 10000
Epoch 354 training complete
Cost on training data: 0.004215150732240222
Accuracy on evaluation data: 8455 / 10000
Epoch 355 training complete
Cost on training data: 0.004201686478348546
Accuracy on evaluation data: 8458 / 10000
Epoch 356 training complete
Cost on training data: 0.004188213373837653
Accuracy on evaluation data: 8458 / 10000
Epoch 357 training complete
Cost on training data: 0.004174708057344338
Accuracy on evaluation data: 8455 / 10000
Epoch 358 training complete
Cost on training data: 0.004161246420505441
Accuracy on evaluation data: 8455 / 10000
Epoch 359 training complete
Cost on training data: 0.004147894111635248
Accuracy on evaluation data: 8455 / 10000
Epoch 360 training complete
Cost on training data: 0.004134682185277353
Accuracy on evaluation data: 8455 / 10000
Epoch 361 training complete
Cost on training data: 0.004121624502562014
Accuracy on evaluation data: 8458 / 10000
Epoch 362 training complete
Cost on training data: 0.004108598803852993
Accuracy on evaluation data: 8455 / 10000
Epoch 363 training complete
Cost on training data: 0.004095710810742628
Accuracy on evaluation data: 8456 / 10000
Epoch 364 training complete
Cost on training data: 0.004082718920143551
Accuracy on evaluation data: 8457 / 10000
Epoch 365 training complete
Cost on training data: 0.004070051476235988
Accuracy on evaluation data: 8458 / 10000
Epoch 366 training complete
Cost on training data: 0.004056938332299395
Accuracy on evaluation data: 8456 / 10000
Epoch 367 training complete
Cost on training data: 0.0040443297875509
Accuracy on evaluation data: 8456 / 10000
Epoch 368 training complete
Cost on training data: 0.0040318003206056005
Accuracy on evaluation data: 8456 / 10000
Epoch 369 training complete
Cost on training data: 0.004019181753384823
Accuracy on evaluation data: 8457 / 10000
Epoch 370 training complete
Cost on training data: 0.004006705267628003
Accuracy on evaluation data: 8457 / 10000
Epoch 371 training complete
Cost on training data: 0.003994326146730657
Accuracy on evaluation data: 8457 / 10000
Epoch 372 training complete
Cost on training data: 0.003981902892525214
Accuracy on evaluation data: 8458 / 10000
Epoch 373 training complete
Cost on training data: 0.00396964517256825
Accuracy on evaluation data: 8457 / 10000
Epoch 374 training complete
Cost on training data: 0.003957553909764387
Accuracy on evaluation data: 8458 / 10000
Epoch 375 training complete
Cost on training data: 0.003945315482560462
Accuracy on evaluation data: 8457 / 10000
Epoch 376 training complete
Cost on training data: 0.003933276759706994
Accuracy on evaluation data: 8457 / 10000
Epoch 377 training complete
Cost on training data: 0.003921424852837589
Accuracy on evaluation data: 8459 / 10000
Epoch 378 training complete
Cost on training data: 0.003909363708513628
Accuracy on evaluation data: 8459 / 10000
Epoch 379 training complete
Cost on training data: 0.00389741880422459
Accuracy on evaluation data: 8457 / 10000
Epoch 380 training complete
Cost on training data: 0.003885686878142363
Accuracy on evaluation data: 8459 / 10000
Epoch 381 training complete
Cost on training data: 0.003873948097592198
Accuracy on evaluation data: 8460 / 10000
Epoch 382 training complete
Cost on training data: 0.0038623321179242914
Accuracy on evaluation data: 8461 / 10000
Epoch 383 training complete
Cost on training data: 0.0038507623652346496
Accuracy on evaluation data: 8462 / 10000
Epoch 384 training complete
Cost on training data: 0.003839250421204888
Accuracy on evaluation data: 8462 / 10000
Epoch 385 training complete
Cost on training data: 0.0038278098301889668
Accuracy on evaluation data: 8462 / 10000
Epoch 386 training complete
Cost on training data: 0.003816341730791096
Accuracy on evaluation data: 8462 / 10000
Epoch 387 training complete
Cost on training data: 0.003805058387591798
Accuracy on evaluation data: 8462 / 10000
Epoch 388 training complete
Cost on training data: 0.003793795821766855
Accuracy on evaluation data: 8464 / 10000
Epoch 389 training complete
Cost on training data: 0.0037826058079641384
Accuracy on evaluation data: 8463 / 10000
Epoch 390 training complete
Cost on training data: 0.003771464060147009
Accuracy on evaluation data: 8463 / 10000
Epoch 391 training complete
Cost on training data: 0.0037604229342176718
Accuracy on evaluation data: 8464 / 10000
Epoch 392 training complete
Cost on training data: 0.0037494234527342524
Accuracy on evaluation data: 8463 / 10000
Epoch 393 training complete
Cost on training data: 0.003738512485251956
Accuracy on evaluation data: 8463 / 10000
Epoch 394 training complete
Cost on training data: 0.003727631899579603
Accuracy on evaluation data: 8463 / 10000
Epoch 395 training complete
Cost on training data: 0.003716802182262589
Accuracy on evaluation data: 8462 / 10000
Epoch 396 training complete
Cost on training data: 0.003706068792614434
Accuracy on evaluation data: 8462 / 10000
Epoch 397 training complete
Cost on training data: 0.0036954481497339015
Accuracy on evaluation data: 8463 / 10000
Epoch 398 training complete
Cost on training data: 0.0036848192810071934
Accuracy on evaluation data: 8463 / 10000
Epoch 399 training complete
Cost on training data: 0.0036742418585741245
Accuracy on evaluation data: 8464 / 10000
CPU times: user 7min 5s, sys: 5min 31s, total: 12min 37s
Wall time: 7min 35s

([],
 [5362,
  6508,
  6996,
  7367,
  7592,
  7708,
  7752,
  7931,
  8037,
  8080,
  8136,
  8149,
  8160,
  8255,
  8258,
  8267,
  8277,
  8281,
  8276,
  8293,
  8319,
  8324,
  8324,
  8370,
  8351,
  8344,
  8345,
  8352,
  8348,
  8374,
  8386,
  8360,
  8361,
  8381,
  8365,
  8376,
  8390,
  8399,
  8389,
  8398,
  8382,
  8400,
  8394,
  8398,
  8388,
  8392,
  8400,
  8396,
  8410,
  8402,
  8398,
  8405,
  8403,
  8399,
  8410,
  8401,
  8403,
  8405,
  8410,
  8416,
  8408,
  8411,
  8410,
  8413,
  8404,
  8417,
  8418,
  8417,
  8420,
  8412,
  8411,
  8420,
  8410,
  8421,
  8420,
  8426,
  8410,
  8423,
  8431,
  8421,
  8424,
  8434,
  8433,
  8437,
  8419,
  8435,
  8435,
  8424,
  8435,
  8437,
  8441,
  8441,
  8430,
  8443,
  8448,
  8441,
  8441,
  8437,
  8448,
  8444,
  8447,
  8448,
  8445,
  8443,
  8454,
  8446,
  8446,
  8444,
  8441,
  8446,
  8447,
  8444,
  8449,
  8446,
  8442,
  8443,
  8446,
  8444,
  8447,
  8439,
  8439,
  8447,
  8444,
  8444,
  8443,
  8441,
  8447,
  8451,
  8444,
  8447,
  8449,
  8447,
  8448,
  8448,
  8445,
  8447,
  8447,
  8448,
  8448,
  8449,
  8448,
  8447,
  8448,
  8448,
  8452,
  8447,
  8447,
  8447,
  8445,
  8450,
  8452,
  8451,
  8444,
  8442,
  8448,
  8446,
  8446,
  8444,
  8445,
  8447,
  8448,
  8450,
  8445,
  8445,
  8443,
  8445,
  8445,
  8445,
  8448,
  8447,
  8449,
  8444,
  8443,
  8444,
  8452,
  8449,
  8446,
  8449,
  8445,
  8444,
  8442,
  8446,
  8447,
  8447,
  8446,
  8444,
  8446,
  8445,
  8449,
  8447,
  8443,
  8443,
  8442,
  8447,
  8448,
  8442,
  8448,
  8446,
  8449,
  8446,
  8444,
  8448,
  8447,
  8445,
  8444,
  8447,
  8443,
  8446,
  8447,
  8446,
  8448,
  8450,
  8446,
  8447,
  8447,
  8449,
  8448,
  8449,
  8449,
  8449,
  8451,
  8448,
  8445,
  8446,
  8446,
  8450,
  8450,
  8450,
  8451,
  8451,
  8446,
  8449,
  8451,
  8452,
  8453,
  8450,
  8451,
  8454,
  8451,
  8453,
  8450,
  8451,
  8455,
  8449,
  8450,
  8452,
  8450,
  8450,
  8449,
  8449,
  8450,
  8450,
  8448,
  8448,
  8452,
  8448,
  8450,
  8448,
  8449,
  8450,
  8449,
  8452,
  8449,
  8449,
  8451,
  8450,
  8451,
  8452,
  8450,
  8451,
  8450,
  8452,
  8450,
  8451,
  8450,
  8447,
  8447,
  8448,
  8450,
  8450,
  8454,
  8456,
  8457,
  8453,
  8452,
  8453,
  8452,
  8451,
  8451,
  8451,
  8452,
  8452,
  8456,
  8450,
  8452,
  8452,
  8452,
  8452,
  8451,
  8450,
  8451,
  8453,
  8453,
  8454,
  8452,
  8453,
  8453,
  8454,
  8454,
  8454,
  8456,
  8457,
  8456,
  8458,
  8456,
  8456,
  8456,
  8456,
  8457,
  8456,
  8457,
  8460,
  8456,
  8457,
  8459,
  8456,
  8457,
  8456,
  8456,
  8458,
  8459,
  8457,
  8455,
  8456,
  8454,
  8454,
  8454,
  8454,
  8454,
  8455,
  8454,
  8455,
  8454,
  8453,
  8454,
  8454,
  8458,
  8455,
  8453,
  8453,
  8455,
  8454,
  8457,
  8456,
  8455,
  8458,
  8458,
  8455,
  8455,
  8455,
  8455,
  8458,
  8455,
  8456,
  8457,
  8458,
  8456,
  8456,
  8456,
  8457,
  8457,
  8457,
  8458,
  8457,
  8458,
  8457,
  8457,
  8459,
  8459,
  8457,
  8459,
  8460,
  8461,
  8462,
  8462,
  8462,
  8462,
  8462,
  8464,
  8463,
  8463,
  8464,
  8463,
  8463,
  8463,
  8462,
  8462,
  8463,
  8463,
  8464],
 [2.1334873081011976,
  1.6117154896081094,
  1.3435858874341642,
  1.118955170706648,
  0.9901307112582952,
  0.8519259734137855,
  0.7960956201359782,
  0.6720702851986642,
  0.5998755333945173,
  0.5532262347451836,
  0.48301226958062626,
  0.43713906935077373,
  0.4054099079755518,
  0.37177860867561674,
  0.3347236482029461,
  0.3071926589220765,
  0.27895724476902123,
  0.2652053196232567,
  0.24971868688894489,
  0.22578091792710658,
  0.21674447452601472,
  0.19696616273484055,
  0.18367647407713247,
  0.1697081997249623,
  0.15694410574943818,
  0.14931148314777995,
  0.1376326060327698,
  0.12822834529691673,
  0.12334225702609936,
  0.11344356383141305,
  0.10874642376376649,
  0.10232741359823394,
  0.09800310442716333,
  0.09268095912167088,
  0.08985211317214391,
  0.08529139016670571,
  0.08354889001668293,
  0.08003507351151332,
  0.07650548159073085,
  0.0735963249949398,
  0.07054144890418132,
  0.06887572297685532,
  0.0663065350160309,
  0.06408076105973032,
  0.061858574556072424,
  0.06058942537298044,
  0.05873446242587899,
  0.05593898642402481,
  0.055205914338667846,
  0.05313535099205961,
  0.051116209408344354,
  0.049505308014994136,
  0.04786572144766827,
  0.04640639733502174,
  0.045380854077187775,
  0.04412889015222005,
  0.04290577626118409,
  0.04194561198588216,
  0.04079484197387679,
  0.039718523191149574,
  0.038891299256091943,
  0.03801167501946484,
  0.03700659233294767,
  0.03619008401446996,
  0.0354101234112211,
  0.03446877877563418,
  0.03371575942835778,
  0.032671918060232265,
  0.03191790922420547,
  0.031155109011162536,
  0.030581435138702377,
  0.029850450319083684,
  0.02927639825367843,
  0.028877550454782647,
  0.028148984621107827,
  0.027645446138674794,
  0.027157308884780768,
  0.02660532840891866,
  0.02630193671251622,
  0.025685865680059764,
  0.025230066226201827,
  0.024826864455365017,
  0.0243866839572904,
  0.02399530254399858,
  0.023730621576772862,
  0.02325836603295703,
  0.02289041133860511,
  0.022542631111450415,
  0.02219589347435611,
  0.021936901953618273,
  0.02157206047399831,
  0.021223795818202454,
  0.02096752615216147,
  0.02063715475299584,
  0.020356029193518026,
  0.020072665857854867,
  0.019834001926711618,
  0.01954121858171723,
  0.019286776431325092,
  0.01905240925628857,
  0.01880649215622303,
  0.01854682840467356,
  0.018399883172119994,
  0.018090713620616473,
  0.01789813825489991,
  0.01765118746282579,
  0.017473085175173292,
  0.017244494927694504,
  0.01703486435487612,
  0.016859129978839392,
  0.01664474198599423,
  0.016492894164797507,
  0.016277872550429357,
  0.01609761470488398,
  0.01591808120014808,
  0.01574090632673071,
  0.015569273751146593,
  0.015401941782147353,
  0.015238524446743831,
  0.015108287488949757,
  0.014915727919302634,
  0.014765483792022284,
  0.014608352572538171,
  0.014466773855429526,
  0.014318795347529108,
  0.014166768716930093,
  0.01403021765009755,
  0.013889891280038329,
  0.013756143033653877,
  0.013620655462519938,
  0.013493108647379412,
  0.013364012980409488,
  0.013238234613230631,
  0.013117270271950057,
  0.01299690578155512,
  0.012882657709243744,
  0.012758697885839247,
  0.012649429525397628,
  0.012538584374843613,
  0.012422775590650628,
  0.012319840609501332,
  0.012209649003281678,
  0.012102365427467381,
  0.012000446070765454,
  0.011904631671549513,
  0.011805864322997753,
  0.011698740239416737,
  0.011606371241906156,
  0.011514374611308427,
  0.011416940179436775,
  0.011324325638086047,
  0.011232413314906249,
  0.011143023721712417,
  0.011057575745872053,
  0.010968259739096101,
  0.01088128644534935,
  0.010799772166372307,
  0.010717817480123754,
  0.010638889630533489,
  0.010554400509895827,
  0.010475138359528497,
  0.010397616721467969,
  0.010321260115164373,
  0.010244649482948473,
  0.010170491415277361,
  0.01009958618454486,
  0.010024345542467384,
  0.009953348266001014,
  0.009883756280116205,
  0.009811861956010477,
  0.009749156561601519,
  0.009679778377319816,
  0.009609249102981717,
  0.009544288870302616,
  0.009480519623452777,
  0.009416154781809356,
  0.009354569506204469,
  0.00929067436973555,
  0.009229670289513256,
  0.009169356854431477,
  0.009111816213489487,
  0.0090543641894766,
  0.008993560686706871,
  0.008938581789090341,
  0.008880962419123395,
  0.008826301190224393,
  0.008769672548831493,
  0.008716483735853068,
  0.00866460369160469,
  0.00861036903221576,
  0.008557577345388928,
  0.00850933448370844,
  0.008456169763878356,
  0.008404662640716853,
  0.008354622029987917,
  0.00830697908651364,
  0.008257180529236265,
  0.008210014573194865,
  0.008162098131249284,
  0.00811539010760783,
  0.008072431304794821,
  0.00802376202280493,
  0.007979104863057019,
  0.007937625159070238,
  0.007890554712392644,
  0.00784790016459241,
  0.007802929222624051,
  0.007760549692860105,
  0.007718367124741382,
  0.007676850015661369,
  0.007636902480344225,
  0.007594654515272643,
  0.007557266065446237,
  0.007517549897936283,
  0.007475327712505067,
  0.007435933824650646,
  0.007397819828337148,
  0.007359477880672005,
  0.0073208837720296705,
  0.007283805820236689,
  0.007247153976131099,
  0.0072106084576830706,
  0.0071737567683653304,
  0.007137721603632793,
  0.0071018808134562916,
  0.007066331622788004,
  0.007031389918120413,
  0.006996545294324694,
  0.00696280808553848,
  0.0069292201735108635,
  0.0068955718106028935,
  0.006861764530310163,
  0.006828683516656813,
  0.006796285258697305,
  0.006764736003988712,
  0.006731546513205398,
  0.006700374780671107,
  0.006669692110701232,
  0.0066375050596078275,
  0.006607049930111248,
  0.006575578731710099,
  0.006545454671152349,
  0.0065158945218681775,
  0.006485383413394594,
  0.006456175961449934,
  0.006426848105220396,
  0.006398061937999,
  0.0063690181176364775,
  0.0063400142312427694,
  0.00631159377241229,
  0.006283580444020859,
  0.006256424756791315,
  0.006228133859257206,
  0.0062012656237021,
  0.006173386086059297,
  0.006146747634716472,
  0.006119690002133509,
  0.006093457222384939,
  0.0060670225519100715,
  0.006040967453545693,
  0.006016303393940634,
  0.005989868820592518,
  0.005965394043346562,
  0.005938647466701287,
  0.005913173471753071,
  0.00588804629659548,
  0.005863373702903854,
  0.005838688139919397,
  0.005814238787509755,
  0.005789700551838541,
  0.0057654899095211255,
  0.005741943688407352,
  0.005717660700910018,
  0.005694161758288953,
  0.005670904531158865,
  0.005647668912346308,
  0.005624116044239148,
  0.005601062368304801,
  0.00557800370514958,
  0.005555208869651533,
  0.0055328375244215085,
  0.005510002368060477,
  0.0054877717401872585,
  0.0054656957066184534,
  0.0054436599852498045,
  0.005421555530871548,
  0.005399528537032674,
  0.005377943555842746,
  0.005356354186241844,
  0.005334673757338759,
  0.005313056392103426,
  0.005291963549073837,
  0.005270880039216974,
  0.005249934586441436,
  0.005229237625108339,
  0.005208364767316311,
  0.005187662045800864,
  0.005167176146519333,
  0.005147164009746564,
  0.005126728309057134,
  0.005106603389814559,
  0.005086805388927815,
  0.005067189496995626,
  0.005047382562905665,
  0.005028034574598833,
  0.0050084702392899,
  0.004989487783001165,
  0.004970364946417611,
  0.0049517185197455625,
  0.004932909742147225,
  0.004914067483989186,
  0.004895392136366004,
  0.004877447404631756,
  0.004858917942767792,
  0.00484093082756921,
  0.004822977903949536,
  0.0048051048027296156,
  0.004787525172086458,
  0.004769962311108191,
  0.004752465103495572,
  0.004735117711139855,
  0.004718069015654681,
  0.004700954567374447,
  0.004684125026718987,
  0.004667369207845105,
  0.004650662161445141,
  0.004634210332208765,
  0.0046179946139440755,
  0.0046015235949743655,
  0.004585486665367072,
  0.004569374762516196,
  0.004553359198725704,
  0.004537789650018289,
  0.004521863397284219,
  0.0045063059547557245,
  0.004490830953853623,
  0.004475399270124156,
  0.0044601104423474735,
  0.004444892712063275,
  0.0044299373366120266,
  0.00441495550594844,
  0.00439999573863545,
  0.0043851796595118855,
  0.004370509967274126,
  0.004356001437446666,
  0.004341476075756548,
  0.004327375344027786,
  0.004312816833407688,
  0.004298550589725273,
  0.004284437102809583,
  0.004270396041901047,
  0.0042564485888378,
  0.004242700773955589,
  0.004228973608756927,
  0.004215150732240222,
  0.004201686478348546,
  0.004188213373837653,
  0.004174708057344338,
  0.004161246420505441,
  0.004147894111635248,
  0.004134682185277353,
  0.004121624502562014,
  0.004108598803852993,
  0.004095710810742628,
  0.004082718920143551,
  0.004070051476235988,
  0.004056938332299395,
  0.0040443297875509,
  0.0040318003206056005,
  0.004019181753384823,
  0.004006705267628003,
  0.003994326146730657,
  0.003981902892525214,
  0.00396964517256825,
  0.003957553909764387,
  0.003945315482560462,
  0.003933276759706994,
  0.003921424852837589,
  0.003909363708513628,
  0.00389741880422459,
  0.003885686878142363,
  0.003873948097592198,
  0.0038623321179242914,
  0.0038507623652346496,
  0.003839250421204888,
  0.0038278098301889668,
  0.003816341730791096,
  0.003805058387591798,
  0.003793795821766855,
  0.0037826058079641384,
  0.003771464060147009,
  0.0037604229342176718,
  0.0037494234527342524,
  0.003738512485251956,
  0.003727631899579603,
  0.003716802182262589,
  0.003706068792614434,
  0.0036954481497339015,
  0.0036848192810071934,
  0.0036742418585741245],
 [])

# @title
"""
overfitting
~~~~~~~~~~~

Plot graphs to illustrate the problem of overfitting.
"""

# Standard library
import json
import random
import sys

# My library
sys.path.append('../src/')
import mnist_loader
# import network2

# Third-party libraries
import matplotlib.pyplot as plt
import numpy as np


def main(filename, num_epochs,
         training_cost_xmin=200,
         test_accuracy_xmin=200,
         test_cost_xmin=0,
         training_accuracy_xmin=0,
         training_set_size=1000,
         lmbda=0.0):
    """``filename`` is the name of the file where the results will be
    stored.  ``num_epochs`` is the number of epochs to train for.
    ``training_set_size`` is the number of images to train on.
    ``lmbda`` is the regularization parameter.  The other parameters
    set the epochs at which to start plotting on the x axis.
    """
    run_network(filename, num_epochs, training_set_size, lmbda)
    make_plots(filename, num_epochs,
               training_cost_xmin,
               test_accuracy_xmin,
               test_cost_xmin,
               training_accuracy_xmin,
               training_set_size)

def run_network(filename, num_epochs, training_set_size=1000, lmbda=0.0):
    """Train the network for ``num_epochs`` on ``training_set_size``
    images, and store the results in ``filename``.  Those results can
    later be used by ``make_plots``.  Note that the results are stored
    to disk in large part because it's convenient not to have to
    ``run_network`` each time we want to make a plot (it's slow).

    """
    # Make results more easily reproducible
    random.seed(12345678)
    np.random.seed(12345678)
    training_data, validation_data, test_data = mnist_loader.load_data_wrapper()

    # Convert data to lists
    training_data = list(training_data)
    validation_data = list(validation_data)
    test_data = list(test_data)

    net = Network2mat([784, 30, 10], cost=CrossEntropyCost())
    net.large_weight_initializer()
    test_cost, test_accuracy, training_cost, training_accuracy \
        = net.SGD(training_data[:training_set_size], num_epochs, 10, 0.5,
                  evaluation_data=test_data, lmbda = lmbda,
                  monitor_evaluation_cost=True,
                  monitor_evaluation_accuracy=True,
                  monitor_training_cost=True,
                  monitor_training_accuracy=True)
    f = open(filename, "w")
    json.dump([test_cost, test_accuracy, training_cost, training_accuracy], f)
    f.close()

def make_plots(filename, num_epochs,
               training_cost_xmin=200,
               test_accuracy_xmin=200,
               test_cost_xmin=0,
               training_accuracy_xmin=0,
               training_set_size=1000):
    """Load the results from ``filename``, and generate the corresponding
    plots. """
    f = open(filename, "r")
    test_cost, test_accuracy, training_cost, training_accuracy \
        = json.load(f)
    f.close()
    plot_training_cost(training_cost, num_epochs, training_cost_xmin)
    plot_test_accuracy(test_accuracy, num_epochs, test_accuracy_xmin)
    plot_test_cost(test_cost, num_epochs, test_cost_xmin)
    plot_training_accuracy(training_accuracy, num_epochs,
                           training_accuracy_xmin, training_set_size)
    plot_overlay(test_accuracy, training_accuracy, num_epochs,
                 min(test_accuracy_xmin, training_accuracy_xmin),
                 training_set_size)

def plot_training_cost(training_cost, num_epochs, training_cost_xmin):
    fig = plt.figure()
    ax = fig.add_subplot(111)
    ax.plot(np.arange(training_cost_xmin, num_epochs),
            training_cost[training_cost_xmin:num_epochs],
            color='#2A6EA6')
    ax.set_xlim([training_cost_xmin, num_epochs])
    ax.grid(True)
    ax.set_xlabel('Epoch')
    ax.set_title('Cost on the training data')
    plt.show()

def plot_test_accuracy(test_accuracy, num_epochs, test_accuracy_xmin):
    fig = plt.figure()
    ax = fig.add_subplot(111)
    ax.plot(np.arange(test_accuracy_xmin, num_epochs),
            [accuracy/100.0
             for accuracy in test_accuracy[test_accuracy_xmin:num_epochs]],
            color='#2A6EA6')
    ax.set_xlim([test_accuracy_xmin, num_epochs])
    ax.grid(True)
    ax.set_xlabel('Epoch')
    ax.set_title('Accuracy (%) on the test data')
    plt.show()

def plot_test_cost(test_cost, num_epochs, test_cost_xmin):
    fig = plt.figure()
    ax = fig.add_subplot(111)
    ax.plot(np.arange(test_cost_xmin, num_epochs),
            test_cost[test_cost_xmin:num_epochs],
            color='#2A6EA6')
    ax.set_xlim([test_cost_xmin, num_epochs])
    ax.grid(True)
    ax.set_xlabel('Epoch')
    ax.set_title('Cost on the test data')
    plt.show()

def plot_training_accuracy(training_accuracy, num_epochs,
                           training_accuracy_xmin, training_set_size):
    fig = plt.figure()
    ax = fig.add_subplot(111)
    ax.plot(np.arange(training_accuracy_xmin, num_epochs),
            [accuracy*100.0/training_set_size
             for accuracy in training_accuracy[training_accuracy_xmin:num_epochs]],
            color='#2A6EA6')
    ax.set_xlim([training_accuracy_xmin, num_epochs])
    ax.grid(True)
    ax.set_xlabel('Epoch')
    ax.set_title('Accuracy (%) on the training data')
    plt.show()

def plot_overlay(test_accuracy, training_accuracy, num_epochs, xmin,
                 training_set_size):
    fig = plt.figure()
    ax = fig.add_subplot(111)
    ax.plot(np.arange(xmin, num_epochs),
            [accuracy/100.0 for accuracy in test_accuracy],
            color='#2A6EA6',
            label="Accuracy on the test data")
    ax.plot(np.arange(xmin, num_epochs),
            [accuracy*100.0/training_set_size
             for accuracy in training_accuracy],
            color='#FFA933',
            label="Accuracy on the training data")
    ax.grid(True)
    ax.set_xlim([xmin, num_epochs])
    ax.set_xlabel('Epoch')
    ax.set_ylim([90, 100])
    plt.legend(loc="lower right")
    plt.show()

# @title
%%time
if __name__ == "__main__":
    filename = input("Enter a file name: ")
    num_epochs = int(input("Enter the number of epochs to run for: "))
    training_cost_xmin = int(input("training_cost_xmin (suggest 200): "))
    test_accuracy_xmin = int(input("test_accuracy_xmin (suggest 200): "))
    test_cost_xmin = int(input("test_cost_xmin (suggest 0): "))
    training_accuracy_xmin = int(input("training_accuracy_xmin (suggest 0): "))
    training_set_size = int(input("Training set size (suggest 1000): "))
    lmbda = float(input("Enter the regularization parameter, lambda (suggest: 5.0): "))
    main(filename, num_epochs, training_cost_xmin,
         test_accuracy_xmin, test_cost_xmin, training_accuracy_xmin,
         training_set_size, lmbda)

Enter a file name: train1000-2
Enter the number of epochs to run for: 400
training_cost_xmin (suggest 200): 200
test_accuracy_xmin (suggest 200): 200
test_cost_xmin (suggest 0): 0
training_accuracy_xmin (suggest 0): 0
Training set size (suggest 1000): 1000
Enter the regularization parameter, lambda (suggest: 5.0): 0
Epoch 0 training complete
Cost on training data: 1.9075988236669716
Accuracy on training data: 654 / 1000
Cost on evaluation data: 2.15247583066795
Accuracy on evaluation data: 5689 / 10000
Epoch 1 training complete
Cost on training data: 1.3963045571853507
Accuracy on training data: 780 / 1000
Cost on evaluation data: 1.7751670577479657
Accuracy on evaluation data: 6621 / 10000
Epoch 2 training complete
Cost on training data: 1.1861623435401252
Accuracy on training data: 803 / 1000
Cost on evaluation data: 1.5886885172771115
Accuracy on evaluation data: 6936 / 10000
Epoch 3 training complete
Cost on training data: 0.957877685241931
Accuracy on training data: 863 / 1000
Cost on evaluation data: 1.4534095873742074
Accuracy on evaluation data: 7408 / 10000
Epoch 4 training complete
Cost on training data: 0.8336335846220341
Accuracy on training data: 889 / 1000
Cost on evaluation data: 1.379012913103051
Accuracy on evaluation data: 7542 / 10000
Epoch 5 training complete
Cost on training data: 0.7522116437245072
Accuracy on training data: 912 / 1000
Cost on evaluation data: 1.351870710205931
Accuracy on evaluation data: 7597 / 10000
Epoch 6 training complete
Cost on training data: 0.6413963637964828
Accuracy on training data: 930 / 1000
Cost on evaluation data: 1.2747740134443177
Accuracy on evaluation data: 7764 / 10000
Epoch 7 training complete
Cost on training data: 0.5681974347964177
Accuracy on training data: 938 / 1000
Cost on evaluation data: 1.2447921820714714
Accuracy on evaluation data: 7807 / 10000
Epoch 8 training complete
Cost on training data: 0.5226136255046645
Accuracy on training data: 945 / 1000
Cost on evaluation data: 1.230372968043675
Accuracy on evaluation data: 7867 / 10000
Epoch 9 training complete
Cost on training data: 0.4717990360357309
Accuracy on training data: 955 / 1000
Cost on evaluation data: 1.230570058846599
Accuracy on evaluation data: 7906 / 10000
Epoch 10 training complete
Cost on training data: 0.4184664793143509
Accuracy on training data: 960 / 1000
Cost on evaluation data: 1.2366302986961217
Accuracy on evaluation data: 7896 / 10000
Epoch 11 training complete
Cost on training data: 0.41347693897190724
Accuracy on training data: 962 / 1000
Cost on evaluation data: 1.2208620271618362
Accuracy on evaluation data: 7901 / 10000
Epoch 12 training complete
Cost on training data: 0.3521037939208232
Accuracy on training data: 967 / 1000
Cost on evaluation data: 1.2121660360660207
Accuracy on evaluation data: 7955 / 10000
Epoch 13 training complete
Cost on training data: 0.32089877700564823
Accuracy on training data: 971 / 1000
Cost on evaluation data: 1.1979290776913063
Accuracy on evaluation data: 8004 / 10000
Epoch 14 training complete
Cost on training data: 0.3029826295627433
Accuracy on training data: 976 / 1000
Cost on evaluation data: 1.2022012227364014
Accuracy on evaluation data: 7999 / 10000
Epoch 15 training complete
Cost on training data: 0.27898399755520775
Accuracy on training data: 975 / 1000
Cost on evaluation data: 1.2030782451273618
Accuracy on evaluation data: 8046 / 10000
Epoch 16 training complete
Cost on training data: 0.25395570439554827
Accuracy on training data: 980 / 1000
Cost on evaluation data: 1.1804909392296614
Accuracy on evaluation data: 8070 / 10000
Epoch 17 training complete
Cost on training data: 0.2390643036330512
Accuracy on training data: 980 / 1000
Cost on evaluation data: 1.1844694690376982
Accuracy on evaluation data: 8077 / 10000
Epoch 18 training complete
Cost on training data: 0.220000346242534
Accuracy on training data: 980 / 1000
Cost on evaluation data: 1.1959815107293714
Accuracy on evaluation data: 8060 / 10000
Epoch 19 training complete
Cost on training data: 0.20275966619724267
Accuracy on training data: 986 / 1000
Cost on evaluation data: 1.1889368752266372
Accuracy on evaluation data: 8107 / 10000
Epoch 20 training complete
Cost on training data: 0.18974717206036806
Accuracy on training data: 984 / 1000
Cost on evaluation data: 1.1908627012916029
Accuracy on evaluation data: 8116 / 10000
Epoch 21 training complete
Cost on training data: 0.17978632075415268
Accuracy on training data: 988 / 1000
Cost on evaluation data: 1.206416371201434
Accuracy on evaluation data: 8138 / 10000
Epoch 22 training complete
Cost on training data: 0.1696180191589008
Accuracy on training data: 986 / 1000
Cost on evaluation data: 1.200500528937712
Accuracy on evaluation data: 8133 / 10000
Epoch 23 training complete
Cost on training data: 0.16066829167854627
Accuracy on training data: 990 / 1000
Cost on evaluation data: 1.1916032495370887
Accuracy on evaluation data: 8125 / 10000
Epoch 24 training complete
Cost on training data: 0.15979479548060732
Accuracy on training data: 990 / 1000
Cost on evaluation data: 1.2142359462462253
Accuracy on evaluation data: 8084 / 10000
Epoch 25 training complete
Cost on training data: 0.14645192808497318
Accuracy on training data: 990 / 1000
Cost on evaluation data: 1.2030957487435563
Accuracy on evaluation data: 8133 / 10000
Epoch 26 training complete
Cost on training data: 0.14361544306842858
Accuracy on training data: 991 / 1000
Cost on evaluation data: 1.2132069735876854
Accuracy on evaluation data: 8120 / 10000
Epoch 27 training complete
Cost on training data: 0.13293324613026067
Accuracy on training data: 994 / 1000
Cost on evaluation data: 1.2110816888558171
Accuracy on evaluation data: 8136 / 10000
Epoch 28 training complete
Cost on training data: 0.12809597159620803
Accuracy on training data: 994 / 1000
Cost on evaluation data: 1.221906351326292
Accuracy on evaluation data: 8112 / 10000
Epoch 29 training complete
Cost on training data: 0.12121702604589271
Accuracy on training data: 996 / 1000
Cost on evaluation data: 1.217029749855143
Accuracy on evaluation data: 8143 / 10000
Epoch 30 training complete
Cost on training data: 0.12155180353404105
Accuracy on training data: 994 / 1000
Cost on evaluation data: 1.257542166840531
Accuracy on evaluation data: 8132 / 10000
Epoch 31 training complete
Cost on training data: 0.11233999927446671
Accuracy on training data: 996 / 1000
Cost on evaluation data: 1.2295626466574059
Accuracy on evaluation data: 8143 / 10000
Epoch 32 training complete
Cost on training data: 0.10746363413839789
Accuracy on training data: 996 / 1000
Cost on evaluation data: 1.2264917284804158
Accuracy on evaluation data: 8133 / 10000
Epoch 33 training complete
Cost on training data: 0.1048020832388671
Accuracy on training data: 998 / 1000
Cost on evaluation data: 1.2447689499079413
Accuracy on evaluation data: 8131 / 10000
Epoch 34 training complete
Cost on training data: 0.09944015208409258
Accuracy on training data: 999 / 1000
Cost on evaluation data: 1.2476905216865046
Accuracy on evaluation data: 8132 / 10000
Epoch 35 training complete
Cost on training data: 0.09715708214560319
Accuracy on training data: 997 / 1000
Cost on evaluation data: 1.2481767157761254
Accuracy on evaluation data: 8135 / 10000
Epoch 36 training complete
Cost on training data: 0.09226811429853458
Accuracy on training data: 998 / 1000
Cost on evaluation data: 1.2470100151134607
Accuracy on evaluation data: 8137 / 10000
Epoch 37 training complete
Cost on training data: 0.09030924500475625
Accuracy on training data: 999 / 1000
Cost on evaluation data: 1.2554195954781409
Accuracy on evaluation data: 8160 / 10000
Epoch 38 training complete
Cost on training data: 0.08682752591099029
Accuracy on training data: 999 / 1000
Cost on evaluation data: 1.2622992445083403
Accuracy on evaluation data: 8147 / 10000
Epoch 39 training complete
Cost on training data: 0.08384814938783021
Accuracy on training data: 999 / 1000
Cost on evaluation data: 1.2588102938211014
Accuracy on evaluation data: 8142 / 10000
Epoch 40 training complete
Cost on training data: 0.08157297192245669
Accuracy on training data: 999 / 1000
Cost on evaluation data: 1.2631299632261273
Accuracy on evaluation data: 8143 / 10000
Epoch 41 training complete
Cost on training data: 0.08020944311658823
Accuracy on training data: 998 / 1000
Cost on evaluation data: 1.263860917909261
Accuracy on evaluation data: 8159 / 10000
Epoch 42 training complete
Cost on training data: 0.077013592547164
Accuracy on training data: 998 / 1000
Cost on evaluation data: 1.2727564175269843
Accuracy on evaluation data: 8157 / 10000
Epoch 43 training complete
Cost on training data: 0.07501717577506369
Accuracy on training data: 998 / 1000
Cost on evaluation data: 1.2616965135181386
Accuracy on evaluation data: 8157 / 10000
Epoch 44 training complete
Cost on training data: 0.07256130905007778
Accuracy on training data: 999 / 1000
Cost on evaluation data: 1.2752269079869292
Accuracy on evaluation data: 8161 / 10000
Epoch 45 training complete
Cost on training data: 0.0704409498768075
Accuracy on training data: 999 / 1000
Cost on evaluation data: 1.2801470505634474
Accuracy on evaluation data: 8160 / 10000
Epoch 46 training complete
Cost on training data: 0.06863010836166256
Accuracy on training data: 999 / 1000
Cost on evaluation data: 1.2789606476891509
Accuracy on evaluation data: 8173 / 10000
Epoch 47 training complete
Cost on training data: 0.06648076082379294
Accuracy on training data: 999 / 1000
Cost on evaluation data: 1.279253509388061
Accuracy on evaluation data: 8168 / 10000
Epoch 48 training complete
Cost on training data: 0.06497266858128088
Accuracy on training data: 999 / 1000
Cost on evaluation data: 1.2754182080246854
Accuracy on evaluation data: 8173 / 10000
Epoch 49 training complete
Cost on training data: 0.06297874601918942
Accuracy on training data: 999 / 1000
Cost on evaluation data: 1.2831719062170928
Accuracy on evaluation data: 8180 / 10000
Epoch 50 training complete
Cost on training data: 0.06244284088384119
Accuracy on training data: 999 / 1000
Cost on evaluation data: 1.292793526990941
Accuracy on evaluation data: 8168 / 10000
Epoch 51 training complete
Cost on training data: 0.06062325729258808
Accuracy on training data: 998 / 1000
Cost on evaluation data: 1.28716872806879
Accuracy on evaluation data: 8189 / 10000
Epoch 52 training complete
Cost on training data: 0.05844892715523189
Accuracy on training data: 999 / 1000
Cost on evaluation data: 1.2905043673963854
Accuracy on evaluation data: 8176 / 10000
Epoch 53 training complete
Cost on training data: 0.057037591237637436
Accuracy on training data: 999 / 1000
Cost on evaluation data: 1.289332558328879
Accuracy on evaluation data: 8190 / 10000
Epoch 54 training complete
Cost on training data: 0.05586136017969852
Accuracy on training data: 999 / 1000
Cost on evaluation data: 1.289541071379457
Accuracy on evaluation data: 8187 / 10000
Epoch 55 training complete
Cost on training data: 0.05443515023471758
Accuracy on training data: 999 / 1000
Cost on evaluation data: 1.2990823487570156
Accuracy on evaluation data: 8194 / 10000
Epoch 56 training complete
Cost on training data: 0.05319308117057278
Accuracy on training data: 999 / 1000
Cost on evaluation data: 1.3000239962698485
Accuracy on evaluation data: 8194 / 10000
Epoch 57 training complete
Cost on training data: 0.05161039604476325
Accuracy on training data: 999 / 1000
Cost on evaluation data: 1.3043509672082498
Accuracy on evaluation data: 8198 / 10000
Epoch 58 training complete
Cost on training data: 0.05083871731572437
Accuracy on training data: 999 / 1000
Cost on evaluation data: 1.3022715294113845
Accuracy on evaluation data: 8192 / 10000
Epoch 59 training complete
Cost on training data: 0.04976307728372554
Accuracy on training data: 999 / 1000
Cost on evaluation data: 1.3027124826623515
Accuracy on evaluation data: 8204 / 10000
Epoch 60 training complete
Cost on training data: 0.04889624941336269
Accuracy on training data: 999 / 1000
Cost on evaluation data: 1.311817808052567
Accuracy on evaluation data: 8205 / 10000
Epoch 61 training complete
Cost on training data: 0.047355358467197774
Accuracy on training data: 999 / 1000
Cost on evaluation data: 1.3084747578048694
Accuracy on evaluation data: 8209 / 10000
Epoch 62 training complete
Cost on training data: 0.046690844247366935
Accuracy on training data: 999 / 1000
Cost on evaluation data: 1.3165304367735435
Accuracy on evaluation data: 8199 / 10000
Epoch 63 training complete
Cost on training data: 0.04549230589169085
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.3184831225330333
Accuracy on evaluation data: 8224 / 10000
Epoch 64 training complete
Cost on training data: 0.04439579507040581
Accuracy on training data: 999 / 1000
Cost on evaluation data: 1.3176685968090527
Accuracy on evaluation data: 8220 / 10000
Epoch 65 training complete
Cost on training data: 0.04380936264543519
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.3204777091429898
Accuracy on evaluation data: 8224 / 10000
Epoch 66 training complete
Cost on training data: 0.04295522249463422
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.322150656281547
Accuracy on evaluation data: 8227 / 10000
Epoch 67 training complete
Cost on training data: 0.04217965757386699
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.3312918605319168
Accuracy on evaluation data: 8202 / 10000
Epoch 68 training complete
Cost on training data: 0.04111553983491183
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.3268372062185354
Accuracy on evaluation data: 8218 / 10000
Epoch 69 training complete
Cost on training data: 0.04011753237381811
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.3293706774002778
Accuracy on evaluation data: 8221 / 10000
Epoch 70 training complete
Cost on training data: 0.03940713880724735
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.330451004552
Accuracy on evaluation data: 8222 / 10000
Epoch 71 training complete
Cost on training data: 0.03864022634880402
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.3244472119366424
Accuracy on evaluation data: 8227 / 10000
Epoch 72 training complete
Cost on training data: 0.03819322155100375
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.3356105636135454
Accuracy on evaluation data: 8226 / 10000
Epoch 73 training complete
Cost on training data: 0.03795210269276921
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.334786683448296
Accuracy on evaluation data: 8231 / 10000
Epoch 74 training complete
Cost on training data: 0.03665713709932965
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.3317450719054986
Accuracy on evaluation data: 8238 / 10000
Epoch 75 training complete
Cost on training data: 0.036224344977773305
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.3405366155623732
Accuracy on evaluation data: 8235 / 10000
Epoch 76 training complete
Cost on training data: 0.03594795902601789
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.3460334504537177
Accuracy on evaluation data: 8220 / 10000
Epoch 77 training complete
Cost on training data: 0.035029786923706556
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.3443514365483307
Accuracy on evaluation data: 8222 / 10000
Epoch 78 training complete
Cost on training data: 0.03422378002812577
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.3470272182018044
Accuracy on evaluation data: 8220 / 10000
Epoch 79 training complete
Cost on training data: 0.03377978719514206
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.3426779369036395
Accuracy on evaluation data: 8226 / 10000
Epoch 80 training complete
Cost on training data: 0.033138308715480395
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.3447548765895374
Accuracy on evaluation data: 8232 / 10000
Epoch 81 training complete
Cost on training data: 0.03268605845639425
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.3450526932437048
Accuracy on evaluation data: 8251 / 10000
Epoch 82 training complete
Cost on training data: 0.03209119586792485
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.3496497414944504
Accuracy on evaluation data: 8228 / 10000
Epoch 83 training complete
Cost on training data: 0.031779354602804455
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.3560790776204301
Accuracy on evaluation data: 8234 / 10000
Epoch 84 training complete
Cost on training data: 0.031158275092867366
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.3523924565970855
Accuracy on evaluation data: 8230 / 10000
Epoch 85 training complete
Cost on training data: 0.030822552964318604
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.3580212024739295
Accuracy on evaluation data: 8232 / 10000
Epoch 86 training complete
Cost on training data: 0.030286835132469672
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.3530244927734938
Accuracy on evaluation data: 8242 / 10000
Epoch 87 training complete
Cost on training data: 0.029829457325905237
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.3558996500427245
Accuracy on evaluation data: 8237 / 10000
Epoch 88 training complete
Cost on training data: 0.0294069597301358
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.3617463601866189
Accuracy on evaluation data: 8231 / 10000
Epoch 89 training complete
Cost on training data: 0.02905029395246835
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.3630566055654625
Accuracy on evaluation data: 8238 / 10000
Epoch 90 training complete
Cost on training data: 0.028582596150161302
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.3643957249849021
Accuracy on evaluation data: 8233 / 10000
Epoch 91 training complete
Cost on training data: 0.028324914762123725
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.3677404770063595
Accuracy on evaluation data: 8248 / 10000
Epoch 92 training complete
Cost on training data: 0.02781108236483961
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.3717879363863614
Accuracy on evaluation data: 8229 / 10000
Epoch 93 training complete
Cost on training data: 0.027439935185696013
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.3733372913909754
Accuracy on evaluation data: 8240 / 10000
Epoch 94 training complete
Cost on training data: 0.027090856085152184
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.3693605126806525
Accuracy on evaluation data: 8246 / 10000
Epoch 95 training complete
Cost on training data: 0.026796102134257774
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.3765848871958497
Accuracy on evaluation data: 8239 / 10000
Epoch 96 training complete
Cost on training data: 0.02634836198227556
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.375382197351273
Accuracy on evaluation data: 8241 / 10000
Epoch 97 training complete
Cost on training data: 0.026166660210878483
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.3764636309870855
Accuracy on evaluation data: 8247 / 10000
Epoch 98 training complete
Cost on training data: 0.02570743482697377
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.3795183694956978
Accuracy on evaluation data: 8241 / 10000
Epoch 99 training complete
Cost on training data: 0.02546962734965154
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.3802676621668803
Accuracy on evaluation data: 8245 / 10000
Epoch 100 training complete
Cost on training data: 0.025135070352877868
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.3831611835252005
Accuracy on evaluation data: 8254 / 10000
Epoch 101 training complete
Cost on training data: 0.024857003022192948
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.387021430560139
Accuracy on evaluation data: 8255 / 10000
Epoch 102 training complete
Cost on training data: 0.024415660700624307
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.3869078184979873
Accuracy on evaluation data: 8246 / 10000
Epoch 103 training complete
Cost on training data: 0.024189005338626866
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.3861328385240215
Accuracy on evaluation data: 8260 / 10000
Epoch 104 training complete
Cost on training data: 0.023930930374165098
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.3892384906478257
Accuracy on evaluation data: 8254 / 10000
Epoch 105 training complete
Cost on training data: 0.02361683607315844
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.3905883067110325
Accuracy on evaluation data: 8254 / 10000
Epoch 106 training complete
Cost on training data: 0.023304495566227972
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.3926414354817855
Accuracy on evaluation data: 8255 / 10000
Epoch 107 training complete
Cost on training data: 0.023042273978151624
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.3918417501112124
Accuracy on evaluation data: 8252 / 10000
Epoch 108 training complete
Cost on training data: 0.022777293795288162
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.3946393780268975
Accuracy on evaluation data: 8251 / 10000
Epoch 109 training complete
Cost on training data: 0.022534764059725383
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.397671645052809
Accuracy on evaluation data: 8255 / 10000
Epoch 110 training complete
Cost on training data: 0.02223438575040059
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.397158494603118
Accuracy on evaluation data: 8255 / 10000
Epoch 111 training complete
Cost on training data: 0.02208784166772305
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.3961993057243174
Accuracy on evaluation data: 8253 / 10000
Epoch 112 training complete
Cost on training data: 0.021762273396763464
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.402263097584314
Accuracy on evaluation data: 8256 / 10000
Epoch 113 training complete
Cost on training data: 0.02148852198993692
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.4025027526609668
Accuracy on evaluation data: 8262 / 10000
Epoch 114 training complete
Cost on training data: 0.02130258436273831
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.4062743842167909
Accuracy on evaluation data: 8257 / 10000
Epoch 115 training complete
Cost on training data: 0.02102236595522121
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.4080591973252832
Accuracy on evaluation data: 8264 / 10000
Epoch 116 training complete
Cost on training data: 0.020844312713047174
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.410027990877223
Accuracy on evaluation data: 8255 / 10000
Epoch 117 training complete
Cost on training data: 0.02058027372039338
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.4101225365206775
Accuracy on evaluation data: 8260 / 10000
Epoch 118 training complete
Cost on training data: 0.020341870116987782
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.410622305822487
Accuracy on evaluation data: 8258 / 10000
Epoch 119 training complete
Cost on training data: 0.020193724092252695
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.4145837512197723
Accuracy on evaluation data: 8256 / 10000
Epoch 120 training complete
Cost on training data: 0.01992633186260965
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.4134389838644985
Accuracy on evaluation data: 8257 / 10000
Epoch 121 training complete
Cost on training data: 0.019757311487731036
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.4129552080415986
Accuracy on evaluation data: 8257 / 10000
Epoch 122 training complete
Cost on training data: 0.01954288456104957
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.4177607576725875
Accuracy on evaluation data: 8255 / 10000
Epoch 123 training complete
Cost on training data: 0.019333246805970948
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.419387030296146
Accuracy on evaluation data: 8265 / 10000
Epoch 124 training complete
Cost on training data: 0.01914627828533984
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.4233036082510475
Accuracy on evaluation data: 8262 / 10000
Epoch 125 training complete
Cost on training data: 0.018970773036744916
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.4207850308481844
Accuracy on evaluation data: 8265 / 10000
Epoch 126 training complete
Cost on training data: 0.018783260619228612
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.4229071574897914
Accuracy on evaluation data: 8258 / 10000
Epoch 127 training complete
Cost on training data: 0.01861739769238104
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.425685222963037
Accuracy on evaluation data: 8264 / 10000
Epoch 128 training complete
Cost on training data: 0.018435287515398965
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.4283066511203715
Accuracy on evaluation data: 8263 / 10000
Epoch 129 training complete
Cost on training data: 0.018286089959563163
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.4293618030946014
Accuracy on evaluation data: 8265 / 10000
Epoch 130 training complete
Cost on training data: 0.01812446522708623
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.4310825139939893
Accuracy on evaluation data: 8267 / 10000
Epoch 131 training complete
Cost on training data: 0.017961369730303548
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.4282144296482868
Accuracy on evaluation data: 8265 / 10000
Epoch 132 training complete
Cost on training data: 0.017760121510890896
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.4322339476144366
Accuracy on evaluation data: 8266 / 10000
Epoch 133 training complete
Cost on training data: 0.017646533059025166
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.434450880282216
Accuracy on evaluation data: 8267 / 10000
Epoch 134 training complete
Cost on training data: 0.017475043716433714
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.434732922635616
Accuracy on evaluation data: 8268 / 10000
Epoch 135 training complete
Cost on training data: 0.01734564633373677
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.437643377410993
Accuracy on evaluation data: 8270 / 10000
Epoch 136 training complete
Cost on training data: 0.017147511489556407
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.4346645687058328
Accuracy on evaluation data: 8268 / 10000
Epoch 137 training complete
Cost on training data: 0.017009193152674285
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.435957488303919
Accuracy on evaluation data: 8263 / 10000
Epoch 138 training complete
Cost on training data: 0.016858186884646947
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.4380956950907162
Accuracy on evaluation data: 8265 / 10000
Epoch 139 training complete
Cost on training data: 0.016720646928311267
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.4410238942110818
Accuracy on evaluation data: 8268 / 10000
Epoch 140 training complete
Cost on training data: 0.016598430655668503
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.4449023958522522
Accuracy on evaluation data: 8270 / 10000
Epoch 141 training complete
Cost on training data: 0.016428012815438174
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.4440189903773473
Accuracy on evaluation data: 8270 / 10000
Epoch 142 training complete
Cost on training data: 0.016298803110622186
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.4423420615804867
Accuracy on evaluation data: 8266 / 10000
Epoch 143 training complete
Cost on training data: 0.01617218740738815
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.4439811983458035
Accuracy on evaluation data: 8271 / 10000
Epoch 144 training complete
Cost on training data: 0.016043214646873288
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.446720940768649
Accuracy on evaluation data: 8269 / 10000
Epoch 145 training complete
Cost on training data: 0.015902712760033973
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.4478508201337381
Accuracy on evaluation data: 8266 / 10000
Epoch 146 training complete
Cost on training data: 0.015785338295322714
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.4473790496382208
Accuracy on evaluation data: 8266 / 10000
Epoch 147 training complete
Cost on training data: 0.015655610960450456
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.450634094099597
Accuracy on evaluation data: 8264 / 10000
Epoch 148 training complete
Cost on training data: 0.015538388797205305
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.45006822529315
Accuracy on evaluation data: 8274 / 10000
Epoch 149 training complete
Cost on training data: 0.015410941936529143
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.4524579486639575
Accuracy on evaluation data: 8273 / 10000
Epoch 150 training complete
Cost on training data: 0.0152857747012854
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.4539467327389253
Accuracy on evaluation data: 8271 / 10000
Epoch 151 training complete
Cost on training data: 0.015171456001645385
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.4562911591979164
Accuracy on evaluation data: 8281 / 10000
Epoch 152 training complete
Cost on training data: 0.015059041724475097
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.4578653681468134
Accuracy on evaluation data: 8271 / 10000
Epoch 153 training complete
Cost on training data: 0.014930753674301758
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.4585771793423556
Accuracy on evaluation data: 8271 / 10000
Epoch 154 training complete
Cost on training data: 0.014823944184983852
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.459296871379549
Accuracy on evaluation data: 8267 / 10000
Epoch 155 training complete
Cost on training data: 0.014711387373920718
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.4611042398343053
Accuracy on evaluation data: 8274 / 10000
Epoch 156 training complete
Cost on training data: 0.01460827251896262
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.4638433915171474
Accuracy on evaluation data: 8276 / 10000
Epoch 157 training complete
Cost on training data: 0.014493847710064806
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.4638396054779341
Accuracy on evaluation data: 8275 / 10000
Epoch 158 training complete
Cost on training data: 0.014409518604683896
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.463682311355057
Accuracy on evaluation data: 8271 / 10000
Epoch 159 training complete
Cost on training data: 0.014285795296330341
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.4660814996201126
Accuracy on evaluation data: 8282 / 10000
Epoch 160 training complete
Cost on training data: 0.01417753956565658
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.4661121797970065
Accuracy on evaluation data: 8271 / 10000
Epoch 161 training complete
Cost on training data: 0.014073144730616381
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.4691084073977518
Accuracy on evaluation data: 8280 / 10000
Epoch 162 training complete
Cost on training data: 0.013974520787591172
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.4678470727178272
Accuracy on evaluation data: 8281 / 10000
Epoch 163 training complete
Cost on training data: 0.01387479292599548
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.4689206912517845
Accuracy on evaluation data: 8274 / 10000
Epoch 164 training complete
Cost on training data: 0.013804255227650433
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.4701044323366772
Accuracy on evaluation data: 8270 / 10000
Epoch 165 training complete
Cost on training data: 0.01367398135085149
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.4723420386662769
Accuracy on evaluation data: 8274 / 10000
Epoch 166 training complete
Cost on training data: 0.013586645142879546
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.474399565865029
Accuracy on evaluation data: 8269 / 10000
Epoch 167 training complete
Cost on training data: 0.013482023497314505
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.4746436031466308
Accuracy on evaluation data: 8274 / 10000
Epoch 168 training complete
Cost on training data: 0.013388518698839932
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.4762899794337074
Accuracy on evaluation data: 8268 / 10000
Epoch 169 training complete
Cost on training data: 0.013296379898317731
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.476170073641381
Accuracy on evaluation data: 8274 / 10000
Epoch 170 training complete
Cost on training data: 0.013206612541539937
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.4762451289483594
Accuracy on evaluation data: 8283 / 10000
Epoch 171 training complete
Cost on training data: 0.01311791812918744
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.4814631982788284
Accuracy on evaluation data: 8279 / 10000
Epoch 172 training complete
Cost on training data: 0.013026801812712813
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.480090087570288
Accuracy on evaluation data: 8281 / 10000
Epoch 173 training complete
Cost on training data: 0.012956105744968105
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.4819430698183
Accuracy on evaluation data: 8277 / 10000
Epoch 174 training complete
Cost on training data: 0.01284616635895308
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.4831807265130927
Accuracy on evaluation data: 8276 / 10000
Epoch 175 training complete
Cost on training data: 0.012768774884022768
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.4846064456092476
Accuracy on evaluation data: 8272 / 10000
Epoch 176 training complete
Cost on training data: 0.01267280560620696
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.4872779405228755
Accuracy on evaluation data: 8272 / 10000
Epoch 177 training complete
Cost on training data: 0.0125942958592458
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.487256124755324
Accuracy on evaluation data: 8277 / 10000
Epoch 178 training complete
Cost on training data: 0.012509456326562496
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.488762669336739
Accuracy on evaluation data: 8277 / 10000
Epoch 179 training complete
Cost on training data: 0.012426678021030951
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.4893045636363336
Accuracy on evaluation data: 8274 / 10000
Epoch 180 training complete
Cost on training data: 0.012350459166636844
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.4908376388339077
Accuracy on evaluation data: 8271 / 10000
Epoch 181 training complete
Cost on training data: 0.012272095528212327
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.490305823188322
Accuracy on evaluation data: 8273 / 10000
Epoch 182 training complete
Cost on training data: 0.012195359196871949
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.4915766684742984
Accuracy on evaluation data: 8274 / 10000
Epoch 183 training complete
Cost on training data: 0.01211008938134288
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.4934363638091046
Accuracy on evaluation data: 8276 / 10000
Epoch 184 training complete
Cost on training data: 0.012034043691115366
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.4958805180457355
Accuracy on evaluation data: 8275 / 10000
Epoch 185 training complete
Cost on training data: 0.011959648231842602
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.4953044018341621
Accuracy on evaluation data: 8275 / 10000
Epoch 186 training complete
Cost on training data: 0.01188271539501657
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.4982862905659715
Accuracy on evaluation data: 8278 / 10000
Epoch 187 training complete
Cost on training data: 0.011805570974375115
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.4984599963716536
Accuracy on evaluation data: 8271 / 10000
Epoch 188 training complete
Cost on training data: 0.011739130248085065
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.4999222456417727
Accuracy on evaluation data: 8276 / 10000
Epoch 189 training complete
Cost on training data: 0.011662019634965269
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.5022406106655926
Accuracy on evaluation data: 8273 / 10000
Epoch 190 training complete
Cost on training data: 0.011588585105491449
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.501004429202565
Accuracy on evaluation data: 8275 / 10000
Epoch 191 training complete
Cost on training data: 0.011513318029394645
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.5027807152449528
Accuracy on evaluation data: 8271 / 10000
Epoch 192 training complete
Cost on training data: 0.01144913787556783
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.5042102278247018
Accuracy on evaluation data: 8277 / 10000
Epoch 193 training complete
Cost on training data: 0.011380722462723288
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.5043459284240497
Accuracy on evaluation data: 8276 / 10000
Epoch 194 training complete
Cost on training data: 0.011310841748858403
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.5057309393285034
Accuracy on evaluation data: 8274 / 10000
Epoch 195 training complete
Cost on training data: 0.011237910520928911
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.5098251083004055
Accuracy on evaluation data: 8277 / 10000
Epoch 196 training complete
Cost on training data: 0.011167825740054814
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.5083522521607542
Accuracy on evaluation data: 8273 / 10000
Epoch 197 training complete
Cost on training data: 0.011098893190490879
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.509689117678321
Accuracy on evaluation data: 8278 / 10000
Epoch 198 training complete
Cost on training data: 0.011033184264359744
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.5094626586832172
Accuracy on evaluation data: 8273 / 10000
Epoch 199 training complete
Cost on training data: 0.010969241073128723
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.5111954514857135
Accuracy on evaluation data: 8274 / 10000
Epoch 200 training complete
Cost on training data: 0.010901688250409856
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.5133778293044429
Accuracy on evaluation data: 8281 / 10000
Epoch 201 training complete
Cost on training data: 0.010836919854817734
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.512899045469996
Accuracy on evaluation data: 8276 / 10000
Epoch 202 training complete
Cost on training data: 0.01077363280325336
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.5141862424925208
Accuracy on evaluation data: 8283 / 10000
Epoch 203 training complete
Cost on training data: 0.010710238786932684
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.516021833075121
Accuracy on evaluation data: 8278 / 10000
Epoch 204 training complete
Cost on training data: 0.01064802523183221
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.5174300992779384
Accuracy on evaluation data: 8274 / 10000
Epoch 205 training complete
Cost on training data: 0.01058728437304496
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.5175545538577204
Accuracy on evaluation data: 8284 / 10000
Epoch 206 training complete
Cost on training data: 0.010524093452105442
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.5171366684972472
Accuracy on evaluation data: 8284 / 10000
Epoch 207 training complete
Cost on training data: 0.010464047409633297
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.5183459691008394
Accuracy on evaluation data: 8279 / 10000
Epoch 208 training complete
Cost on training data: 0.010405035820926401
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.5199626708782477
Accuracy on evaluation data: 8289 / 10000
Epoch 209 training complete
Cost on training data: 0.010346821567523935
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.5201538511505353
Accuracy on evaluation data: 8281 / 10000
Epoch 210 training complete
Cost on training data: 0.010286450474074083
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.5209364473701017
Accuracy on evaluation data: 8280 / 10000
Epoch 211 training complete
Cost on training data: 0.010229782911356983
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.5221262852400896
Accuracy on evaluation data: 8283 / 10000
Epoch 212 training complete
Cost on training data: 0.010174373980374391
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.5232705932012571
Accuracy on evaluation data: 8282 / 10000
Epoch 213 training complete
Cost on training data: 0.010117792555784311
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.5239829918497494
Accuracy on evaluation data: 8285 / 10000
Epoch 214 training complete
Cost on training data: 0.01006160721237948
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.5255351952080605
Accuracy on evaluation data: 8288 / 10000
Epoch 215 training complete
Cost on training data: 0.010005994402084191
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.5264244103313802
Accuracy on evaluation data: 8285 / 10000
Epoch 216 training complete
Cost on training data: 0.009945515986459671
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.5269862218334829
Accuracy on evaluation data: 8288 / 10000
Epoch 217 training complete
Cost on training data: 0.009893803245005467
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.526672626053213
Accuracy on evaluation data: 8290 / 10000
Epoch 218 training complete
Cost on training data: 0.009839790219893514
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.5276038641202723
Accuracy on evaluation data: 8288 / 10000
Epoch 219 training complete
Cost on training data: 0.009783638865064177
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.529722582070876
Accuracy on evaluation data: 8289 / 10000
Epoch 220 training complete
Cost on training data: 0.009729608663313141
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.5307905012686494
Accuracy on evaluation data: 8282 / 10000
Epoch 221 training complete
Cost on training data: 0.009676012006254286
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.5315590400491936
Accuracy on evaluation data: 8284 / 10000
Epoch 222 training complete
Cost on training data: 0.00962409709422437
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.5322365418043955
Accuracy on evaluation data: 8284 / 10000
Epoch 223 training complete
Cost on training data: 0.009572481113296616
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.5337122190780228
Accuracy on evaluation data: 8285 / 10000
Epoch 224 training complete
Cost on training data: 0.009520489595211335
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.5351252517660996
Accuracy on evaluation data: 8288 / 10000
Epoch 225 training complete
Cost on training data: 0.00946919852042269
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.5349021748364333
Accuracy on evaluation data: 8290 / 10000
Epoch 226 training complete
Cost on training data: 0.009419269761729595
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.5361619412199559
Accuracy on evaluation data: 8290 / 10000
Epoch 227 training complete
Cost on training data: 0.00937021194318918
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.5364592230633645
Accuracy on evaluation data: 8283 / 10000
Epoch 228 training complete
Cost on training data: 0.009321031078397397
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.5372351166971725
Accuracy on evaluation data: 8288 / 10000
Epoch 229 training complete
Cost on training data: 0.009272947056972252
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.5369664558159832
Accuracy on evaluation data: 8286 / 10000
Epoch 230 training complete
Cost on training data: 0.009224948364530684
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.5388243478923795
Accuracy on evaluation data: 8288 / 10000
Epoch 231 training complete
Cost on training data: 0.00917793292602622
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.5396987958085369
Accuracy on evaluation data: 8287 / 10000
Epoch 232 training complete
Cost on training data: 0.009129481267253584
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.5398790048659416
Accuracy on evaluation data: 8287 / 10000
Epoch 233 training complete
Cost on training data: 0.0090851786527295
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.5403340732822746
Accuracy on evaluation data: 8292 / 10000
Epoch 234 training complete
Cost on training data: 0.009043266960554519
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.5430000141831488
Accuracy on evaluation data: 8291 / 10000
Epoch 235 training complete
Cost on training data: 0.008995195976958914
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.5420335420285523
Accuracy on evaluation data: 8291 / 10000
Epoch 236 training complete
Cost on training data: 0.008950670950450762
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.5423219876631749
Accuracy on evaluation data: 8289 / 10000
Epoch 237 training complete
Cost on training data: 0.008905987780674442
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.5443883818726836
Accuracy on evaluation data: 8292 / 10000
Epoch 238 training complete
Cost on training data: 0.00886147955832109
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.5450416932433921
Accuracy on evaluation data: 8294 / 10000
Epoch 239 training complete
Cost on training data: 0.008819365937289459
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.5460247044115711
Accuracy on evaluation data: 8285 / 10000
Epoch 240 training complete
Cost on training data: 0.008777622814209884
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.5470096421353945
Accuracy on evaluation data: 8286 / 10000
Epoch 241 training complete
Cost on training data: 0.008733713683806727
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.5474543288446543
Accuracy on evaluation data: 8284 / 10000
Epoch 242 training complete
Cost on training data: 0.008692693217847578
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.5477454350223538
Accuracy on evaluation data: 8288 / 10000
Epoch 243 training complete
Cost on training data: 0.00864955033988838
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.5484336171997362
Accuracy on evaluation data: 8290 / 10000
Epoch 244 training complete
Cost on training data: 0.008608598619274009
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.549349833676881
Accuracy on evaluation data: 8289 / 10000
Epoch 245 training complete
Cost on training data: 0.008570220026530765
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.5493258744155092
Accuracy on evaluation data: 8293 / 10000
Epoch 246 training complete
Cost on training data: 0.008528773791209502
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.5506108702814045
Accuracy on evaluation data: 8289 / 10000
Epoch 247 training complete
Cost on training data: 0.008488499093182864
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.550905591527031
Accuracy on evaluation data: 8292 / 10000
Epoch 248 training complete
Cost on training data: 0.0084502661222921
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.5522673119612387
Accuracy on evaluation data: 8291 / 10000
Epoch 249 training complete
Cost on training data: 0.008412317911493328
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.5529453503210753
Accuracy on evaluation data: 8292 / 10000
Epoch 250 training complete
Cost on training data: 0.008372352432644605
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.5535656715063015
Accuracy on evaluation data: 8287 / 10000
Epoch 251 training complete
Cost on training data: 0.008332599091548133
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.5551894582490937
Accuracy on evaluation data: 8286 / 10000
Epoch 252 training complete
Cost on training data: 0.008294105919023428
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.5554505735309163
Accuracy on evaluation data: 8292 / 10000
Epoch 253 training complete
Cost on training data: 0.008259215320506665
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.556161364977507
Accuracy on evaluation data: 8292 / 10000
Epoch 254 training complete
Cost on training data: 0.008220615694691165
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.5575670688131764
Accuracy on evaluation data: 8291 / 10000
Epoch 255 training complete
Cost on training data: 0.008183659456734401
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.5566841456215208
Accuracy on evaluation data: 8293 / 10000
Epoch 256 training complete
Cost on training data: 0.008145243765449391
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.5580784947499111
Accuracy on evaluation data: 8293 / 10000
Epoch 257 training complete
Cost on training data: 0.00811132497633309
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.5591491505288406
Accuracy on evaluation data: 8290 / 10000
Epoch 258 training complete
Cost on training data: 0.008072867444806033
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.559205777860412
Accuracy on evaluation data: 8293 / 10000
Epoch 259 training complete
Cost on training data: 0.008039137471661302
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.5595608495494915
Accuracy on evaluation data: 8295 / 10000
Epoch 260 training complete
Cost on training data: 0.008002776941955229
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.5599915645171585
Accuracy on evaluation data: 8295 / 10000
Epoch 261 training complete
Cost on training data: 0.007966135462807927
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.5612742191533502
Accuracy on evaluation data: 8293 / 10000
Epoch 262 training complete
Cost on training data: 0.007932927803246505
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.5617044661896524
Accuracy on evaluation data: 8290 / 10000
Epoch 263 training complete
Cost on training data: 0.007896814017856647
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.5639061614090142
Accuracy on evaluation data: 8295 / 10000
Epoch 264 training complete
Cost on training data: 0.007862314047186325
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.5638887080090829
Accuracy on evaluation data: 8300 / 10000
Epoch 265 training complete
Cost on training data: 0.007829161245286636
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.564039570976322
Accuracy on evaluation data: 8298 / 10000
Epoch 266 training complete
Cost on training data: 0.007794196803171016
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.5650650519250455
Accuracy on evaluation data: 8294 / 10000
Epoch 267 training complete
Cost on training data: 0.007761089658909686
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.5661425998548535
Accuracy on evaluation data: 8297 / 10000
Epoch 268 training complete
Cost on training data: 0.007728443999357951
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.565807574673517
Accuracy on evaluation data: 8296 / 10000
Epoch 269 training complete
Cost on training data: 0.007694339065644186
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.5668031493773735
Accuracy on evaluation data: 8294 / 10000
Epoch 270 training complete
Cost on training data: 0.007662401185503222
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.568106770122055
Accuracy on evaluation data: 8300 / 10000
Epoch 271 training complete
Cost on training data: 0.007628572916914694
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.5687437285597305
Accuracy on evaluation data: 8296 / 10000
Epoch 272 training complete
Cost on training data: 0.007596682083618303
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.569268873401768
Accuracy on evaluation data: 8298 / 10000
Epoch 273 training complete
Cost on training data: 0.007564724510252255
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.569618285417743
Accuracy on evaluation data: 8298 / 10000
Epoch 274 training complete
Cost on training data: 0.00753279685228016
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.57063427330117
Accuracy on evaluation data: 8297 / 10000
Epoch 275 training complete
Cost on training data: 0.0075011932152492895
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.5711367692299203
Accuracy on evaluation data: 8297 / 10000
Epoch 276 training complete
Cost on training data: 0.007469686884988832
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.5721961590335034
Accuracy on evaluation data: 8298 / 10000
Epoch 277 training complete
Cost on training data: 0.0074397778934368166
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.573410871832064
Accuracy on evaluation data: 8299 / 10000
Epoch 278 training complete
Cost on training data: 0.007407899575713503
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.5734911482889113
Accuracy on evaluation data: 8299 / 10000
Epoch 279 training complete
Cost on training data: 0.007378360422680599
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.5744310738185394
Accuracy on evaluation data: 8295 / 10000
Epoch 280 training complete
Cost on training data: 0.00734723654093919
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.5743071528668362
Accuracy on evaluation data: 8300 / 10000
Epoch 281 training complete
Cost on training data: 0.0073182950033790755
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.5744745904157171
Accuracy on evaluation data: 8298 / 10000
Epoch 282 training complete
Cost on training data: 0.007289688300786332
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.5750136164633
Accuracy on evaluation data: 8295 / 10000
Epoch 283 training complete
Cost on training data: 0.007258697864985666
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.5766372111264622
Accuracy on evaluation data: 8294 / 10000
Epoch 284 training complete
Cost on training data: 0.007228247824557503
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.5779167560477239
Accuracy on evaluation data: 8295 / 10000
Epoch 285 training complete
Cost on training data: 0.0071993309724694525
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.5778660066099806
Accuracy on evaluation data: 8294 / 10000
Epoch 286 training complete
Cost on training data: 0.0071700828069124894
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.5788231418840584
Accuracy on evaluation data: 8294 / 10000
Epoch 287 training complete
Cost on training data: 0.007141382897501142
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.5797525834529964
Accuracy on evaluation data: 8294 / 10000
Epoch 288 training complete
Cost on training data: 0.007112755232208839
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.5794581545545405
Accuracy on evaluation data: 8296 / 10000
Epoch 289 training complete
Cost on training data: 0.007084432856325476
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.580591489351644
Accuracy on evaluation data: 8296 / 10000
Epoch 290 training complete
Cost on training data: 0.007055870763156264
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.5812627251939715
Accuracy on evaluation data: 8296 / 10000
Epoch 291 training complete
Cost on training data: 0.007028506324634178
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.5821939184770741
Accuracy on evaluation data: 8296 / 10000
Epoch 292 training complete
Cost on training data: 0.0070009498896063185
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.5825582951259767
Accuracy on evaluation data: 8297 / 10000
Epoch 293 training complete
Cost on training data: 0.006974487368698741
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.583824629303999
Accuracy on evaluation data: 8295 / 10000
Epoch 294 training complete
Cost on training data: 0.006946937966661494
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.583767970377277
Accuracy on evaluation data: 8291 / 10000
Epoch 295 training complete
Cost on training data: 0.006919016871455062
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.584793865062959
Accuracy on evaluation data: 8294 / 10000
Epoch 296 training complete
Cost on training data: 0.00689320902697137
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.5851442725875566
Accuracy on evaluation data: 8298 / 10000
Epoch 297 training complete
Cost on training data: 0.006865269809284962
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.5858903754774274
Accuracy on evaluation data: 8294 / 10000
Epoch 298 training complete
Cost on training data: 0.006838294745804023
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.5861635949572546
Accuracy on evaluation data: 8296 / 10000
Epoch 299 training complete
Cost on training data: 0.006811919263123583
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.5866647471415776
Accuracy on evaluation data: 8296 / 10000
Epoch 300 training complete
Cost on training data: 0.006785582799807724
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.5880810533631184
Accuracy on evaluation data: 8296 / 10000
Epoch 301 training complete
Cost on training data: 0.006760091004834835
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.5884613348067216
Accuracy on evaluation data: 8296 / 10000
Epoch 302 training complete
Cost on training data: 0.006734380375664931
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.589441767429376
Accuracy on evaluation data: 8299 / 10000
Epoch 303 training complete
Cost on training data: 0.006707966319262621
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.5889667838339951
Accuracy on evaluation data: 8296 / 10000
Epoch 304 training complete
Cost on training data: 0.006683337659338772
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.5899947445197147
Accuracy on evaluation data: 8295 / 10000
Epoch 305 training complete
Cost on training data: 0.006657937273495311
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.5904112229081537
Accuracy on evaluation data: 8294 / 10000
Epoch 306 training complete
Cost on training data: 0.006632656524273505
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.591234861717765
Accuracy on evaluation data: 8293 / 10000
Epoch 307 training complete
Cost on training data: 0.006607694239544852
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.5918131096636272
Accuracy on evaluation data: 8292 / 10000
Epoch 308 training complete
Cost on training data: 0.006582527007917741
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.591949681545194
Accuracy on evaluation data: 8293 / 10000
Epoch 309 training complete
Cost on training data: 0.006557784507463533
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.5931461376998346
Accuracy on evaluation data: 8296 / 10000
Epoch 310 training complete
Cost on training data: 0.006532846019595167
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.5934244099146166
Accuracy on evaluation data: 8293 / 10000
Epoch 311 training complete
Cost on training data: 0.0065083430951528905
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.594333469805743
Accuracy on evaluation data: 8293 / 10000
Epoch 312 training complete
Cost on training data: 0.006484709878571502
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.5952254737488258
Accuracy on evaluation data: 8293 / 10000
Epoch 313 training complete
Cost on training data: 0.006461170226183056
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.5964903807368758
Accuracy on evaluation data: 8296 / 10000
Epoch 314 training complete
Cost on training data: 0.006437303030948467
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.5964405487165547
Accuracy on evaluation data: 8294 / 10000
Epoch 315 training complete
Cost on training data: 0.0064127525277012315
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.596278985661997
Accuracy on evaluation data: 8295 / 10000
Epoch 316 training complete
Cost on training data: 0.006389409876967011
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.597688882057781
Accuracy on evaluation data: 8293 / 10000
Epoch 317 training complete
Cost on training data: 0.006366289655937858
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.5981273490843853
Accuracy on evaluation data: 8292 / 10000
Epoch 318 training complete
Cost on training data: 0.006342981082541249
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.5983906887950496
Accuracy on evaluation data: 8292 / 10000
Epoch 319 training complete
Cost on training data: 0.0063203398977905246
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.5986559702471281
Accuracy on evaluation data: 8295 / 10000
Epoch 320 training complete
Cost on training data: 0.0062962109518768135
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.5996711483418125
Accuracy on evaluation data: 8294 / 10000
Epoch 321 training complete
Cost on training data: 0.006273808047758421
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.6000424002265707
Accuracy on evaluation data: 8294 / 10000
Epoch 322 training complete
Cost on training data: 0.00625151095440419
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.6012701332494756
Accuracy on evaluation data: 8296 / 10000
Epoch 323 training complete
Cost on training data: 0.006228060550527366
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.6009994905314584
Accuracy on evaluation data: 8295 / 10000
Epoch 324 training complete
Cost on training data: 0.006206012839317615
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.6017933799357573
Accuracy on evaluation data: 8295 / 10000
Epoch 325 training complete
Cost on training data: 0.006183551822276642
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.6026720476181193
Accuracy on evaluation data: 8295 / 10000
Epoch 326 training complete
Cost on training data: 0.00616156253286002
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.6031862436345017
Accuracy on evaluation data: 8297 / 10000
Epoch 327 training complete
Cost on training data: 0.0061393561034788335
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.6034609817542438
Accuracy on evaluation data: 8297 / 10000
Epoch 328 training complete
Cost on training data: 0.006117597590134794
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.6047903939728643
Accuracy on evaluation data: 8298 / 10000
Epoch 329 training complete
Cost on training data: 0.006095757719829368
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.6049676902579775
Accuracy on evaluation data: 8298 / 10000
Epoch 330 training complete
Cost on training data: 0.006074009583956521
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.606296965410682
Accuracy on evaluation data: 8299 / 10000
Epoch 331 training complete
Cost on training data: 0.006052353491692505
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.6064514540051067
Accuracy on evaluation data: 8300 / 10000
Epoch 332 training complete
Cost on training data: 0.006031630415465684
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.606345865444866
Accuracy on evaluation data: 8298 / 10000
Epoch 333 training complete
Cost on training data: 0.0060105752904866605
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.6064939849144657
Accuracy on evaluation data: 8298 / 10000
Epoch 334 training complete
Cost on training data: 0.005989124166275899
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.607247295769548
Accuracy on evaluation data: 8296 / 10000
Epoch 335 training complete
Cost on training data: 0.005968453256550267
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.6072743048906684
Accuracy on evaluation data: 8295 / 10000
Epoch 336 training complete
Cost on training data: 0.005947808976751187
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.6077297907547556
Accuracy on evaluation data: 8293 / 10000
Epoch 337 training complete
Cost on training data: 0.005926509371158197
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.6093465285897326
Accuracy on evaluation data: 8297 / 10000
Epoch 338 training complete
Cost on training data: 0.005905956551024666
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.6098180526405625
Accuracy on evaluation data: 8296 / 10000
Epoch 339 training complete
Cost on training data: 0.005885509731717664
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.6107506260942306
Accuracy on evaluation data: 8296 / 10000
Epoch 340 training complete
Cost on training data: 0.0058650594071129785
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.6109526589784067
Accuracy on evaluation data: 8298 / 10000
Epoch 341 training complete
Cost on training data: 0.005844745282616136
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.6122580019575428
Accuracy on evaluation data: 8294 / 10000
Epoch 342 training complete
Cost on training data: 0.005824959573318467
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.6125485880277985
Accuracy on evaluation data: 8293 / 10000
Epoch 343 training complete
Cost on training data: 0.005804486235662657
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.613061937913507
Accuracy on evaluation data: 8294 / 10000
Epoch 344 training complete
Cost on training data: 0.005784820721646074
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.6131458751515746
Accuracy on evaluation data: 8295 / 10000
Epoch 345 training complete
Cost on training data: 0.005765186384621952
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.6142646007571384
Accuracy on evaluation data: 8295 / 10000
Epoch 346 training complete
Cost on training data: 0.005745120567645598
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.6145386574705065
Accuracy on evaluation data: 8294 / 10000
Epoch 347 training complete
Cost on training data: 0.005726436921016914
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.6149872376103747
Accuracy on evaluation data: 8295 / 10000
Epoch 348 training complete
Cost on training data: 0.005706246623579335
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.6154425306340765
Accuracy on evaluation data: 8296 / 10000
Epoch 349 training complete
Cost on training data: 0.005686681617061784
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.615973840493396
Accuracy on evaluation data: 8298 / 10000
Epoch 350 training complete
Cost on training data: 0.005667618354450181
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.6168644901395945
Accuracy on evaluation data: 8298 / 10000
Epoch 351 training complete
Cost on training data: 0.0056484299253665455
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.6175424416430357
Accuracy on evaluation data: 8298 / 10000
Epoch 352 training complete
Cost on training data: 0.005629276987089924
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.6176808317169489
Accuracy on evaluation data: 8298 / 10000
Epoch 353 training complete
Cost on training data: 0.0056105744976983314
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.6180581123087519
Accuracy on evaluation data: 8299 / 10000
Epoch 354 training complete
Cost on training data: 0.005592218262876864
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.6187874868947216
Accuracy on evaluation data: 8297 / 10000
Epoch 355 training complete
Cost on training data: 0.0055732862042998
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.6194588538992134
Accuracy on evaluation data: 8298 / 10000
Epoch 356 training complete
Cost on training data: 0.005555112903494204
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.619157698876115
Accuracy on evaluation data: 8299 / 10000
Epoch 357 training complete
Cost on training data: 0.005536896018830674
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.6202643900487748
Accuracy on evaluation data: 8296 / 10000
Epoch 358 training complete
Cost on training data: 0.005518116027962652
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.6206752003129121
Accuracy on evaluation data: 8299 / 10000
Epoch 359 training complete
Cost on training data: 0.005500670560409385
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.6207342571663153
Accuracy on evaluation data: 8300 / 10000
Epoch 360 training complete
Cost on training data: 0.005481677725089759
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.621618064414015
Accuracy on evaluation data: 8298 / 10000
Epoch 361 training complete
Cost on training data: 0.005464226171988452
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.622652578125993
Accuracy on evaluation data: 8296 / 10000
Epoch 362 training complete
Cost on training data: 0.005446400210087946
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.6228148596804397
Accuracy on evaluation data: 8297 / 10000
Epoch 363 training complete
Cost on training data: 0.005428454624956246
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.623025821697966
Accuracy on evaluation data: 8300 / 10000
Epoch 364 training complete
Cost on training data: 0.0054104286964459
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.6237028791233803
Accuracy on evaluation data: 8298 / 10000
Epoch 365 training complete
Cost on training data: 0.00539301430658295
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.6243806093788618
Accuracy on evaluation data: 8299 / 10000
Epoch 366 training complete
Cost on training data: 0.005375498262897481
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.6243901540279466
Accuracy on evaluation data: 8300 / 10000
Epoch 367 training complete
Cost on training data: 0.005358489797376338
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.624971238427544
Accuracy on evaluation data: 8299 / 10000
Epoch 368 training complete
Cost on training data: 0.005340784988125832
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.6254374122861526
Accuracy on evaluation data: 8298 / 10000
Epoch 369 training complete
Cost on training data: 0.005324152676756311
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.6254644712753177
Accuracy on evaluation data: 8302 / 10000
Epoch 370 training complete
Cost on training data: 0.005307064677220559
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.6262740501600876
Accuracy on evaluation data: 8301 / 10000
Epoch 371 training complete
Cost on training data: 0.00529048244829506
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.6263261924176353
Accuracy on evaluation data: 8301 / 10000
Epoch 372 training complete
Cost on training data: 0.005273317255472328
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.6270137416842052
Accuracy on evaluation data: 8299 / 10000
Epoch 373 training complete
Cost on training data: 0.005256813242155231
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.6273298132854352
Accuracy on evaluation data: 8300 / 10000
Epoch 374 training complete
Cost on training data: 0.005240131927593864
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.6277954063184625
Accuracy on evaluation data: 8303 / 10000
Epoch 375 training complete
Cost on training data: 0.005223825211361032
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.6286123134962385
Accuracy on evaluation data: 8299 / 10000
Epoch 376 training complete
Cost on training data: 0.005207332331341755
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.6288090022631614
Accuracy on evaluation data: 8303 / 10000
Epoch 377 training complete
Cost on training data: 0.00519125015797241
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.6291172069092357
Accuracy on evaluation data: 8303 / 10000
Epoch 378 training complete
Cost on training data: 0.005174970251628666
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.6303892442232097
Accuracy on evaluation data: 8300 / 10000
Epoch 379 training complete
Cost on training data: 0.005159061104723816
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.6307672740156367
Accuracy on evaluation data: 8301 / 10000
Epoch 380 training complete
Cost on training data: 0.005142849386959845
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.6309609266221168
Accuracy on evaluation data: 8302 / 10000
Epoch 381 training complete
Cost on training data: 0.00512710871117804
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.6312935288622605
Accuracy on evaluation data: 8302 / 10000
Epoch 382 training complete
Cost on training data: 0.005111354331814471
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.6315350820484409
Accuracy on evaluation data: 8303 / 10000
Epoch 383 training complete
Cost on training data: 0.005096038644866422
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.633040252535604
Accuracy on evaluation data: 8302 / 10000
Epoch 384 training complete
Cost on training data: 0.0050798822309203705
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.6331708446266922
Accuracy on evaluation data: 8302 / 10000
Epoch 385 training complete
Cost on training data: 0.005064595858651629
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.6339067912521021
Accuracy on evaluation data: 8302 / 10000
Epoch 386 training complete
Cost on training data: 0.00504906391920897
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.63412329890395
Accuracy on evaluation data: 8303 / 10000
Epoch 387 training complete
Cost on training data: 0.0050336951762420725
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.6347667245639317
Accuracy on evaluation data: 8303 / 10000
Epoch 388 training complete
Cost on training data: 0.005018508540358671
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.6352536246919185
Accuracy on evaluation data: 8305 / 10000
Epoch 389 training complete
Cost on training data: 0.005003339736179524
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.635440492252135
Accuracy on evaluation data: 8303 / 10000
Epoch 390 training complete
Cost on training data: 0.004988257563787323
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.6360078298787881
Accuracy on evaluation data: 8303 / 10000
Epoch 391 training complete
Cost on training data: 0.004973524991423581
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.637184475033575
Accuracy on evaluation data: 8302 / 10000
Epoch 392 training complete
Cost on training data: 0.004958559338911718
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.6375423803935851
Accuracy on evaluation data: 8303 / 10000
Epoch 393 training complete
Cost on training data: 0.004943845879734059
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.6368120798376302
Accuracy on evaluation data: 8305 / 10000
Epoch 394 training complete
Cost on training data: 0.0049289738244940534
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.6375102867468778
Accuracy on evaluation data: 8306 / 10000
Epoch 395 training complete
Cost on training data: 0.004914685866062012
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.6380359684293087
Accuracy on evaluation data: 8304 / 10000
Epoch 396 training complete
Cost on training data: 0.004899697730094281
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.6385443849673784
Accuracy on evaluation data: 8305 / 10000
Epoch 397 training complete
Cost on training data: 0.004885364227840051
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.639186180627196
Accuracy on evaluation data: 8304 / 10000
Epoch 398 training complete
Cost on training data: 0.00487109817644281
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.6393365941848124
Accuracy on evaluation data: 8305 / 10000
Epoch 399 training complete
Cost on training data: 0.004856754489618354
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.6398650020783767
Accuracy on evaluation data: 8305 / 10000

CPU times: user 14min 38s, sys: 11min 39s, total: 26min 17s
Wall time: 15min 50s

training_cost_xmin=200
test_accuracy_xmin=200
test_cost_xmin=0
training_accuracy_xmin=0
training_set_size=1000
lmbda=5.0

# local filename: train1000-lamdba5
make_plots('/content/train1000', 400,
               training_cost_xmin,
               test_accuracy_xmin,
               test_cost_xmin,
               training_accuracy_xmin,
               training_set_size)

training_cost_xmin=200
test_accuracy_xmin=200
test_cost_xmin=0
training_accuracy_xmin=0
training_set_size=1000
lmbda=0.0

# local filename: train1000
make_plots('/content/train1000-2', 400,
               training_cost_xmin,
               test_accuracy_xmin,
               test_cost_xmin,
               training_accuracy_xmin,
               training_set_size)

%%time
if __name__ == "__main__":
    filename = input("Enter a file name: ")
    num_epochs = int(input("Enter the number of epochs to run for: "))
    training_cost_xmin = int(input("training_cost_xmin (suggest 200): "))
    test_accuracy_xmin = int(input("test_accuracy_xmin (suggest 200): "))
    test_cost_xmin = int(input("test_cost_xmin (suggest 0): "))
    training_accuracy_xmin = int(input("training_accuracy_xmin (suggest 0): "))
    training_set_size = int(input("Training set size (suggest 1000): "))
    lmbda = float(input("Enter the regularization parameter, lambda (suggest: 5.0): "))
    main(filename, num_epochs, training_cost_xmin,
         test_accuracy_xmin, test_cost_xmin, training_accuracy_xmin,
         training_set_size, lmbda)

Enter a file name: train50000
Enter the number of epochs to run for: 30
training_cost_xmin (suggest 200): 15
test_accuracy_xmin (suggest 200): 15
test_cost_xmin (suggest 0): 0
training_accuracy_xmin (suggest 0): 0
Training set size (suggest 1000): 50000
Enter the regularization parameter, lambda (suggest: 5.0): 0
Epoch 0 training complete
Cost on training data: 0.565487411645113
Accuracy on training data: 45774 / 50000
Cost on evaluation data: 0.5623484668671077
Accuracy on evaluation data: 9147 / 10000
Epoch 1 training complete
Cost on training data: 0.4743487323018099
Accuracy on training data: 46415 / 50000
Cost on evaluation data: 0.49780077854652244
Accuracy on evaluation data: 9206 / 10000
Epoch 2 training complete
Cost on training data: 0.4241966751129283
Accuracy on training data: 46822 / 50000
Cost on evaluation data: 0.459207819060238
Accuracy on evaluation data: 9303 / 10000
Epoch 3 training complete
Cost on training data: 0.37634258350749356
Accuracy on training data: 47143 / 50000
Cost on evaluation data: 0.4322746488815946
Accuracy on evaluation data: 9343 / 10000
Epoch 4 training complete
Cost on training data: 0.33737051844768595
Accuracy on training data: 47527 / 50000
Cost on evaluation data: 0.4113851509361249
Accuracy on evaluation data: 9379 / 10000
Epoch 5 training complete
Cost on training data: 0.3382252087812596
Accuracy on training data: 47557 / 50000
Cost on evaluation data: 0.41546570994205817
Accuracy on evaluation data: 9370 / 10000
Epoch 6 training complete
Cost on training data: 0.2960054626692339
Accuracy on training data: 47814 / 50000
Cost on evaluation data: 0.38841472122981535
Accuracy on evaluation data: 9419 / 10000
Epoch 7 training complete
Cost on training data: 0.3110609033056994
Accuracy on training data: 47730 / 50000
Cost on evaluation data: 0.3992007929940586
Accuracy on evaluation data: 9416 / 10000
Epoch 8 training complete
Cost on training data: 0.2915602632430604
Accuracy on training data: 47859 / 50000
Cost on evaluation data: 0.39488486195152966
Accuracy on evaluation data: 9397 / 10000
Epoch 9 training complete
Cost on training data: 0.2667638206942151
Accuracy on training data: 48138 / 50000
Cost on evaluation data: 0.374039623897005
Accuracy on evaluation data: 9473 / 10000
Epoch 10 training complete
Cost on training data: 0.27078803048995786
Accuracy on training data: 48067 / 50000
Cost on evaluation data: 0.3984224935183234
Accuracy on evaluation data: 9423 / 10000
Epoch 11 training complete
Cost on training data: 0.26626139438990726
Accuracy on training data: 48122 / 50000
Cost on evaluation data: 0.38517197513213136
Accuracy on evaluation data: 9436 / 10000
Epoch 12 training complete
Cost on training data: 0.2541741680259756
Accuracy on training data: 48249 / 50000
Cost on evaluation data: 0.3762106123238996
Accuracy on evaluation data: 9448 / 10000
Epoch 13 training complete
Cost on training data: 0.24751161475558864
Accuracy on training data: 48319 / 50000
Cost on evaluation data: 0.3839525171381957
Accuracy on evaluation data: 9459 / 10000
Epoch 14 training complete
Cost on training data: 0.22923263801903093
Accuracy on training data: 48417 / 50000
Cost on evaluation data: 0.3669446685984708
Accuracy on evaluation data: 9462 / 10000
Epoch 15 training complete
Cost on training data: 0.22156525044004927
Accuracy on training data: 48494 / 50000
Cost on evaluation data: 0.3593594442957297
Accuracy on evaluation data: 9499 / 10000
Epoch 16 training complete
Cost on training data: 0.22133254144627743
Accuracy on training data: 48504 / 50000
Cost on evaluation data: 0.3586621737977863
Accuracy on evaluation data: 9525 / 10000
Epoch 17 training complete
Cost on training data: 0.20921983346732087
Accuracy on training data: 48601 / 50000
Cost on evaluation data: 0.3600117441081664
Accuracy on evaluation data: 9497 / 10000
Epoch 18 training complete
Cost on training data: 0.20592653169237327
Accuracy on training data: 48637 / 50000
Cost on evaluation data: 0.35395119496548
Accuracy on evaluation data: 9505 / 10000
Epoch 19 training complete
Cost on training data: 0.19824505059661693
Accuracy on training data: 48692 / 50000
Cost on evaluation data: 0.36864000308210604
Accuracy on evaluation data: 9506 / 10000
Epoch 20 training complete
Cost on training data: 0.2082025238766053
Accuracy on training data: 48621 / 50000
Cost on evaluation data: 0.3754736669857341
Accuracy on evaluation data: 9480 / 10000
Epoch 21 training complete
Cost on training data: 0.20323017315719372
Accuracy on training data: 48658 / 50000
Cost on evaluation data: 0.3768741163764822
Accuracy on evaluation data: 9493 / 10000
Epoch 22 training complete
Cost on training data: 0.19966263392082073
Accuracy on training data: 48629 / 50000
Cost on evaluation data: 0.37276643167074736
Accuracy on evaluation data: 9492 / 10000
Epoch 23 training complete
Cost on training data: 0.19885709346335595
Accuracy on training data: 48673 / 50000
Cost on evaluation data: 0.3808985433917027
Accuracy on evaluation data: 9474 / 10000
Epoch 24 training complete
Cost on training data: 0.18032189693172532
Accuracy on training data: 48840 / 50000
Cost on evaluation data: 0.3591711188637095
Accuracy on evaluation data: 9500 / 10000
Epoch 25 training complete
Cost on training data: 0.17471177293884144
Accuracy on training data: 48861 / 50000
Cost on evaluation data: 0.35192917918287636
Accuracy on evaluation data: 9516 / 10000
Epoch 26 training complete
Cost on training data: 0.18026505053099987
Accuracy on training data: 48844 / 50000
Cost on evaluation data: 0.36468757183451384
Accuracy on evaluation data: 9505 / 10000
Epoch 27 training complete
Cost on training data: 0.18464408045211
Accuracy on training data: 48783 / 50000
Cost on evaluation data: 0.37161590935708827
Accuracy on evaluation data: 9506 / 10000
Epoch 28 training complete
Cost on training data: 0.19594147314018678
Accuracy on training data: 48724 / 50000
Cost on evaluation data: 0.38136079101938825
Accuracy on evaluation data: 9514 / 10000
Epoch 29 training complete
Cost on training data: 0.17359688345261265
Accuracy on training data: 48915 / 50000
Cost on evaluation data: 0.3619564888355245
Accuracy on evaluation data: 9507 / 10000

CPU times: user 5min 53s, sys: 3min 57s, total: 9min 50s
Wall time: 7min 41s

training_cost_xmin=15
test_accuracy_xmin=15
test_cost_xmin=0
training_accuracy_xmin=0
training_set_size=50000
lmbda=0.0

make_plots('/content/train50000', 30,
               training_cost_xmin,
               test_accuracy_xmin,
               test_cost_xmin,
               training_accuracy_xmin,
               training_set_size)

# net = Network2mat([784, 30, 10], cost=CrossEntropyCost)
# net.large_weight_initializer()
# net.SGD(training_data[:1000], 400, 10, 0.5,
#         evaluation_data=test_data, lmbda = 0.1,
#         monitor_evaluation_cost=True, monitor_evaluation_accuracy=True,
#         monitor_training_cost=True, monitor_training_accuracy=True)

%%time
if __name__ == "__main__":
    filename = input("Enter a file name: ")
    num_epochs = int(input("Enter the number of epochs to run for: "))
    training_cost_xmin = int(input("training_cost_xmin (suggest 200): "))
    test_accuracy_xmin = int(input("test_accuracy_xmin (suggest 200): "))
    test_cost_xmin = int(input("test_cost_xmin (suggest 0): "))
    training_accuracy_xmin = int(input("training_accuracy_xmin (suggest 0): "))
    training_set_size = int(input("Training set size (suggest 1000): "))
    lmbda = float(input("Enter the regularization parameter, lambda (suggest: 5.0): "))
    main(filename, num_epochs, training_cost_xmin,
         test_accuracy_xmin, test_cost_xmin, training_accuracy_xmin,
         training_set_size, lmbda)

Enter a file name: train1000-lambda_point1
Enter the number of epochs to run for: 400
training_cost_xmin (suggest 200): 200
test_accuracy_xmin (suggest 200): 200
test_cost_xmin (suggest 0): 0
training_accuracy_xmin (suggest 0): 0
Training set size (suggest 1000): 1000
Enter the regularization parameter, lambda (suggest: 5.0): 0.1
Epoch 0 training complete
Cost on training data: 3.070360696676798
Accuracy on training data: 654 / 1000
Cost on evaluation data: 2.267426805451026
Accuracy on evaluation data: 5689 / 10000
Epoch 1 training complete
Cost on training data: 2.5499498812714574
Accuracy on training data: 779 / 1000
Cost on evaluation data: 1.8874601622712677
Accuracy on evaluation data: 6627 / 10000
Epoch 2 training complete
Cost on training data: 2.3300385720251486
Accuracy on training data: 801 / 1000
Cost on evaluation data: 1.6980853395461923
Accuracy on evaluation data: 6951 / 10000
Epoch 3 training complete
Cost on training data: 2.09308027964456
Accuracy on training data: 863 / 1000
Cost on evaluation data: 1.558920509036749
Accuracy on evaluation data: 7414 / 10000
Epoch 4 training complete
Cost on training data: 1.959190834576042
Accuracy on training data: 890 / 1000
Cost on evaluation data: 1.4806930558341391
Accuracy on evaluation data: 7560 / 10000
Epoch 5 training complete
Cost on training data: 1.8701389458524686
Accuracy on training data: 911 / 1000
Cost on evaluation data: 1.4505745942997819
Accuracy on evaluation data: 7622 / 10000
Epoch 6 training complete
Cost on training data: 1.7506274221524538
Accuracy on training data: 933 / 1000
Cost on evaluation data: 1.3703546508680589
Accuracy on evaluation data: 7787 / 10000
Epoch 7 training complete
Cost on training data: 1.6674245348944023
Accuracy on training data: 939 / 1000
Cost on evaluation data: 1.3360211887461975
Accuracy on evaluation data: 7858 / 10000
Epoch 8 training complete
Cost on training data: 1.6157130804251154
Accuracy on training data: 946 / 1000
Cost on evaluation data: 1.3219846238241026
Accuracy on evaluation data: 7880 / 10000
Epoch 9 training complete
Cost on training data: 1.554686268195161
Accuracy on training data: 955 / 1000
Cost on evaluation data: 1.315796372936809
Accuracy on evaluation data: 7935 / 10000
Epoch 10 training complete
Cost on training data: 1.4949204835817977
Accuracy on training data: 961 / 1000
Cost on evaluation data: 1.320230219689416
Accuracy on evaluation data: 7929 / 10000
Epoch 11 training complete
Cost on training data: 1.4818460096460242
Accuracy on training data: 964 / 1000
Cost on evaluation data: 1.30039531832619
Accuracy on evaluation data: 7944 / 10000
Epoch 12 training complete
Cost on training data: 1.4137442275209868
Accuracy on training data: 967 / 1000
Cost on evaluation data: 1.2875201012045665
Accuracy on evaluation data: 8000 / 10000
Epoch 13 training complete
Cost on training data: 1.3731826171058648
Accuracy on training data: 971 / 1000
Cost on evaluation data: 1.2676480508606796
Accuracy on evaluation data: 8046 / 10000
Epoch 14 training complete
Cost on training data: 1.3463558779205007
Accuracy on training data: 974 / 1000
Cost on evaluation data: 1.2705197922835934
Accuracy on evaluation data: 8055 / 10000
Epoch 15 training complete
Cost on training data: 1.3164924028939695
Accuracy on training data: 975 / 1000
Cost on evaluation data: 1.2665984376116841
Accuracy on evaluation data: 8117 / 10000
Epoch 16 training complete
Cost on training data: 1.2817951437471875
Accuracy on training data: 981 / 1000
Cost on evaluation data: 1.2378727008655925
Accuracy on evaluation data: 8121 / 10000
Epoch 17 training complete
Cost on training data: 1.2614812043125732
Accuracy on training data: 982 / 1000
Cost on evaluation data: 1.236990536073006
Accuracy on evaluation data: 8127 / 10000
Epoch 18 training complete
Cost on training data: 1.2317981130366493
Accuracy on training data: 981 / 1000
Cost on evaluation data: 1.2416701168090345
Accuracy on evaluation data: 8116 / 10000
Epoch 19 training complete
Cost on training data: 1.2092999770497324
Accuracy on training data: 985 / 1000
Cost on evaluation data: 1.2345992370599745
Accuracy on evaluation data: 8179 / 10000
Epoch 20 training complete
Cost on training data: 1.1891372287177537
Accuracy on training data: 985 / 1000
Cost on evaluation data: 1.231776145659345
Accuracy on evaluation data: 8185 / 10000
Epoch 21 training complete
Cost on training data: 1.1708545314330734
Accuracy on training data: 988 / 1000
Cost on evaluation data: 1.243009048023027
Accuracy on evaluation data: 8197 / 10000
Epoch 22 training complete
Cost on training data: 1.1532019095305681
Accuracy on training data: 987 / 1000
Cost on evaluation data: 1.2302276620841754
Accuracy on evaluation data: 8209 / 10000
Epoch 23 training complete
Cost on training data: 1.1362273197668147
Accuracy on training data: 990 / 1000
Cost on evaluation data: 1.219180684096704
Accuracy on evaluation data: 8201 / 10000
Epoch 24 training complete
Cost on training data: 1.1271831452904617
Accuracy on training data: 990 / 1000
Cost on evaluation data: 1.2373382475294543
Accuracy on evaluation data: 8171 / 10000
Epoch 25 training complete
Cost on training data: 1.1068411285820956
Accuracy on training data: 989 / 1000
Cost on evaluation data: 1.221990180563778
Accuracy on evaluation data: 8242 / 10000
Epoch 26 training complete
Cost on training data: 1.0954351860053821
Accuracy on training data: 991 / 1000
Cost on evaluation data: 1.2263361751773023
Accuracy on evaluation data: 8225 / 10000
Epoch 27 training complete
Cost on training data: 1.0770384742595405
Accuracy on training data: 994 / 1000
Cost on evaluation data: 1.221169926679863
Accuracy on evaluation data: 8248 / 10000
Epoch 28 training complete
Cost on training data: 1.0647476972021186
Accuracy on training data: 993 / 1000
Cost on evaluation data: 1.2268031620380757
Accuracy on evaluation data: 8233 / 10000
Epoch 29 training complete
Cost on training data: 1.0504298589808183
Accuracy on training data: 995 / 1000
Cost on evaluation data: 1.218352952579399
Accuracy on evaluation data: 8252 / 10000
Epoch 30 training complete
Cost on training data: 1.0431744378239685
Accuracy on training data: 994 / 1000
Cost on evaluation data: 1.249762387453506
Accuracy on evaluation data: 8235 / 10000
Epoch 31 training complete
Cost on training data: 1.0270402407258246
Accuracy on training data: 996 / 1000
Cost on evaluation data: 1.2164889227325593
Accuracy on evaluation data: 8271 / 10000
Epoch 32 training complete
Cost on training data: 1.0145262510213389
Accuracy on training data: 997 / 1000
Cost on evaluation data: 1.2067352650513907
Accuracy on evaluation data: 8284 / 10000
Epoch 33 training complete
Cost on training data: 1.005263678102775
Accuracy on training data: 997 / 1000
Cost on evaluation data: 1.2145690708533032
Accuracy on evaluation data: 8274 / 10000
Epoch 34 training complete
Cost on training data: 0.9924583626742032
Accuracy on training data: 999 / 1000
Cost on evaluation data: 1.2126795283986687
Accuracy on evaluation data: 8288 / 10000
Epoch 35 training complete
Cost on training data: 0.9830586317364072
Accuracy on training data: 997 / 1000
Cost on evaluation data: 1.2119285236493695
Accuracy on evaluation data: 8300 / 10000
Epoch 36 training complete
Cost on training data: 0.9705926102024827
Accuracy on training data: 998 / 1000
Cost on evaluation data: 1.2033980838931593
Accuracy on evaluation data: 8305 / 10000
Epoch 37 training complete
Cost on training data: 0.9614662788410542
Accuracy on training data: 998 / 1000
Cost on evaluation data: 1.209260429527872
Accuracy on evaluation data: 8303 / 10000
Epoch 38 training complete
Cost on training data: 0.9504859623649382
Accuracy on training data: 999 / 1000
Cost on evaluation data: 1.209794237046139
Accuracy on evaluation data: 8298 / 10000
Epoch 39 training complete
Cost on training data: 0.9409620715462244
Accuracy on training data: 998 / 1000
Cost on evaluation data: 1.2013984540125793
Accuracy on evaluation data: 8314 / 10000
Epoch 40 training complete
Cost on training data: 0.931409062408318
Accuracy on training data: 999 / 1000
Cost on evaluation data: 1.2041893743294771
Accuracy on evaluation data: 8292 / 10000
Epoch 41 training complete
Cost on training data: 0.9233605694063073
Accuracy on training data: 999 / 1000
Cost on evaluation data: 1.1957668550497065
Accuracy on evaluation data: 8316 / 10000
Epoch 42 training complete
Cost on training data: 0.9135406247829807
Accuracy on training data: 998 / 1000
Cost on evaluation data: 1.208849699811312
Accuracy on evaluation data: 8332 / 10000
Epoch 43 training complete
Cost on training data: 0.9045779503370333
Accuracy on training data: 999 / 1000
Cost on evaluation data: 1.1905092238989747
Accuracy on evaluation data: 8344 / 10000
Epoch 44 training complete
Cost on training data: 0.894746526106206
Accuracy on training data: 999 / 1000
Cost on evaluation data: 1.1952179540903338
Accuracy on evaluation data: 8339 / 10000
Epoch 45 training complete
Cost on training data: 0.8859516418619959
Accuracy on training data: 999 / 1000
Cost on evaluation data: 1.2001018188519033
Accuracy on evaluation data: 8329 / 10000
Epoch 46 training complete
Cost on training data: 0.8775834772799677
Accuracy on training data: 999 / 1000
Cost on evaluation data: 1.1956884045628653
Accuracy on evaluation data: 8362 / 10000
Epoch 47 training complete
Cost on training data: 0.868318997843876
Accuracy on training data: 999 / 1000
Cost on evaluation data: 1.1878009617275824
Accuracy on evaluation data: 8363 / 10000
Epoch 48 training complete
Cost on training data: 0.8602327271854222
Accuracy on training data: 999 / 1000
Cost on evaluation data: 1.1790846301736781
Accuracy on evaluation data: 8371 / 10000
Epoch 49 training complete
Cost on training data: 0.8518103930713838
Accuracy on training data: 999 / 1000
Cost on evaluation data: 1.1859390790245632
Accuracy on evaluation data: 8356 / 10000
Epoch 50 training complete
Cost on training data: 0.8452577691690797
Accuracy on training data: 999 / 1000
Cost on evaluation data: 1.189137429036529
Accuracy on evaluation data: 8354 / 10000
Epoch 51 training complete
Cost on training data: 0.8371444093461841
Accuracy on training data: 999 / 1000
Cost on evaluation data: 1.1798285194326754
Accuracy on evaluation data: 8374 / 10000
Epoch 52 training complete
Cost on training data: 0.8287714969068899
Accuracy on training data: 999 / 1000
Cost on evaluation data: 1.1756404086913743
Accuracy on evaluation data: 8374 / 10000
Epoch 53 training complete
Cost on training data: 0.8206680240068689
Accuracy on training data: 999 / 1000
Cost on evaluation data: 1.1743344319866673
Accuracy on evaluation data: 8380 / 10000
Epoch 54 training complete
Cost on training data: 0.8133380574119782
Accuracy on training data: 999 / 1000
Cost on evaluation data: 1.1679987805090895
Accuracy on evaluation data: 8395 / 10000
Epoch 55 training complete
Cost on training data: 0.805778782386357
Accuracy on training data: 999 / 1000
Cost on evaluation data: 1.172367941841726
Accuracy on evaluation data: 8408 / 10000
Epoch 56 training complete
Cost on training data: 0.7987269778265124
Accuracy on training data: 999 / 1000
Cost on evaluation data: 1.1761319882370984
Accuracy on evaluation data: 8401 / 10000
Epoch 57 training complete
Cost on training data: 0.7906693290400094
Accuracy on training data: 999 / 1000
Cost on evaluation data: 1.1724772566768158
Accuracy on evaluation data: 8400 / 10000
Epoch 58 training complete
Cost on training data: 0.7839454856281693
Accuracy on training data: 999 / 1000
Cost on evaluation data: 1.1674510374324722
Accuracy on evaluation data: 8410 / 10000
Epoch 59 training complete
Cost on training data: 0.7768349139088552
Accuracy on training data: 999 / 1000
Cost on evaluation data: 1.1602928677115827
Accuracy on evaluation data: 8414 / 10000
Epoch 60 training complete
Cost on training data: 0.7701948501121045
Accuracy on training data: 999 / 1000
Cost on evaluation data: 1.1649740554734405
Accuracy on evaluation data: 8401 / 10000
Epoch 61 training complete
Cost on training data: 0.76265131896608
Accuracy on training data: 999 / 1000
Cost on evaluation data: 1.1610097442259038
Accuracy on evaluation data: 8420 / 10000
Epoch 62 training complete
Cost on training data: 0.7563411017977935
Accuracy on training data: 999 / 1000
Cost on evaluation data: 1.1677569586756251
Accuracy on evaluation data: 8387 / 10000
Epoch 63 training complete
Cost on training data: 0.7487698963499093
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.1667465701026938
Accuracy on evaluation data: 8411 / 10000
Epoch 64 training complete
Cost on training data: 0.7420017001531081
Accuracy on training data: 999 / 1000
Cost on evaluation data: 1.1609317453576495
Accuracy on evaluation data: 8420 / 10000
Epoch 65 training complete
Cost on training data: 0.7357105122958817
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.155752894849745
Accuracy on evaluation data: 8425 / 10000
Epoch 66 training complete
Cost on training data: 0.7289733522730543
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.1577199669551295
Accuracy on evaluation data: 8432 / 10000
Epoch 67 training complete
Cost on training data: 0.7235684943760071
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.1688554113909086
Accuracy on evaluation data: 8398 / 10000
Epoch 68 training complete
Cost on training data: 0.7159325158524454
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.154304690658254
Accuracy on evaluation data: 8426 / 10000
Epoch 69 training complete
Cost on training data: 0.7093997734589172
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.1574750196566603
Accuracy on evaluation data: 8414 / 10000
Epoch 70 training complete
Cost on training data: 0.7035583367839017
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.1516106340542782
Accuracy on evaluation data: 8442 / 10000
Epoch 71 training complete
Cost on training data: 0.6968430288757788
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.142106898612304
Accuracy on evaluation data: 8435 / 10000
Epoch 72 training complete
Cost on training data: 0.6909832833275426
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.1529215728801063
Accuracy on evaluation data: 8421 / 10000
Epoch 73 training complete
Cost on training data: 0.6855772122537542
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.1457975205409108
Accuracy on evaluation data: 8434 / 10000
Epoch 74 training complete
Cost on training data: 0.6786027156334329
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.135481103621664
Accuracy on evaluation data: 8459 / 10000
Epoch 75 training complete
Cost on training data: 0.6727448600233423
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.1437997618199078
Accuracy on evaluation data: 8442 / 10000
Epoch 76 training complete
Cost on training data: 0.6672691866478249
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.1442576706424303
Accuracy on evaluation data: 8422 / 10000
Epoch 77 training complete
Cost on training data: 0.6616917968102124
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.1435320800257693
Accuracy on evaluation data: 8435 / 10000
Epoch 78 training complete
Cost on training data: 0.6552344417618767
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.145340054605354
Accuracy on evaluation data: 8445 / 10000
Epoch 79 training complete
Cost on training data: 0.6497423315673907
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.1320900400485996
Accuracy on evaluation data: 8443 / 10000
Epoch 80 training complete
Cost on training data: 0.6439332583387082
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.13272644353737
Accuracy on evaluation data: 8460 / 10000
Epoch 81 training complete
Cost on training data: 0.6383980997901139
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.1334349258652565
Accuracy on evaluation data: 8450 / 10000
Epoch 82 training complete
Cost on training data: 0.6328834371928292
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.1276393352544032
Accuracy on evaluation data: 8464 / 10000
Epoch 83 training complete
Cost on training data: 0.6275276812766616
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.1369029185673407
Accuracy on evaluation data: 8458 / 10000
Epoch 84 training complete
Cost on training data: 0.6221252884667142
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.1251081483350736
Accuracy on evaluation data: 8477 / 10000
Epoch 85 training complete
Cost on training data: 0.6167786748976244
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.129389445581057
Accuracy on evaluation data: 8462 / 10000
Epoch 86 training complete
Cost on training data: 0.6112779255542808
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.1196692186773112
Accuracy on evaluation data: 8463 / 10000
Epoch 87 training complete
Cost on training data: 0.6060701874474181
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.1166357542958638
Accuracy on evaluation data: 8471 / 10000
Epoch 88 training complete
Cost on training data: 0.6010666249184339
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.125102894622701
Accuracy on evaluation data: 8469 / 10000
Epoch 89 training complete
Cost on training data: 0.5959465659903667
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.1213750906869682
Accuracy on evaluation data: 8470 / 10000
Epoch 90 training complete
Cost on training data: 0.5906568544392168
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.1259289331964941
Accuracy on evaluation data: 8465 / 10000
Epoch 91 training complete
Cost on training data: 0.5861264127955454
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.1184106580461106
Accuracy on evaluation data: 8477 / 10000
Epoch 92 training complete
Cost on training data: 0.580671935411445
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.1184002783363547
Accuracy on evaluation data: 8470 / 10000
Epoch 93 training complete
Cost on training data: 0.5760981459781768
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.1228835056066304
Accuracy on evaluation data: 8475 / 10000
Epoch 94 training complete
Cost on training data: 0.5709105081459104
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.1060498607910458
Accuracy on evaluation data: 8481 / 10000
Epoch 95 training complete
Cost on training data: 0.5660342620187431
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.1124311855112545
Accuracy on evaluation data: 8468 / 10000
Epoch 96 training complete
Cost on training data: 0.5612067120458198
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.1075789221289434
Accuracy on evaluation data: 8483 / 10000
Epoch 97 training complete
Cost on training data: 0.5568060296911028
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.105766761279775
Accuracy on evaluation data: 8469 / 10000
Epoch 98 training complete
Cost on training data: 0.5518568991926573
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.1072128439657136
Accuracy on evaluation data: 8481 / 10000
Epoch 99 training complete
Cost on training data: 0.5472253893243857
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.1061384536719019
Accuracy on evaluation data: 8494 / 10000
Epoch 100 training complete
Cost on training data: 0.5426711441807877
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.100378031811514
Accuracy on evaluation data: 8500 / 10000
Epoch 101 training complete
Cost on training data: 0.5380397636741305
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.1002199662004
Accuracy on evaluation data: 8499 / 10000
Epoch 102 training complete
Cost on training data: 0.5335646434856679
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.0988213785480394
Accuracy on evaluation data: 8490 / 10000
Epoch 103 training complete
Cost on training data: 0.5294047294109829
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.0900107533164984
Accuracy on evaluation data: 8504 / 10000
Epoch 104 training complete
Cost on training data: 0.52523416065568
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.0964760601101404
Accuracy on evaluation data: 8507 / 10000
Epoch 105 training complete
Cost on training data: 0.5207395152501778
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.0927849650735937
Accuracy on evaluation data: 8492 / 10000
Epoch 106 training complete
Cost on training data: 0.5163693446450212
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.0901597358410224
Accuracy on evaluation data: 8507 / 10000
Epoch 107 training complete
Cost on training data: 0.5120147366887396
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.0844035394638436
Accuracy on evaluation data: 8509 / 10000
Epoch 108 training complete
Cost on training data: 0.5079181077933959
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.0856628699864233
Accuracy on evaluation data: 8514 / 10000
Epoch 109 training complete
Cost on training data: 0.5036345285230972
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.0849985286006492
Accuracy on evaluation data: 8513 / 10000
Epoch 110 training complete
Cost on training data: 0.49979401285433644
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.079939108150978
Accuracy on evaluation data: 8523 / 10000
Epoch 111 training complete
Cost on training data: 0.49589769801639955
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.0731377554486958
Accuracy on evaluation data: 8500 / 10000
Epoch 112 training complete
Cost on training data: 0.49160978873305156
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.080378646446618
Accuracy on evaluation data: 8513 / 10000
Epoch 113 training complete
Cost on training data: 0.4876300112581351
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.0745828097900179
Accuracy on evaluation data: 8508 / 10000
Epoch 114 training complete
Cost on training data: 0.4836723405761315
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.075208901561714
Accuracy on evaluation data: 8520 / 10000
Epoch 115 training complete
Cost on training data: 0.47973197290061925
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.0686971799792375
Accuracy on evaluation data: 8520 / 10000
Epoch 116 training complete
Cost on training data: 0.4763398529076438
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.0707577092006326
Accuracy on evaluation data: 8515 / 10000
Epoch 117 training complete
Cost on training data: 0.4721794847942432
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.0663157688165716
Accuracy on evaluation data: 8539 / 10000
Epoch 118 training complete
Cost on training data: 0.4685008823816642
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.0645619067938241
Accuracy on evaluation data: 8518 / 10000
Epoch 119 training complete
Cost on training data: 0.4649434739239494
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.0712796107664448
Accuracy on evaluation data: 8515 / 10000
Epoch 120 training complete
Cost on training data: 0.46083726276079234
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.0630403877519798
Accuracy on evaluation data: 8539 / 10000
Epoch 121 training complete
Cost on training data: 0.45737056060447684
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.058825461535758
Accuracy on evaluation data: 8524 / 10000
Epoch 122 training complete
Cost on training data: 0.45376923712463946
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.059786012154478
Accuracy on evaluation data: 8542 / 10000
Epoch 123 training complete
Cost on training data: 0.4501138092475525
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.054497592028444
Accuracy on evaluation data: 8526 / 10000
Epoch 124 training complete
Cost on training data: 0.44651199771842986
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.0565889019220311
Accuracy on evaluation data: 8535 / 10000
Epoch 125 training complete
Cost on training data: 0.4429338795642369
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.0542271093617395
Accuracy on evaluation data: 8543 / 10000
Epoch 126 training complete
Cost on training data: 0.43965173142168684
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.054063176870209
Accuracy on evaluation data: 8529 / 10000
Epoch 127 training complete
Cost on training data: 0.43615191533552083
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.05726970234447
Accuracy on evaluation data: 8533 / 10000
Epoch 128 training complete
Cost on training data: 0.4326840876960149
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.0541307579736678
Accuracy on evaluation data: 8555 / 10000
Epoch 129 training complete
Cost on training data: 0.4292515478091215
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.0452310988523206
Accuracy on evaluation data: 8566 / 10000
Epoch 130 training complete
Cost on training data: 0.42599583041749456
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.044747823372503
Accuracy on evaluation data: 8559 / 10000
Epoch 131 training complete
Cost on training data: 0.4228333858996333
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.042498469069074
Accuracy on evaluation data: 8534 / 10000
Epoch 132 training complete
Cost on training data: 0.4192813211802949
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.0431359244934744
Accuracy on evaluation data: 8559 / 10000
Epoch 133 training complete
Cost on training data: 0.4160666888406107
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.0409472442784973
Accuracy on evaluation data: 8548 / 10000
Epoch 134 training complete
Cost on training data: 0.4128396953510987
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.038907117079897
Accuracy on evaluation data: 8564 / 10000
Epoch 135 training complete
Cost on training data: 0.40976503154808963
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.0382415495506565
Accuracy on evaluation data: 8570 / 10000
Epoch 136 training complete
Cost on training data: 0.4064628935065046
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.0334902529629897
Accuracy on evaluation data: 8552 / 10000
Epoch 137 training complete
Cost on training data: 0.4033967976834742
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.032090639686283
Accuracy on evaluation data: 8555 / 10000
Epoch 138 training complete
Cost on training data: 0.40031038484944415
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.0284082484237094
Accuracy on evaluation data: 8573 / 10000
Epoch 139 training complete
Cost on training data: 0.3972970573611687
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.0258324426923557
Accuracy on evaluation data: 8556 / 10000
Epoch 140 training complete
Cost on training data: 0.3943584506185519
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.0298310624445937
Accuracy on evaluation data: 8560 / 10000
Epoch 141 training complete
Cost on training data: 0.39120683113290244
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.028337639590758
Accuracy on evaluation data: 8566 / 10000
Epoch 142 training complete
Cost on training data: 0.38848087310504403
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.018580436270432
Accuracy on evaluation data: 8557 / 10000
Epoch 143 training complete
Cost on training data: 0.38536270281883955
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.022723966140331
Accuracy on evaluation data: 8577 / 10000
Epoch 144 training complete
Cost on training data: 0.3824613198142168
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.0224744277019158
Accuracy on evaluation data: 8567 / 10000
Epoch 145 training complete
Cost on training data: 0.3797024443953414
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.0221023701873893
Accuracy on evaluation data: 8565 / 10000
Epoch 146 training complete
Cost on training data: 0.3768851487747798
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.0163006934093424
Accuracy on evaluation data: 8572 / 10000
Epoch 147 training complete
Cost on training data: 0.37397519712488925
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.016980417190237
Accuracy on evaluation data: 8577 / 10000
Epoch 148 training complete
Cost on training data: 0.37137686970197914
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.011710160417635
Accuracy on evaluation data: 8580 / 10000
Epoch 149 training complete
Cost on training data: 0.3685688099007376
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.014571757311663
Accuracy on evaluation data: 8570 / 10000
Epoch 150 training complete
Cost on training data: 0.3657936421174424
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.0108394155438951
Accuracy on evaluation data: 8572 / 10000
Epoch 151 training complete
Cost on training data: 0.36312853286413804
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.0128783771821217
Accuracy on evaluation data: 8571 / 10000
Epoch 152 training complete
Cost on training data: 0.36085510824244477
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.0097383655086116
Accuracy on evaluation data: 8566 / 10000
Epoch 153 training complete
Cost on training data: 0.35799258689175095
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.0108136294487764
Accuracy on evaluation data: 8575 / 10000
Epoch 154 training complete
Cost on training data: 0.35538710006655067
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.0077854021898414
Accuracy on evaluation data: 8585 / 10000
Epoch 155 training complete
Cost on training data: 0.35271170449808054
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.0054234676795113
Accuracy on evaluation data: 8585 / 10000
Epoch 156 training complete
Cost on training data: 0.35018199235213654
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.0040649538199333
Accuracy on evaluation data: 8587 / 10000
Epoch 157 training complete
Cost on training data: 0.34766872465283727
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 1.00090907590078
Accuracy on evaluation data: 8584 / 10000
Epoch 158 training complete
Cost on training data: 0.34531116354604563
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.9984777760002546
Accuracy on evaluation data: 8590 / 10000
Epoch 159 training complete
Cost on training data: 0.3426584939464257
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.9970503420863798
Accuracy on evaluation data: 8591 / 10000
Epoch 160 training complete
Cost on training data: 0.3403031228838666
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.9955190435795641
Accuracy on evaluation data: 8588 / 10000
Epoch 161 training complete
Cost on training data: 0.3378355694728108
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.9949213751773776
Accuracy on evaluation data: 8595 / 10000
Epoch 162 training complete
Cost on training data: 0.3356056767370044
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.9900862138776373
Accuracy on evaluation data: 8591 / 10000
Epoch 163 training complete
Cost on training data: 0.33310719669294325
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.9913091622946635
Accuracy on evaluation data: 8590 / 10000
Epoch 164 training complete
Cost on training data: 0.33085299998072476
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.9921913237587365
Accuracy on evaluation data: 8591 / 10000
Epoch 165 training complete
Cost on training data: 0.3285090890048322
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.9901275774854743
Accuracy on evaluation data: 8595 / 10000
Epoch 166 training complete
Cost on training data: 0.3263825835415328
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.9863752173064689
Accuracy on evaluation data: 8599 / 10000
Epoch 167 training complete
Cost on training data: 0.3237915297385139
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.9846798984106111
Accuracy on evaluation data: 8596 / 10000
Epoch 168 training complete
Cost on training data: 0.32158844793255303
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.9854608993112698
Accuracy on evaluation data: 8591 / 10000
Epoch 169 training complete
Cost on training data: 0.3195464556953592
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.9835123271244862
Accuracy on evaluation data: 8598 / 10000
Epoch 170 training complete
Cost on training data: 0.317141138389106
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.9796438058708128
Accuracy on evaluation data: 8608 / 10000
Epoch 171 training complete
Cost on training data: 0.3150041408200562
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.9850629369437014
Accuracy on evaluation data: 8604 / 10000
Epoch 172 training complete
Cost on training data: 0.31282405970860505
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.9775500072502445
Accuracy on evaluation data: 8604 / 10000
Epoch 173 training complete
Cost on training data: 0.31063170022820646
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.9750909696904299
Accuracy on evaluation data: 8597 / 10000
Epoch 174 training complete
Cost on training data: 0.3086164420005054
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.9735322648485687
Accuracy on evaluation data: 8616 / 10000
Epoch 175 training complete
Cost on training data: 0.30659936417506584
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.9737922837158631
Accuracy on evaluation data: 8612 / 10000
Epoch 176 training complete
Cost on training data: 0.30429730075902967
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.9730490636218293
Accuracy on evaluation data: 8597 / 10000
Epoch 177 training complete
Cost on training data: 0.30236591054689943
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.9717680417877619
Accuracy on evaluation data: 8605 / 10000
Epoch 178 training complete
Cost on training data: 0.30022352996640783
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.9657337660887518
Accuracy on evaluation data: 8619 / 10000
Epoch 179 training complete
Cost on training data: 0.2982864533786486
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.9673564428068605
Accuracy on evaluation data: 8609 / 10000
Epoch 180 training complete
Cost on training data: 0.2963812191834984
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.9659019178588237
Accuracy on evaluation data: 8613 / 10000
Epoch 181 training complete
Cost on training data: 0.2943137630449074
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.962307111801244
Accuracy on evaluation data: 8621 / 10000
Epoch 182 training complete
Cost on training data: 0.2922992628006298
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.9627495828237937
Accuracy on evaluation data: 8622 / 10000
Epoch 183 training complete
Cost on training data: 0.2902443604580467
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.9604034454622192
Accuracy on evaluation data: 8619 / 10000
Epoch 184 training complete
Cost on training data: 0.28854875075350334
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.9601704289619261
Accuracy on evaluation data: 8627 / 10000
Epoch 185 training complete
Cost on training data: 0.2863777436727088
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.957753524939023
Accuracy on evaluation data: 8615 / 10000
Epoch 186 training complete
Cost on training data: 0.2845138185768666
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.9618269604373911
Accuracy on evaluation data: 8620 / 10000
Epoch 187 training complete
Cost on training data: 0.28283263133510983
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.953499131707952
Accuracy on evaluation data: 8629 / 10000
Epoch 188 training complete
Cost on training data: 0.280866056155417
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.9575321711864925
Accuracy on evaluation data: 8613 / 10000
Epoch 189 training complete
Cost on training data: 0.279058864259464
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.9567347439271653
Accuracy on evaluation data: 8624 / 10000
Epoch 190 training complete
Cost on training data: 0.2771139561623108
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.9525171937974678
Accuracy on evaluation data: 8630 / 10000
Epoch 191 training complete
Cost on training data: 0.27536900577029405
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.949788560785232
Accuracy on evaluation data: 8621 / 10000
Epoch 192 training complete
Cost on training data: 0.27375144392891065
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.9504958774810262
Accuracy on evaluation data: 8624 / 10000
Epoch 193 training complete
Cost on training data: 0.2718968371231152
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.9464862060950203
Accuracy on evaluation data: 8639 / 10000
Epoch 194 training complete
Cost on training data: 0.27029286454149304
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.9492755218281824
Accuracy on evaluation data: 8644 / 10000
Epoch 195 training complete
Cost on training data: 0.26847142641013705
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.9566018777828571
Accuracy on evaluation data: 8627 / 10000
Epoch 196 training complete
Cost on training data: 0.2666335939921046
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.9477481981511001
Accuracy on evaluation data: 8636 / 10000
Epoch 197 training complete
Cost on training data: 0.2651374125542186
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.9396362842646786
Accuracy on evaluation data: 8652 / 10000
Epoch 198 training complete
Cost on training data: 0.2632965290836348
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.9407665642837855
Accuracy on evaluation data: 8641 / 10000
Epoch 199 training complete
Cost on training data: 0.2616826647558872
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.9441674938361294
Accuracy on evaluation data: 8639 / 10000
Epoch 200 training complete
Cost on training data: 0.25988938503472225
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.9383478224987111
Accuracy on evaluation data: 8636 / 10000
Epoch 201 training complete
Cost on training data: 0.2584688652869881
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.9381589456557016
Accuracy on evaluation data: 8641 / 10000
Epoch 202 training complete
Cost on training data: 0.25675261711143976
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.9363087006261198
Accuracy on evaluation data: 8633 / 10000
Epoch 203 training complete
Cost on training data: 0.2549806002368845
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.9374467003094413
Accuracy on evaluation data: 8647 / 10000
Epoch 204 training complete
Cost on training data: 0.25344269793200874
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.931210670511748
Accuracy on evaluation data: 8649 / 10000
Epoch 205 training complete
Cost on training data: 0.2521553929400422
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.9311850191669658
Accuracy on evaluation data: 8660 / 10000
Epoch 206 training complete
Cost on training data: 0.25037697554788413
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.9313284762598302
Accuracy on evaluation data: 8665 / 10000
Epoch 207 training complete
Cost on training data: 0.2489173769908963
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.9347901300663953
Accuracy on evaluation data: 8651 / 10000
Epoch 208 training complete
Cost on training data: 0.24729540873151434
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.9323403598425785
Accuracy on evaluation data: 8667 / 10000
Epoch 209 training complete
Cost on training data: 0.24579436325741222
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.9283407320427631
Accuracy on evaluation data: 8649 / 10000
Epoch 210 training complete
Cost on training data: 0.24424169041095997
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.923566411053962
Accuracy on evaluation data: 8677 / 10000
Epoch 211 training complete
Cost on training data: 0.24269927248517825
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.9268513094560215
Accuracy on evaluation data: 8671 / 10000
Epoch 212 training complete
Cost on training data: 0.24121942182590764
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.9216509435151032
Accuracy on evaluation data: 8667 / 10000
Epoch 213 training complete
Cost on training data: 0.23990350598194607
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.9273240920041805
Accuracy on evaluation data: 8663 / 10000
Epoch 214 training complete
Cost on training data: 0.23823197200383942
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.921769830037589
Accuracy on evaluation data: 8662 / 10000
Epoch 215 training complete
Cost on training data: 0.23701995434122158
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.9224243125300559
Accuracy on evaluation data: 8660 / 10000
Epoch 216 training complete
Cost on training data: 0.23536747054157547
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.9184186623253383
Accuracy on evaluation data: 8676 / 10000
Epoch 217 training complete
Cost on training data: 0.23419937121139953
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.9202427079076817
Accuracy on evaluation data: 8664 / 10000
Epoch 218 training complete
Cost on training data: 0.2330044036927178
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.925146712482847
Accuracy on evaluation data: 8676 / 10000
Epoch 219 training complete
Cost on training data: 0.2312596724716881
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.9206080162930965
Accuracy on evaluation data: 8673 / 10000
Epoch 220 training complete
Cost on training data: 0.23031645833158104
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.9186762344355945
Accuracy on evaluation data: 8693 / 10000
Epoch 221 training complete
Cost on training data: 0.22852373208896215
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.9157567645926983
Accuracy on evaluation data: 8681 / 10000
Epoch 222 training complete
Cost on training data: 0.22714725274498995
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.9135863072322205
Accuracy on evaluation data: 8673 / 10000
Epoch 223 training complete
Cost on training data: 0.22587134615750346
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.914875845634017
Accuracy on evaluation data: 8692 / 10000
Epoch 224 training complete
Cost on training data: 0.22471543142772396
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.9140169572443706
Accuracy on evaluation data: 8694 / 10000
Epoch 225 training complete
Cost on training data: 0.22319764742878007
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.9109334878257309
Accuracy on evaluation data: 8681 / 10000
Epoch 226 training complete
Cost on training data: 0.22203460880767
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.9118069504960488
Accuracy on evaluation data: 8679 / 10000
Epoch 227 training complete
Cost on training data: 0.22077112517778738
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.9100672062235233
Accuracy on evaluation data: 8693 / 10000
Epoch 228 training complete
Cost on training data: 0.21945204892126075
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.9076287588568888
Accuracy on evaluation data: 8690 / 10000
Epoch 229 training complete
Cost on training data: 0.21836887455102844
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.9138478062382027
Accuracy on evaluation data: 8687 / 10000
Epoch 230 training complete
Cost on training data: 0.21688496595709547
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.908990815376184
Accuracy on evaluation data: 8701 / 10000
Epoch 231 training complete
Cost on training data: 0.2158607257598287
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.9030413552962572
Accuracy on evaluation data: 8694 / 10000
Epoch 232 training complete
Cost on training data: 0.21442570199725358
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.9057739751178543
Accuracy on evaluation data: 8701 / 10000
Epoch 233 training complete
Cost on training data: 0.21335977468874118
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.905471356610055
Accuracy on evaluation data: 8686 / 10000
Epoch 234 training complete
Cost on training data: 0.21213295908548604
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.9068125487436939
Accuracy on evaluation data: 8696 / 10000
Epoch 235 training complete
Cost on training data: 0.21116580709214125
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.8985111163433342
Accuracy on evaluation data: 8691 / 10000
Epoch 236 training complete
Cost on training data: 0.20974135291887233
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.9013553847590637
Accuracy on evaluation data: 8704 / 10000
Epoch 237 training complete
Cost on training data: 0.20902771097558934
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.905843516512316
Accuracy on evaluation data: 8701 / 10000
Epoch 238 training complete
Cost on training data: 0.20752110861891707
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.8975060263870962
Accuracy on evaluation data: 8702 / 10000
Epoch 239 training complete
Cost on training data: 0.20666548641945548
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.897458046069742
Accuracy on evaluation data: 8701 / 10000
Epoch 240 training complete
Cost on training data: 0.20516917587776332
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.8972825714057207
Accuracy on evaluation data: 8704 / 10000
Epoch 241 training complete
Cost on training data: 0.20404513591729895
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.8958285017955858
Accuracy on evaluation data: 8704 / 10000
Epoch 242 training complete
Cost on training data: 0.20308501563760334
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.8933725254857114
Accuracy on evaluation data: 8715 / 10000
Epoch 243 training complete
Cost on training data: 0.20207403845826008
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.8945412115057829
Accuracy on evaluation data: 8710 / 10000
Epoch 244 training complete
Cost on training data: 0.20090103199861262
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.896354523311267
Accuracy on evaluation data: 8698 / 10000
Epoch 245 training complete
Cost on training data: 0.19985771509972403
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.8910236688436194
Accuracy on evaluation data: 8714 / 10000
Epoch 246 training complete
Cost on training data: 0.19880992614009171
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.8959842012066697
Accuracy on evaluation data: 8708 / 10000
Epoch 247 training complete
Cost on training data: 0.19762300820897988
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.8896306922687188
Accuracy on evaluation data: 8714 / 10000
Epoch 248 training complete
Cost on training data: 0.19665061492181354
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.8871209620294733
Accuracy on evaluation data: 8717 / 10000
Epoch 249 training complete
Cost on training data: 0.19571459431875518
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.8871345609980276
Accuracy on evaluation data: 8711 / 10000
Epoch 250 training complete
Cost on training data: 0.19467215390331755
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.8827273599503881
Accuracy on evaluation data: 8726 / 10000
Epoch 251 training complete
Cost on training data: 0.19369183099718398
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.8874416355569478
Accuracy on evaluation data: 8718 / 10000
Epoch 252 training complete
Cost on training data: 0.192713459763124
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.885017442705497
Accuracy on evaluation data: 8719 / 10000
Epoch 253 training complete
Cost on training data: 0.19165592854161145
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.8796660589499661
Accuracy on evaluation data: 8719 / 10000
Epoch 254 training complete
Cost on training data: 0.1906992830498837
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.8852527111928207
Accuracy on evaluation data: 8717 / 10000
Epoch 255 training complete
Cost on training data: 0.18987857405109324
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.8768442183460146
Accuracy on evaluation data: 8729 / 10000
Epoch 256 training complete
Cost on training data: 0.18876869005230054
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.8803530344668762
Accuracy on evaluation data: 8724 / 10000
Epoch 257 training complete
Cost on training data: 0.18790291522859634
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.8796840253468329
Accuracy on evaluation data: 8724 / 10000
Epoch 258 training complete
Cost on training data: 0.18680638466839836
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.8814789381055727
Accuracy on evaluation data: 8722 / 10000
Epoch 259 training complete
Cost on training data: 0.1858223497920047
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.8782524710274955
Accuracy on evaluation data: 8722 / 10000
Epoch 260 training complete
Cost on training data: 0.1851331972053311
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.8818642229781822
Accuracy on evaluation data: 8714 / 10000
Epoch 261 training complete
Cost on training data: 0.18405047616237283
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.8766672977860988
Accuracy on evaluation data: 8735 / 10000
Epoch 262 training complete
Cost on training data: 0.1832290889453867
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.873742031222319
Accuracy on evaluation data: 8727 / 10000
Epoch 263 training complete
Cost on training data: 0.18237111496653696
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.8781558197293345
Accuracy on evaluation data: 8718 / 10000
Epoch 264 training complete
Cost on training data: 0.18158797740176535
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.8787899570140779
Accuracy on evaluation data: 8731 / 10000
Epoch 265 training complete
Cost on training data: 0.1804788401626141
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.8784740042938518
Accuracy on evaluation data: 8727 / 10000
Epoch 266 training complete
Cost on training data: 0.17961304241880377
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.8745434667728088
Accuracy on evaluation data: 8728 / 10000
Epoch 267 training complete
Cost on training data: 0.17880555004096815
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.8747834270544793
Accuracy on evaluation data: 8732 / 10000
Epoch 268 training complete
Cost on training data: 0.17803010361153762
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.8753472940556462
Accuracy on evaluation data: 8730 / 10000
Epoch 269 training complete
Cost on training data: 0.17709268685236854
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.8702290007130803
Accuracy on evaluation data: 8746 / 10000
Epoch 270 training complete
Cost on training data: 0.1761624007930124
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.8724291169250498
Accuracy on evaluation data: 8731 / 10000
Epoch 271 training complete
Cost on training data: 0.17526785299409112
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.8695080400735214
Accuracy on evaluation data: 8742 / 10000
Epoch 272 training complete
Cost on training data: 0.17440818970390012
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.868413955057592
Accuracy on evaluation data: 8742 / 10000
Epoch 273 training complete
Cost on training data: 0.17371291586772553
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.8661118930541547
Accuracy on evaluation data: 8730 / 10000
Epoch 274 training complete
Cost on training data: 0.17281687516691144
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.8722480394357472
Accuracy on evaluation data: 8750 / 10000
Epoch 275 training complete
Cost on training data: 0.17212397937429325
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.8708098226644576
Accuracy on evaluation data: 8747 / 10000
Epoch 276 training complete
Cost on training data: 0.17130834243087967
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.8680567324915787
Accuracy on evaluation data: 8738 / 10000
Epoch 277 training complete
Cost on training data: 0.1705983452888632
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.8679780190483897
Accuracy on evaluation data: 8733 / 10000
Epoch 278 training complete
Cost on training data: 0.16976784290285465
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.8687125166801766
Accuracy on evaluation data: 8741 / 10000
Epoch 279 training complete
Cost on training data: 0.16897487447558304
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.868062550136856
Accuracy on evaluation data: 8741 / 10000
Epoch 280 training complete
Cost on training data: 0.1681435488497453
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.8664250486527446
Accuracy on evaluation data: 8739 / 10000
Epoch 281 training complete
Cost on training data: 0.16752242651900234
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.8596361075275913
Accuracy on evaluation data: 8742 / 10000
Epoch 282 training complete
Cost on training data: 0.16694517728433494
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.8644967911578578
Accuracy on evaluation data: 8746 / 10000
Epoch 283 training complete
Cost on training data: 0.16591901283459615
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.8661258031447138
Accuracy on evaluation data: 8745 / 10000
Epoch 284 training complete
Cost on training data: 0.16545278495789495
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.8691742560718051
Accuracy on evaluation data: 8742 / 10000
Epoch 285 training complete
Cost on training data: 0.16461081030928087
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.860170074662745
Accuracy on evaluation data: 8752 / 10000
Epoch 286 training complete
Cost on training data: 0.16387809903379055
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.8593284942714079
Accuracy on evaluation data: 8744 / 10000
Epoch 287 training complete
Cost on training data: 0.16304514019924352
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.8592619320684322
Accuracy on evaluation data: 8748 / 10000
Epoch 288 training complete
Cost on training data: 0.16235159857895218
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.8593827738015778
Accuracy on evaluation data: 8739 / 10000
Epoch 289 training complete
Cost on training data: 0.16168332722360398
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.8603761701070151
Accuracy on evaluation data: 8751 / 10000
Epoch 290 training complete
Cost on training data: 0.16085056685944527
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.8561631786061727
Accuracy on evaluation data: 8749 / 10000
Epoch 291 training complete
Cost on training data: 0.16028388092220697
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.861405446259314
Accuracy on evaluation data: 8749 / 10000
Epoch 292 training complete
Cost on training data: 0.1597981086525645
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.8499553395549216
Accuracy on evaluation data: 8761 / 10000
Epoch 293 training complete
Cost on training data: 0.15891831916068125
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.8561751683714229
Accuracy on evaluation data: 8764 / 10000
Epoch 294 training complete
Cost on training data: 0.15841224299383208
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.8570601534398546
Accuracy on evaluation data: 8758 / 10000
Epoch 295 training complete
Cost on training data: 0.15763790106980125
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.8591204786877937
Accuracy on evaluation data: 8750 / 10000
Epoch 296 training complete
Cost on training data: 0.1570692587700943
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.8516669886751229
Accuracy on evaluation data: 8762 / 10000
Epoch 297 training complete
Cost on training data: 0.1564893556655105
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.8604790151106874
Accuracy on evaluation data: 8752 / 10000
Epoch 298 training complete
Cost on training data: 0.15581712150175053
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.8590736400347022
Accuracy on evaluation data: 8754 / 10000
Epoch 299 training complete
Cost on training data: 0.1550418737771008
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.8563272926522457
Accuracy on evaluation data: 8758 / 10000
Epoch 300 training complete
Cost on training data: 0.1546069113991705
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.8633989711839911
Accuracy on evaluation data: 8750 / 10000
Epoch 301 training complete
Cost on training data: 0.15384522961996192
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.8575040517327326
Accuracy on evaluation data: 8755 / 10000
Epoch 302 training complete
Cost on training data: 0.153355404670292
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.8548103414629044
Accuracy on evaluation data: 8756 / 10000
Epoch 303 training complete
Cost on training data: 0.15256887885695644
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.848563865397737
Accuracy on evaluation data: 8773 / 10000
Epoch 304 training complete
Cost on training data: 0.1520211753401304
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.8581834888737044
Accuracy on evaluation data: 8750 / 10000
Epoch 305 training complete
Cost on training data: 0.15153503383253425
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.8500322556026624
Accuracy on evaluation data: 8763 / 10000
Epoch 306 training complete
Cost on training data: 0.15132816513799746
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.8546914421092543
Accuracy on evaluation data: 8766 / 10000
Epoch 307 training complete
Cost on training data: 0.15018184983382027
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.8470091349826819
Accuracy on evaluation data: 8770 / 10000
Epoch 308 training complete
Cost on training data: 0.14979032889717372
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.8529945270926104
Accuracy on evaluation data: 8765 / 10000
Epoch 309 training complete
Cost on training data: 0.14919603036390208
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.8451131173525848
Accuracy on evaluation data: 8764 / 10000
Epoch 310 training complete
Cost on training data: 0.1486027239692737
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.8509555995075965
Accuracy on evaluation data: 8764 / 10000
Epoch 311 training complete
Cost on training data: 0.14797657037804773
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.8511266568604978
Accuracy on evaluation data: 8768 / 10000
Epoch 312 training complete
Cost on training data: 0.1477608150699406
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.8574676773062803
Accuracy on evaluation data: 8763 / 10000
Epoch 313 training complete
Cost on training data: 0.14692974751943663
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.850444061083833
Accuracy on evaluation data: 8768 / 10000
Epoch 314 training complete
Cost on training data: 0.14668362017644135
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.8474883114403449
Accuracy on evaluation data: 8762 / 10000
Epoch 315 training complete
Cost on training data: 0.14582269857815605
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.8471343928275197
Accuracy on evaluation data: 8773 / 10000
Epoch 316 training complete
Cost on training data: 0.14528504532811662
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.8486264785179806
Accuracy on evaluation data: 8773 / 10000
Epoch 317 training complete
Cost on training data: 0.14491095205977828
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.8451266002607654
Accuracy on evaluation data: 8775 / 10000
Epoch 318 training complete
Cost on training data: 0.14426993117418524
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.8433642733803693
Accuracy on evaluation data: 8780 / 10000
Epoch 319 training complete
Cost on training data: 0.1438321603511557
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.8468581179612625
Accuracy on evaluation data: 8758 / 10000
Epoch 320 training complete
Cost on training data: 0.14317215897867297
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.840482956925582
Accuracy on evaluation data: 8785 / 10000
Epoch 321 training complete
Cost on training data: 0.14292495379330583
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.8448835571000327
Accuracy on evaluation data: 8778 / 10000
Epoch 322 training complete
Cost on training data: 0.14209766562468454
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.8415624299216345
Accuracy on evaluation data: 8780 / 10000
Epoch 323 training complete
Cost on training data: 0.14184968306277826
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.8438949761469066
Accuracy on evaluation data: 8778 / 10000
Epoch 324 training complete
Cost on training data: 0.14121532608975104
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.8435455043809489
Accuracy on evaluation data: 8773 / 10000
Epoch 325 training complete
Cost on training data: 0.1407352278292615
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.8459187913893154
Accuracy on evaluation data: 8772 / 10000
Epoch 326 training complete
Cost on training data: 0.14056005979323763
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.8468729253172412
Accuracy on evaluation data: 8791 / 10000
Epoch 327 training complete
Cost on training data: 0.13984309974868192
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.8438960543392997
Accuracy on evaluation data: 8776 / 10000
Epoch 328 training complete
Cost on training data: 0.1393031840077545
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.8461526625852205
Accuracy on evaluation data: 8772 / 10000
Epoch 329 training complete
Cost on training data: 0.13893443875108552
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.8466725252038557
Accuracy on evaluation data: 8774 / 10000
Epoch 330 training complete
Cost on training data: 0.13836333971624212
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.8412929901870647
Accuracy on evaluation data: 8775 / 10000
Epoch 331 training complete
Cost on training data: 0.13798121398605698
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.8379412428434512
Accuracy on evaluation data: 8773 / 10000
Epoch 332 training complete
Cost on training data: 0.13756579456342607
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.839826115428956
Accuracy on evaluation data: 8785 / 10000
Epoch 333 training complete
Cost on training data: 0.13699774549693122
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.8360784938337461
Accuracy on evaluation data: 8791 / 10000
Epoch 334 training complete
Cost on training data: 0.13661254851615787
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.843131035966578
Accuracy on evaluation data: 8782 / 10000
Epoch 335 training complete
Cost on training data: 0.13618621814987159
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.8417472732606027
Accuracy on evaluation data: 8783 / 10000
Epoch 336 training complete
Cost on training data: 0.13565012325818337
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.8419977742433912
Accuracy on evaluation data: 8783 / 10000
Epoch 337 training complete
Cost on training data: 0.13525905839785057
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.8410148641516034
Accuracy on evaluation data: 8780 / 10000
Epoch 338 training complete
Cost on training data: 0.13491559339179013
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.8423164484264181
Accuracy on evaluation data: 8779 / 10000
Epoch 339 training complete
Cost on training data: 0.1344323535815333
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.842664809038009
Accuracy on evaluation data: 8785 / 10000
Epoch 340 training complete
Cost on training data: 0.13404731061822198
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.8405227423735879
Accuracy on evaluation data: 8779 / 10000
Epoch 341 training complete
Cost on training data: 0.13355373396820433
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.8396472815854342
Accuracy on evaluation data: 8785 / 10000
Epoch 342 training complete
Cost on training data: 0.13307040752759386
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.8385306816693389
Accuracy on evaluation data: 8786 / 10000
Epoch 343 training complete
Cost on training data: 0.13275218974931693
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.838663780165443
Accuracy on evaluation data: 8783 / 10000
Epoch 344 training complete
Cost on training data: 0.1324433897836284
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.8322896913418766
Accuracy on evaluation data: 8796 / 10000
Epoch 345 training complete
Cost on training data: 0.13209933579839758
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.8450391804977546
Accuracy on evaluation data: 8784 / 10000
Epoch 346 training complete
Cost on training data: 0.1314316701511752
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.8356146399803589
Accuracy on evaluation data: 8796 / 10000
Epoch 347 training complete
Cost on training data: 0.13131368532748672
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.8357746382980542
Accuracy on evaluation data: 8780 / 10000
Epoch 348 training complete
Cost on training data: 0.13072936057769857
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.8369195080731275
Accuracy on evaluation data: 8780 / 10000
Epoch 349 training complete
Cost on training data: 0.130349814479051
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.8421027358067932
Accuracy on evaluation data: 8783 / 10000
Epoch 350 training complete
Cost on training data: 0.1300845659978494
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.8376856239448361
Accuracy on evaluation data: 8785 / 10000
Epoch 351 training complete
Cost on training data: 0.1298038673776775
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.8456000839095424
Accuracy on evaluation data: 8777 / 10000
Epoch 352 training complete
Cost on training data: 0.1292694531459024
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.8378961092644421
Accuracy on evaluation data: 8788 / 10000
Epoch 353 training complete
Cost on training data: 0.1289583127009526
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.8303405004961687
Accuracy on evaluation data: 8797 / 10000
Epoch 354 training complete
Cost on training data: 0.12838220214374454
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.8376466419269839
Accuracy on evaluation data: 8788 / 10000
Epoch 355 training complete
Cost on training data: 0.1281504276141033
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.842210245651563
Accuracy on evaluation data: 8775 / 10000
Epoch 356 training complete
Cost on training data: 0.12795474293066292
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.8345268646356802
Accuracy on evaluation data: 8786 / 10000
Epoch 357 training complete
Cost on training data: 0.12732501909625765
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.8336268109642112
Accuracy on evaluation data: 8797 / 10000
Epoch 358 training complete
Cost on training data: 0.12728450912521477
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.841932351706898
Accuracy on evaluation data: 8770 / 10000
Epoch 359 training complete
Cost on training data: 0.12716156293969375
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.8367820574512923
Accuracy on evaluation data: 8776 / 10000
Epoch 360 training complete
Cost on training data: 0.12638695145313797
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.8321287637891325
Accuracy on evaluation data: 8790 / 10000
Epoch 361 training complete
Cost on training data: 0.12615606989434983
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.8353553242508271
Accuracy on evaluation data: 8798 / 10000
Epoch 362 training complete
Cost on training data: 0.1257316679583279
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.8383729452223393
Accuracy on evaluation data: 8798 / 10000
Epoch 363 training complete
Cost on training data: 0.12535505181304776
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.8424090170979928
Accuracy on evaluation data: 8787 / 10000
Epoch 364 training complete
Cost on training data: 0.12498550456407347
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.8317105456597119
Accuracy on evaluation data: 8797 / 10000
Epoch 365 training complete
Cost on training data: 0.12477479503094986
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.8308041582396846
Accuracy on evaluation data: 8800 / 10000
Epoch 366 training complete
Cost on training data: 0.12436206198020541
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.8376529645035693
Accuracy on evaluation data: 8780 / 10000
Epoch 367 training complete
Cost on training data: 0.12393343236225658
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.8352680732768047
Accuracy on evaluation data: 8795 / 10000
Epoch 368 training complete
Cost on training data: 0.12359744371206685
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.8325036849006247
Accuracy on evaluation data: 8791 / 10000
Epoch 369 training complete
Cost on training data: 0.12338945974522555
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.8352216074832646
Accuracy on evaluation data: 8797 / 10000
Epoch 370 training complete
Cost on training data: 0.12334604145825466
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.8431169662289458
Accuracy on evaluation data: 8793 / 10000
Epoch 371 training complete
Cost on training data: 0.12283331725188486
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.8285677028643451
Accuracy on evaluation data: 8790 / 10000
Epoch 372 training complete
Cost on training data: 0.12247976087574089
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.8359789613879433
Accuracy on evaluation data: 8792 / 10000
Epoch 373 training complete
Cost on training data: 0.12240715213924068
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.8288199442690634
Accuracy on evaluation data: 8806 / 10000
Epoch 374 training complete
Cost on training data: 0.1218214815554502
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.835155191862855
Accuracy on evaluation data: 8798 / 10000
Epoch 375 training complete
Cost on training data: 0.1216793161345654
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.8363447777000258
Accuracy on evaluation data: 8800 / 10000
Epoch 376 training complete
Cost on training data: 0.1212549411994653
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.8302505190580958
Accuracy on evaluation data: 8801 / 10000
Epoch 377 training complete
Cost on training data: 0.12091637896576013
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.8276392341166917
Accuracy on evaluation data: 8793 / 10000
Epoch 378 training complete
Cost on training data: 0.12063856179848394
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.8318815791445036
Accuracy on evaluation data: 8801 / 10000
Epoch 379 training complete
Cost on training data: 0.12024383608696004
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.8332620110540632
Accuracy on evaluation data: 8803 / 10000
Epoch 380 training complete
Cost on training data: 0.1200298853153678
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.8304495324699973
Accuracy on evaluation data: 8799 / 10000
Epoch 381 training complete
Cost on training data: 0.11993416981604951
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.830252214438044
Accuracy on evaluation data: 8787 / 10000
Epoch 382 training complete
Cost on training data: 0.11969626290203031
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.8273790633700642
Accuracy on evaluation data: 8804 / 10000
Epoch 383 training complete
Cost on training data: 0.11916854601410923
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.8307873439050035
Accuracy on evaluation data: 8797 / 10000
Epoch 384 training complete
Cost on training data: 0.1188639298321219
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.8333581634199259
Accuracy on evaluation data: 8799 / 10000
Epoch 385 training complete
Cost on training data: 0.11868051764206795
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.8330874905761427
Accuracy on evaluation data: 8801 / 10000
Epoch 386 training complete
Cost on training data: 0.11850167669051143
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.8307881689869424
Accuracy on evaluation data: 8802 / 10000
Epoch 387 training complete
Cost on training data: 0.11811432976982253
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.8327728592264831
Accuracy on evaluation data: 8798 / 10000
Epoch 388 training complete
Cost on training data: 0.11796020316729892
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.8348151818920478
Accuracy on evaluation data: 8783 / 10000
Epoch 389 training complete
Cost on training data: 0.11753848255812442
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.8320888422582724
Accuracy on evaluation data: 8795 / 10000
Epoch 390 training complete
Cost on training data: 0.11725473032176129
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.8332554737252845
Accuracy on evaluation data: 8801 / 10000
Epoch 391 training complete
Cost on training data: 0.11708829042162623
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.8349636796974668
Accuracy on evaluation data: 8802 / 10000
Epoch 392 training complete
Cost on training data: 0.11678559008525219
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.8319947481378631
Accuracy on evaluation data: 8792 / 10000
Epoch 393 training complete
Cost on training data: 0.11669829208430223
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.8269181803307477
Accuracy on evaluation data: 8801 / 10000
Epoch 394 training complete
Cost on training data: 0.11641542629859208
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.832593000267452
Accuracy on evaluation data: 8804 / 10000
Epoch 395 training complete
Cost on training data: 0.11597176808243001
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.830180287413146
Accuracy on evaluation data: 8802 / 10000
Epoch 396 training complete
Cost on training data: 0.11573759472250579
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.828598271123198
Accuracy on evaluation data: 8795 / 10000
Epoch 397 training complete
Cost on training data: 0.11567845127433003
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.8303041352390372
Accuracy on evaluation data: 8808 / 10000
Epoch 398 training complete
Cost on training data: 0.1153003438591176
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.825470604759123
Accuracy on evaluation data: 8799 / 10000
Epoch 399 training complete
Cost on training data: 0.11499844911758361
Accuracy on training data: 1000 / 1000
Cost on evaluation data: 0.8284975824916161
Accuracy on evaluation data: 8800 / 10000

CPU times: user 14min 42s, sys: 11min 45s, total: 26min 28s
Wall time: 16min 41s

training_cost_xmin=200
test_accuracy_xmin=200
test_cost_xmin=0
training_accuracy_xmin=0
training_set_size=1000
lmbda=0.1

make_plots('/content/train1000-lambda_point1', 400,
               training_cost_xmin,
               test_accuracy_xmin,
               test_cost_xmin,
               training_accuracy_xmin,
               training_set_size)

%%time
if __name__ == "__main__":
    filename = input("Enter a file name: ")
    num_epochs = int(input("Enter the number of epochs to run for: "))
    training_cost_xmin = int(input("training_cost_xmin (suggest 200): "))
    test_accuracy_xmin = int(input("test_accuracy_xmin (suggest 200): "))
    test_cost_xmin = int(input("test_cost_xmin (suggest 0): "))
    training_accuracy_xmin = int(input("training_accuracy_xmin (suggest 0): "))
    training_set_size = int(input("Training set size (suggest 1000): "))
    lmbda = float(input("Enter the regularization parameter, lambda (suggest: 5.0): "))
    main(filename, num_epochs, training_cost_xmin,
         test_accuracy_xmin, test_cost_xmin, training_accuracy_xmin,
         training_set_size, lmbda)

Enter a file name: train50000-lambda5
Enter the number of epochs to run for: 30
training_cost_xmin (suggest 200): 15
test_accuracy_xmin (suggest 200): 15
test_cost_xmin (suggest 0): 0
training_accuracy_xmin (suggest 0): 0
Training set size (suggest 1000): 50000
Enter the regularization parameter, lambda (suggest: 5.0): 5
Epoch 0 training complete
Cost on training data: 1.2734727241718988
Accuracy on training data: 46164 / 50000
Cost on evaluation data: 4.2905433057644355
Accuracy on evaluation data: 9216 / 10000
Epoch 1 training complete
Cost on training data: 0.9419156008205658
Accuracy on training data: 46721 / 50000
Cost on evaluation data: 2.9450097579589958
Accuracy on evaluation data: 9279 / 10000
Epoch 2 training complete
Cost on training data: 0.7565864531292588
Accuracy on training data: 46957 / 50000
Cost on evaluation data: 2.155298294411187
Accuracy on evaluation data: 9336 / 10000
Epoch 3 training complete
Cost on training data: 0.6028287666804373
Accuracy on training data: 47445 / 50000
Cost on evaluation data: 1.6458785962838665
Accuracy on evaluation data: 9445 / 10000
Epoch 4 training complete
Cost on training data: 0.5124487479761338
Accuracy on training data: 47766 / 50000
Cost on evaluation data: 1.3640756256352797
Accuracy on evaluation data: 9465 / 10000
Epoch 5 training complete
Cost on training data: 0.4988030577938535
Accuracy on training data: 47580 / 50000
Cost on evaluation data: 1.2157115394650067
Accuracy on evaluation data: 9455 / 10000
Epoch 6 training complete
Cost on training data: 0.43609154230748126
Accuracy on training data: 47986 / 50000
Cost on evaluation data: 1.11371122465704
Accuracy on evaluation data: 9496 / 10000
Epoch 7 training complete
Cost on training data: 0.42985246104424635
Accuracy on training data: 48046 / 50000
Cost on evaluation data: 1.056222328708533
Accuracy on evaluation data: 9526 / 10000
Epoch 8 training complete
Cost on training data: 0.4167811747375013
Accuracy on training data: 48013 / 50000
Cost on evaluation data: 1.0273016132708597
Accuracy on evaluation data: 9512 / 10000
Epoch 9 training complete
Cost on training data: 0.4034474632442032
Accuracy on training data: 48226 / 50000
Cost on evaluation data: 0.9997864707487703
Accuracy on evaluation data: 9573 / 10000
Epoch 10 training complete
Cost on training data: 0.38505878309243957
Accuracy on training data: 48318 / 50000
Cost on evaluation data: 0.9861193088556921
Accuracy on evaluation data: 9566 / 10000
Epoch 11 training complete
Cost on training data: 0.4405024885937533
Accuracy on training data: 47801 / 50000
Cost on evaluation data: 1.0344830258067927
Accuracy on evaluation data: 9499 / 10000
Epoch 12 training complete
Cost on training data: 0.42016577221228857
Accuracy on training data: 47956 / 50000
Cost on evaluation data: 1.01032844850172
Accuracy on evaluation data: 9479 / 10000
Epoch 13 training complete
Cost on training data: 0.4245495018669271
Accuracy on training data: 47991 / 50000
Cost on evaluation data: 1.025928470139144
Accuracy on evaluation data: 9523 / 10000
Epoch 14 training complete
Cost on training data: 0.37420541639298843
Accuracy on training data: 48428 / 50000
Cost on evaluation data: 0.9717311023828072
Accuracy on evaluation data: 9600 / 10000
Epoch 15 training complete
Cost on training data: 0.3727184218730436
Accuracy on training data: 48442 / 50000
Cost on evaluation data: 0.9684745026763021
Accuracy on evaluation data: 9582 / 10000
Epoch 16 training complete
Cost on training data: 0.36119260065521486
Accuracy on training data: 48554 / 50000
Cost on evaluation data: 0.9550291952912006
Accuracy on evaluation data: 9620 / 10000
Epoch 17 training complete
Cost on training data: 0.3913731761457112
Accuracy on training data: 48308 / 50000
Cost on evaluation data: 0.9904390273644688
Accuracy on evaluation data: 9567 / 10000
Epoch 18 training complete
Cost on training data: 0.3985734385684887
Accuracy on training data: 48344 / 50000
Cost on evaluation data: 0.987286695088534
Accuracy on evaluation data: 9552 / 10000
Epoch 19 training complete
Cost on training data: 0.35887388840256534
Accuracy on training data: 48571 / 50000
Cost on evaluation data: 0.9597223344502038
Accuracy on evaluation data: 9607 / 10000
Epoch 20 training complete
Cost on training data: 0.3906438501336217
Accuracy on training data: 48351 / 50000
Cost on evaluation data: 0.9961442351509222
Accuracy on evaluation data: 9569 / 10000
Epoch 21 training complete
Cost on training data: 0.372900408436176
Accuracy on training data: 48495 / 50000
Cost on evaluation data: 0.9673109260072522
Accuracy on evaluation data: 9608 / 10000
Epoch 22 training complete
Cost on training data: 0.36846837975730923
Accuracy on training data: 48499 / 50000
Cost on evaluation data: 0.9639326651932834
Accuracy on evaluation data: 9605 / 10000
Epoch 23 training complete
Cost on training data: 0.3560745890509801
Accuracy on training data: 48536 / 50000
Cost on evaluation data: 0.9566863207067859
Accuracy on evaluation data: 9613 / 10000
Epoch 24 training complete
Cost on training data: 0.3819290164271956
Accuracy on training data: 48422 / 50000
Cost on evaluation data: 0.9732715096591932
Accuracy on evaluation data: 9596 / 10000
Epoch 25 training complete
Cost on training data: 0.3641173885958884
Accuracy on training data: 48573 / 50000
Cost on evaluation data: 0.9610934702051449
Accuracy on evaluation data: 9616 / 10000
Epoch 26 training complete
Cost on training data: 0.3706151345751465
Accuracy on training data: 48479 / 50000
Cost on evaluation data: 0.9671416611522785
Accuracy on evaluation data: 9613 / 10000
Epoch 27 training complete
Cost on training data: 0.38072640598772167
Accuracy on training data: 48417 / 50000
Cost on evaluation data: 0.9801406029906612
Accuracy on evaluation data: 9590 / 10000
Epoch 28 training complete
Cost on training data: 0.37538138281760225
Accuracy on training data: 48466 / 50000
Cost on evaluation data: 0.9707955441233471
Accuracy on evaluation data: 9611 / 10000
Epoch 29 training complete
Cost on training data: 0.3720172066378544
Accuracy on training data: 48491 / 50000
Cost on evaluation data: 0.9704432700280308
Accuracy on evaluation data: 9613 / 10000

CPU times: user 5min 52s, sys: 3min 59s, total: 9min 51s
Wall time: 6min 41s

import numpy as np
import matplotlib.pyplot as plt
from numpy.polynomial import Polynomial

# Data points
x = np.array([0.7, 1.3, 2, 2.7, 3, 3.7, 3.9, 4.1, 4.5, 4.9])
# y = np.array([1.8, 2.2, 4, 5, 6, 7, 7.5, 8, 9.2, 10])
y = np.array([p(i) for i in x])

# Polynomial function
def p(x):
    return (2.20539187e-01 * x**9
            - 5.49142821 * x**8
            + 5.87844045e+01 * x**7
            - 3.53892824e+02 * x**6
            + 1.31549254e+03 * x**5
            - 3.11809836e+03 * x**4
            + 4.69080366e+03 * x**3
            - 4.29612493e+03 * x**2
            + 2.16228823e+03 * x
            - 4.50983951e+02)

# Function to create a plot
def create_plot(ax, x, y, fit_func=None, title=''):
    ax.scatter(x, y, color='black')
    ax.set_xlim(0, 5)
    ax.set_ylim(0, 11)
    ax.set_xlabel('x')
    ax.set_ylabel('y')
    ax.set_title(title)

    if fit_func:
        x_fit = np.linspace(0, 5, 1000)
        y_fit = fit_func(x_fit)
        ax.plot(x_fit, y_fit, color='blue')

# Create a figure with 3 subplots
fig, (ax1, ax2, ax3) = plt.subplots(3, 1, figsize=(5, 10))

# Plot 1: Simple scatter plot
create_plot(ax1, x, y, title='Simple Scatter Plot')

# Plot 2: 9th order polynomial fit
poly = Polynomial.fit(x, y, 9)
create_plot(ax2, x, y, fit_func=p, title='9th Order Polynomial Fit')

# Plot 3: Linear model y = 2x
linear_model = lambda x: 2 * x
create_plot(ax3, x, y, fit_func=linear_model, title='Linear Model: y = 2x')

plt.tight_layout()
plt.show()

# @title
fig, ax = plt.subplots()
create_plot(ax, x, y, title='Simple Scatter Plot')
plt.tight_layout()
plt.show()

# @title
fig, ax = plt.subplots()
create_plot(ax, x, y, fit_func=p, title='9th Order Polynomial Fit')
plt.tight_layout()
plt.show()

# @title
fig, ax = plt.subplots()
create_plot(ax, x, y, fit_func=linear_model, title='Linear Model: y = 2x')
plt.tight_layout()
plt.show()

# @title
"""more_data
~~~~~~~~~~~~

Plot graphs to illustrate the performance of MNIST when different size
training sets are used.

"""

# Standard library
import json
import random
import sys

# My library
sys.path.append('../src/')
import mnist_loader
# import network2

# Third-party libraries
import matplotlib.pyplot as plt
import numpy as np
from sklearn import svm

# The sizes to use for the different training sets
SIZES = [100, 200, 500, 1000, 2000, 5000, 10000, 20000, 50000]

def main():
    run_networks()
    run_svms()
    make_plots()

def run_networks():
    # Make results more easily reproducible
    random.seed(12345678)
    np.random.seed(12345678)
    training_data, validation_data, test_data = mnist_loader.load_data_wrapper()

    # Convert data to lists
    training_data = list(training_data)
    validation_data = list(validation_data)
    test_data = list(test_data)

    net = Network2mat([784, 30, 10], cost=CrossEntropyCost())
    accuracies = []
    for size in SIZES:
        print("\n\nTraining network with data set size %s" % size)
        net.large_weight_initializer()
        num_epochs = int(1500000 / size)
        net.SGD(training_data[:size], num_epochs, 10, 0.5, lmbda = size*0.0001)
        accuracy = net.accuracy(validation_data) / 100.0
        print("Accuracy was %s percent" % accuracy)
        accuracies.append(accuracy)
    f = open("more_data.json", "w")
    json.dump(accuracies, f)
    f.close()

def run_svms():
    svm_training_data, svm_validation_data, svm_test_data \
        = mnist_loader.load_data()
    accuracies = []
    for size in SIZES:
        print("\n\nTraining SVM with data set size %s" % size)
        clf = svm.SVC()
        clf.fit(svm_training_data[0][:size], svm_training_data[1][:size])
        predictions = [int(a) for a in clf.predict(svm_validation_data[0])]
        accuracy = sum(int(a == y) for a, y in
                       zip(predictions, svm_validation_data[1])) / 100.0
        print("Accuracy was %s percent" % accuracy)
        accuracies.append(accuracy)
    f = open("more_data_svm.json", "w")
    json.dump(accuracies, f)
    f.close()

def make_plots():
    f = open("more_data.json", "r")
    accuracies = json.load(f)
    f.close()
    f = open("more_data_svm.json", "r")
    svm_accuracies = json.load(f)
    f.close()
    make_linear_plot(accuracies)
    make_log_plot(accuracies)
    make_combined_plot(accuracies, svm_accuracies)

def make_linear_plot(accuracies):
    fig = plt.figure()
    ax = fig.add_subplot(111)
    ax.plot(SIZES, accuracies, color='#2A6EA6')
    ax.plot(SIZES, accuracies, "o", color='#FFA933')
    ax.set_xlim(0, 50000)
    ax.set_ylim(60, 100)
    ax.grid(True)
    ax.set_xlabel('Training set size')
    ax.set_title('Accuracy (%) on the validation data')
    plt.show()

def make_log_plot(accuracies):
    fig = plt.figure()
    ax = fig.add_subplot(111)
    ax.plot(SIZES, accuracies, color='#2A6EA6')
    ax.plot(SIZES, accuracies, "o", color='#FFA933')
    ax.set_xlim(100, 50000)
    ax.set_ylim(60, 100)
    ax.set_xscale('log')
    ax.grid(True)
    ax.set_xlabel('Training set size')
    ax.set_title('Accuracy (%) on the validation data')
    plt.show()

def make_combined_plot(accuracies, svm_accuracies):
    fig = plt.figure()
    ax = fig.add_subplot(111)
    ax.plot(SIZES, accuracies, color='#2A6EA6')
    ax.plot(SIZES, accuracies, "o", color='#2A6EA6',
            label='Neural network accuracy (%)')
    ax.plot(SIZES, svm_accuracies, color='#FFA933')
    ax.plot(SIZES, svm_accuracies, "o", color='#FFA933',
            label='SVM accuracy (%)')
    ax.set_xlim(100, 50000)
    ax.set_ylim(25, 100)
    ax.set_xscale('log')
    ax.grid(True)
    ax.set_xlabel('Training set size')
    plt.legend(loc="lower right")
    plt.show()

if __name__ == "__main__":
    main()

Streaming output truncated to the last 5000 lines.
Epoch 866 training complete
Epoch 867 training complete
Epoch 868 training complete
Epoch 869 training complete
Epoch 870 training complete
Epoch 871 training complete
Epoch 872 training complete
Epoch 873 training complete
Epoch 874 training complete
Epoch 875 training complete
Epoch 876 training complete
Epoch 877 training complete
Epoch 878 training complete
Epoch 879 training complete
Epoch 880 training complete
Epoch 881 training complete
Epoch 882 training complete
Epoch 883 training complete
Epoch 884 training complete
Epoch 885 training complete
Epoch 886 training complete
Epoch 887 training complete
Epoch 888 training complete
Epoch 889 training complete
Epoch 890 training complete
Epoch 891 training complete
Epoch 892 training complete
Epoch 893 training complete
Epoch 894 training complete
Epoch 895 training complete
Epoch 896 training complete
Epoch 897 training complete
Epoch 898 training complete
Epoch 899 training complete
Epoch 900 training complete
Epoch 901 training complete
Epoch 902 training complete
Epoch 903 training complete
Epoch 904 training complete
Epoch 905 training complete
Epoch 906 training complete
Epoch 907 training complete
Epoch 908 training complete
Epoch 909 training complete
Epoch 910 training complete
Epoch 911 training complete
Epoch 912 training complete
Epoch 913 training complete
Epoch 914 training complete
Epoch 915 training complete
Epoch 916 training complete
Epoch 917 training complete
Epoch 918 training complete
Epoch 919 training complete
Epoch 920 training complete
Epoch 921 training complete
Epoch 922 training complete
Epoch 923 training complete
Epoch 924 training complete
Epoch 925 training complete
Epoch 926 training complete
Epoch 927 training complete
Epoch 928 training complete
Epoch 929 training complete
Epoch 930 training complete
Epoch 931 training complete
Epoch 932 training complete
Epoch 933 training complete
Epoch 934 training complete
Epoch 935 training complete
Epoch 936 training complete
Epoch 937 training complete
Epoch 938 training complete
Epoch 939 training complete
Epoch 940 training complete
Epoch 941 training complete
Epoch 942 training complete
Epoch 943 training complete
Epoch 944 training complete
Epoch 945 training complete
Epoch 946 training complete
Epoch 947 training complete
Epoch 948 training complete
Epoch 949 training complete
Epoch 950 training complete
Epoch 951 training complete
Epoch 952 training complete
Epoch 953 training complete
Epoch 954 training complete
Epoch 955 training complete
Epoch 956 training complete
Epoch 957 training complete
Epoch 958 training complete
Epoch 959 training complete
Epoch 960 training complete
Epoch 961 training complete
Epoch 962 training complete
Epoch 963 training complete
Epoch 964 training complete
Epoch 965 training complete
Epoch 966 training complete
Epoch 967 training complete
Epoch 968 training complete
Epoch 969 training complete
Epoch 970 training complete
Epoch 971 training complete
Epoch 972 training complete
Epoch 973 training complete
Epoch 974 training complete
Epoch 975 training complete
Epoch 976 training complete
Epoch 977 training complete
Epoch 978 training complete
Epoch 979 training complete
Epoch 980 training complete
Epoch 981 training complete
Epoch 982 training complete
Epoch 983 training complete
Epoch 984 training complete
Epoch 985 training complete
Epoch 986 training complete
Epoch 987 training complete
Epoch 988 training complete
Epoch 989 training complete
Epoch 990 training complete
Epoch 991 training complete
Epoch 992 training complete
Epoch 993 training complete
Epoch 994 training complete
Epoch 995 training complete
Epoch 996 training complete
Epoch 997 training complete
Epoch 998 training complete
Epoch 999 training complete
Epoch 1000 training complete
Epoch 1001 training complete
Epoch 1002 training complete
Epoch 1003 training complete
Epoch 1004 training complete
Epoch 1005 training complete
Epoch 1006 training complete
Epoch 1007 training complete
Epoch 1008 training complete
Epoch 1009 training complete
Epoch 1010 training complete
Epoch 1011 training complete
Epoch 1012 training complete
Epoch 1013 training complete
Epoch 1014 training complete
Epoch 1015 training complete
Epoch 1016 training complete
Epoch 1017 training complete
Epoch 1018 training complete
Epoch 1019 training complete
Epoch 1020 training complete
Epoch 1021 training complete
Epoch 1022 training complete
Epoch 1023 training complete
Epoch 1024 training complete
Epoch 1025 training complete
Epoch 1026 training complete
Epoch 1027 training complete
Epoch 1028 training complete
Epoch 1029 training complete
Epoch 1030 training complete
Epoch 1031 training complete
Epoch 1032 training complete
Epoch 1033 training complete
Epoch 1034 training complete
Epoch 1035 training complete
Epoch 1036 training complete
Epoch 1037 training complete
Epoch 1038 training complete
Epoch 1039 training complete
Epoch 1040 training complete
Epoch 1041 training complete
Epoch 1042 training complete
Epoch 1043 training complete
Epoch 1044 training complete
Epoch 1045 training complete
Epoch 1046 training complete
Epoch 1047 training complete
Epoch 1048 training complete
Epoch 1049 training complete
Epoch 1050 training complete
Epoch 1051 training complete
Epoch 1052 training complete
Epoch 1053 training complete
Epoch 1054 training complete
Epoch 1055 training complete
Epoch 1056 training complete
Epoch 1057 training complete
Epoch 1058 training complete
Epoch 1059 training complete
Epoch 1060 training complete
Epoch 1061 training complete
Epoch 1062 training complete
Epoch 1063 training complete
Epoch 1064 training complete
Epoch 1065 training complete
Epoch 1066 training complete
Epoch 1067 training complete
Epoch 1068 training complete
Epoch 1069 training complete
Epoch 1070 training complete
Epoch 1071 training complete
Epoch 1072 training complete
Epoch 1073 training complete
Epoch 1074 training complete
Epoch 1075 training complete
Epoch 1076 training complete
Epoch 1077 training complete
Epoch 1078 training complete
Epoch 1079 training complete
Epoch 1080 training complete
Epoch 1081 training complete
Epoch 1082 training complete
Epoch 1083 training complete
Epoch 1084 training complete
Epoch 1085 training complete
Epoch 1086 training complete
Epoch 1087 training complete
Epoch 1088 training complete
Epoch 1089 training complete
Epoch 1090 training complete
Epoch 1091 training complete
Epoch 1092 training complete
Epoch 1093 training complete
Epoch 1094 training complete
Epoch 1095 training complete
Epoch 1096 training complete
Epoch 1097 training complete
Epoch 1098 training complete
Epoch 1099 training complete
Epoch 1100 training complete
Epoch 1101 training complete
Epoch 1102 training complete
Epoch 1103 training complete
Epoch 1104 training complete
Epoch 1105 training complete
Epoch 1106 training complete
Epoch 1107 training complete
Epoch 1108 training complete
Epoch 1109 training complete
Epoch 1110 training complete
Epoch 1111 training complete
Epoch 1112 training complete
Epoch 1113 training complete
Epoch 1114 training complete
Epoch 1115 training complete
Epoch 1116 training complete
Epoch 1117 training complete
Epoch 1118 training complete
Epoch 1119 training complete
Epoch 1120 training complete
Epoch 1121 training complete
Epoch 1122 training complete
Epoch 1123 training complete
Epoch 1124 training complete
Epoch 1125 training complete
Epoch 1126 training complete
Epoch 1127 training complete
Epoch 1128 training complete
Epoch 1129 training complete
Epoch 1130 training complete
Epoch 1131 training complete
Epoch 1132 training complete
Epoch 1133 training complete
Epoch 1134 training complete
Epoch 1135 training complete
Epoch 1136 training complete
Epoch 1137 training complete
Epoch 1138 training complete
Epoch 1139 training complete
Epoch 1140 training complete
Epoch 1141 training complete
Epoch 1142 training complete
Epoch 1143 training complete
Epoch 1144 training complete
Epoch 1145 training complete
Epoch 1146 training complete
Epoch 1147 training complete
Epoch 1148 training complete
Epoch 1149 training complete
Epoch 1150 training complete
Epoch 1151 training complete
Epoch 1152 training complete
Epoch 1153 training complete
Epoch 1154 training complete
Epoch 1155 training complete
Epoch 1156 training complete
Epoch 1157 training complete
Epoch 1158 training complete
Epoch 1159 training complete
Epoch 1160 training complete
Epoch 1161 training complete
Epoch 1162 training complete
Epoch 1163 training complete
Epoch 1164 training complete
Epoch 1165 training complete
Epoch 1166 training complete
Epoch 1167 training complete
Epoch 1168 training complete
Epoch 1169 training complete
Epoch 1170 training complete
Epoch 1171 training complete
Epoch 1172 training complete
Epoch 1173 training complete
Epoch 1174 training complete
Epoch 1175 training complete
Epoch 1176 training complete
Epoch 1177 training complete
Epoch 1178 training complete
Epoch 1179 training complete
Epoch 1180 training complete
Epoch 1181 training complete
Epoch 1182 training complete
Epoch 1183 training complete
Epoch 1184 training complete
Epoch 1185 training complete
Epoch 1186 training complete
Epoch 1187 training complete
Epoch 1188 training complete
Epoch 1189 training complete
Epoch 1190 training complete
Epoch 1191 training complete
Epoch 1192 training complete
Epoch 1193 training complete
Epoch 1194 training complete
Epoch 1195 training complete
Epoch 1196 training complete
Epoch 1197 training complete
Epoch 1198 training complete
Epoch 1199 training complete
Epoch 1200 training complete
Epoch 1201 training complete
Epoch 1202 training complete
Epoch 1203 training complete
Epoch 1204 training complete
Epoch 1205 training complete
Epoch 1206 training complete
Epoch 1207 training complete
Epoch 1208 training complete
Epoch 1209 training complete
Epoch 1210 training complete
Epoch 1211 training complete
Epoch 1212 training complete
Epoch 1213 training complete
Epoch 1214 training complete
Epoch 1215 training complete
Epoch 1216 training complete
Epoch 1217 training complete
Epoch 1218 training complete
Epoch 1219 training complete
Epoch 1220 training complete
Epoch 1221 training complete
Epoch 1222 training complete
Epoch 1223 training complete
Epoch 1224 training complete
Epoch 1225 training complete
Epoch 1226 training complete
Epoch 1227 training complete
Epoch 1228 training complete
Epoch 1229 training complete
Epoch 1230 training complete
Epoch 1231 training complete
Epoch 1232 training complete
Epoch 1233 training complete
Epoch 1234 training complete
Epoch 1235 training complete
Epoch 1236 training complete
Epoch 1237 training complete
Epoch 1238 training complete
Epoch 1239 training complete
Epoch 1240 training complete
Epoch 1241 training complete
Epoch 1242 training complete
Epoch 1243 training complete
Epoch 1244 training complete
Epoch 1245 training complete
Epoch 1246 training complete
Epoch 1247 training complete
Epoch 1248 training complete
Epoch 1249 training complete
Epoch 1250 training complete
Epoch 1251 training complete
Epoch 1252 training complete
Epoch 1253 training complete
Epoch 1254 training complete
Epoch 1255 training complete
Epoch 1256 training complete
Epoch 1257 training complete
Epoch 1258 training complete
Epoch 1259 training complete
Epoch 1260 training complete
Epoch 1261 training complete
Epoch 1262 training complete
Epoch 1263 training complete
Epoch 1264 training complete
Epoch 1265 training complete
Epoch 1266 training complete
Epoch 1267 training complete
Epoch 1268 training complete
Epoch 1269 training complete
Epoch 1270 training complete
Epoch 1271 training complete
Epoch 1272 training complete
Epoch 1273 training complete
Epoch 1274 training complete
Epoch 1275 training complete
Epoch 1276 training complete
Epoch 1277 training complete
Epoch 1278 training complete
Epoch 1279 training complete
Epoch 1280 training complete
Epoch 1281 training complete
Epoch 1282 training complete
Epoch 1283 training complete
Epoch 1284 training complete
Epoch 1285 training complete
Epoch 1286 training complete
Epoch 1287 training complete
Epoch 1288 training complete
Epoch 1289 training complete
Epoch 1290 training complete
Epoch 1291 training complete
Epoch 1292 training complete
Epoch 1293 training complete
Epoch 1294 training complete
Epoch 1295 training complete
Epoch 1296 training complete
Epoch 1297 training complete
Epoch 1298 training complete
Epoch 1299 training complete
Epoch 1300 training complete
Epoch 1301 training complete
Epoch 1302 training complete
Epoch 1303 training complete
Epoch 1304 training complete
Epoch 1305 training complete
Epoch 1306 training complete
Epoch 1307 training complete
Epoch 1308 training complete
Epoch 1309 training complete
Epoch 1310 training complete
Epoch 1311 training complete
Epoch 1312 training complete
Epoch 1313 training complete
Epoch 1314 training complete
Epoch 1315 training complete
Epoch 1316 training complete
Epoch 1317 training complete
Epoch 1318 training complete
Epoch 1319 training complete
Epoch 1320 training complete
Epoch 1321 training complete
Epoch 1322 training complete
Epoch 1323 training complete
Epoch 1324 training complete
Epoch 1325 training complete
Epoch 1326 training complete
Epoch 1327 training complete
Epoch 1328 training complete
Epoch 1329 training complete
Epoch 1330 training complete
Epoch 1331 training complete
Epoch 1332 training complete
Epoch 1333 training complete
Epoch 1334 training complete
Epoch 1335 training complete
Epoch 1336 training complete
Epoch 1337 training complete
Epoch 1338 training complete
Epoch 1339 training complete
Epoch 1340 training complete
Epoch 1341 training complete
Epoch 1342 training complete
Epoch 1343 training complete
Epoch 1344 training complete
Epoch 1345 training complete
Epoch 1346 training complete
Epoch 1347 training complete
Epoch 1348 training complete
Epoch 1349 training complete
Epoch 1350 training complete
Epoch 1351 training complete
Epoch 1352 training complete
Epoch 1353 training complete
Epoch 1354 training complete
Epoch 1355 training complete
Epoch 1356 training complete
Epoch 1357 training complete
Epoch 1358 training complete
Epoch 1359 training complete
Epoch 1360 training complete
Epoch 1361 training complete
Epoch 1362 training complete
Epoch 1363 training complete
Epoch 1364 training complete
Epoch 1365 training complete
Epoch 1366 training complete
Epoch 1367 training complete
Epoch 1368 training complete
Epoch 1369 training complete
Epoch 1370 training complete
Epoch 1371 training complete
Epoch 1372 training complete
Epoch 1373 training complete
Epoch 1374 training complete
Epoch 1375 training complete
Epoch 1376 training complete
Epoch 1377 training complete
Epoch 1378 training complete
Epoch 1379 training complete
Epoch 1380 training complete
Epoch 1381 training complete
Epoch 1382 training complete
Epoch 1383 training complete
Epoch 1384 training complete
Epoch 1385 training complete
Epoch 1386 training complete
Epoch 1387 training complete
Epoch 1388 training complete
Epoch 1389 training complete
Epoch 1390 training complete
Epoch 1391 training complete
Epoch 1392 training complete
Epoch 1393 training complete
Epoch 1394 training complete
Epoch 1395 training complete
Epoch 1396 training complete
Epoch 1397 training complete
Epoch 1398 training complete
Epoch 1399 training complete
Epoch 1400 training complete
Epoch 1401 training complete
Epoch 1402 training complete
Epoch 1403 training complete
Epoch 1404 training complete
Epoch 1405 training complete
Epoch 1406 training complete
Epoch 1407 training complete
Epoch 1408 training complete
Epoch 1409 training complete
Epoch 1410 training complete
Epoch 1411 training complete
Epoch 1412 training complete
Epoch 1413 training complete
Epoch 1414 training complete
Epoch 1415 training complete
Epoch 1416 training complete
Epoch 1417 training complete
Epoch 1418 training complete
Epoch 1419 training complete
Epoch 1420 training complete
Epoch 1421 training complete
Epoch 1422 training complete
Epoch 1423 training complete
Epoch 1424 training complete
Epoch 1425 training complete
Epoch 1426 training complete
Epoch 1427 training complete
Epoch 1428 training complete
Epoch 1429 training complete
Epoch 1430 training complete
Epoch 1431 training complete
Epoch 1432 training complete
Epoch 1433 training complete
Epoch 1434 training complete
Epoch 1435 training complete
Epoch 1436 training complete
Epoch 1437 training complete
Epoch 1438 training complete
Epoch 1439 training complete
Epoch 1440 training complete
Epoch 1441 training complete
Epoch 1442 training complete
Epoch 1443 training complete
Epoch 1444 training complete
Epoch 1445 training complete
Epoch 1446 training complete
Epoch 1447 training complete
Epoch 1448 training complete
Epoch 1449 training complete
Epoch 1450 training complete
Epoch 1451 training complete
Epoch 1452 training complete
Epoch 1453 training complete
Epoch 1454 training complete
Epoch 1455 training complete
Epoch 1456 training complete
Epoch 1457 training complete
Epoch 1458 training complete
Epoch 1459 training complete
Epoch 1460 training complete
Epoch 1461 training complete
Epoch 1462 training complete
Epoch 1463 training complete
Epoch 1464 training complete
Epoch 1465 training complete
Epoch 1466 training complete
Epoch 1467 training complete
Epoch 1468 training complete
Epoch 1469 training complete
Epoch 1470 training complete
Epoch 1471 training complete
Epoch 1472 training complete
Epoch 1473 training complete
Epoch 1474 training complete
Epoch 1475 training complete
Epoch 1476 training complete
Epoch 1477 training complete
Epoch 1478 training complete
Epoch 1479 training complete
Epoch 1480 training complete
Epoch 1481 training complete
Epoch 1482 training complete
Epoch 1483 training complete
Epoch 1484 training complete
Epoch 1485 training complete
Epoch 1486 training complete
Epoch 1487 training complete
Epoch 1488 training complete
Epoch 1489 training complete
Epoch 1490 training complete
Epoch 1491 training complete
Epoch 1492 training complete
Epoch 1493 training complete
Epoch 1494 training complete
Epoch 1495 training complete
Epoch 1496 training complete
Epoch 1497 training complete
Epoch 1498 training complete
Epoch 1499 training complete
Epoch 1500 training complete
Epoch 1501 training complete
Epoch 1502 training complete
Epoch 1503 training complete
Epoch 1504 training complete
Epoch 1505 training complete
Epoch 1506 training complete
Epoch 1507 training complete
Epoch 1508 training complete
Epoch 1509 training complete
Epoch 1510 training complete
Epoch 1511 training complete
Epoch 1512 training complete
Epoch 1513 training complete
Epoch 1514 training complete
Epoch 1515 training complete
Epoch 1516 training complete
Epoch 1517 training complete
Epoch 1518 training complete
Epoch 1519 training complete
Epoch 1520 training complete
Epoch 1521 training complete
Epoch 1522 training complete
Epoch 1523 training complete
Epoch 1524 training complete
Epoch 1525 training complete
Epoch 1526 training complete
Epoch 1527 training complete
Epoch 1528 training complete
Epoch 1529 training complete
Epoch 1530 training complete
Epoch 1531 training complete
Epoch 1532 training complete
Epoch 1533 training complete
Epoch 1534 training complete
Epoch 1535 training complete
Epoch 1536 training complete
Epoch 1537 training complete
Epoch 1538 training complete
Epoch 1539 training complete
Epoch 1540 training complete
Epoch 1541 training complete
Epoch 1542 training complete
Epoch 1543 training complete
Epoch 1544 training complete
Epoch 1545 training complete
Epoch 1546 training complete
Epoch 1547 training complete
Epoch 1548 training complete
Epoch 1549 training complete
Epoch 1550 training complete
Epoch 1551 training complete
Epoch 1552 training complete
Epoch 1553 training complete
Epoch 1554 training complete
Epoch 1555 training complete
Epoch 1556 training complete
Epoch 1557 training complete
Epoch 1558 training complete
Epoch 1559 training complete
Epoch 1560 training complete
Epoch 1561 training complete
Epoch 1562 training complete
Epoch 1563 training complete
Epoch 1564 training complete
Epoch 1565 training complete
Epoch 1566 training complete
Epoch 1567 training complete
Epoch 1568 training complete
Epoch 1569 training complete
Epoch 1570 training complete
Epoch 1571 training complete
Epoch 1572 training complete
Epoch 1573 training complete
Epoch 1574 training complete
Epoch 1575 training complete
Epoch 1576 training complete
Epoch 1577 training complete
Epoch 1578 training complete
Epoch 1579 training complete
Epoch 1580 training complete
Epoch 1581 training complete
Epoch 1582 training complete
Epoch 1583 training complete
Epoch 1584 training complete
Epoch 1585 training complete
Epoch 1586 training complete
Epoch 1587 training complete
Epoch 1588 training complete
Epoch 1589 training complete
Epoch 1590 training complete
Epoch 1591 training complete
Epoch 1592 training complete
Epoch 1593 training complete
Epoch 1594 training complete
Epoch 1595 training complete
Epoch 1596 training complete
Epoch 1597 training complete
Epoch 1598 training complete
Epoch 1599 training complete
Epoch 1600 training complete
Epoch 1601 training complete
Epoch 1602 training complete
Epoch 1603 training complete
Epoch 1604 training complete
Epoch 1605 training complete
Epoch 1606 training complete
Epoch 1607 training complete
Epoch 1608 training complete
Epoch 1609 training complete
Epoch 1610 training complete
Epoch 1611 training complete
Epoch 1612 training complete
Epoch 1613 training complete
Epoch 1614 training complete
Epoch 1615 training complete
Epoch 1616 training complete
Epoch 1617 training complete
Epoch 1618 training complete
Epoch 1619 training complete
Epoch 1620 training complete
Epoch 1621 training complete
Epoch 1622 training complete
Epoch 1623 training complete
Epoch 1624 training complete
Epoch 1625 training complete
Epoch 1626 training complete
Epoch 1627 training complete
Epoch 1628 training complete
Epoch 1629 training complete
Epoch 1630 training complete
Epoch 1631 training complete
Epoch 1632 training complete
Epoch 1633 training complete
Epoch 1634 training complete
Epoch 1635 training complete
Epoch 1636 training complete
Epoch 1637 training complete
Epoch 1638 training complete
Epoch 1639 training complete
Epoch 1640 training complete
Epoch 1641 training complete
Epoch 1642 training complete
Epoch 1643 training complete
Epoch 1644 training complete
Epoch 1645 training complete
Epoch 1646 training complete
Epoch 1647 training complete
Epoch 1648 training complete
Epoch 1649 training complete
Epoch 1650 training complete
Epoch 1651 training complete
Epoch 1652 training complete
Epoch 1653 training complete
Epoch 1654 training complete
Epoch 1655 training complete
Epoch 1656 training complete
Epoch 1657 training complete
Epoch 1658 training complete
Epoch 1659 training complete
Epoch 1660 training complete
Epoch 1661 training complete
Epoch 1662 training complete
Epoch 1663 training complete
Epoch 1664 training complete
Epoch 1665 training complete
Epoch 1666 training complete
Epoch 1667 training complete
Epoch 1668 training complete
Epoch 1669 training complete
Epoch 1670 training complete
Epoch 1671 training complete
Epoch 1672 training complete
Epoch 1673 training complete
Epoch 1674 training complete
Epoch 1675 training complete
Epoch 1676 training complete
Epoch 1677 training complete
Epoch 1678 training complete
Epoch 1679 training complete
Epoch 1680 training complete
Epoch 1681 training complete
Epoch 1682 training complete
Epoch 1683 training complete
Epoch 1684 training complete
Epoch 1685 training complete
Epoch 1686 training complete
Epoch 1687 training complete
Epoch 1688 training complete
Epoch 1689 training complete
Epoch 1690 training complete
Epoch 1691 training complete
Epoch 1692 training complete
Epoch 1693 training complete
Epoch 1694 training complete
Epoch 1695 training complete
Epoch 1696 training complete
Epoch 1697 training complete
Epoch 1698 training complete
Epoch 1699 training complete
Epoch 1700 training complete
Epoch 1701 training complete
Epoch 1702 training complete
Epoch 1703 training complete
Epoch 1704 training complete
Epoch 1705 training complete
Epoch 1706 training complete
Epoch 1707 training complete
Epoch 1708 training complete
Epoch 1709 training complete
Epoch 1710 training complete
Epoch 1711 training complete
Epoch 1712 training complete
Epoch 1713 training complete
Epoch 1714 training complete
Epoch 1715 training complete
Epoch 1716 training complete
Epoch 1717 training complete
Epoch 1718 training complete
Epoch 1719 training complete
Epoch 1720 training complete
Epoch 1721 training complete
Epoch 1722 training complete
Epoch 1723 training complete
Epoch 1724 training complete
Epoch 1725 training complete
Epoch 1726 training complete
Epoch 1727 training complete
Epoch 1728 training complete
Epoch 1729 training complete
Epoch 1730 training complete
Epoch 1731 training complete
Epoch 1732 training complete
Epoch 1733 training complete
Epoch 1734 training complete
Epoch 1735 training complete
Epoch 1736 training complete
Epoch 1737 training complete
Epoch 1738 training complete
Epoch 1739 training complete
Epoch 1740 training complete
Epoch 1741 training complete
Epoch 1742 training complete
Epoch 1743 training complete
Epoch 1744 training complete
Epoch 1745 training complete
Epoch 1746 training complete
Epoch 1747 training complete
Epoch 1748 training complete
Epoch 1749 training complete
Epoch 1750 training complete
Epoch 1751 training complete
Epoch 1752 training complete
Epoch 1753 training complete
Epoch 1754 training complete
Epoch 1755 training complete
Epoch 1756 training complete
Epoch 1757 training complete
Epoch 1758 training complete
Epoch 1759 training complete
Epoch 1760 training complete
Epoch 1761 training complete
Epoch 1762 training complete
Epoch 1763 training complete
Epoch 1764 training complete
Epoch 1765 training complete
Epoch 1766 training complete
Epoch 1767 training complete
Epoch 1768 training complete
Epoch 1769 training complete
Epoch 1770 training complete
Epoch 1771 training complete
Epoch 1772 training complete
Epoch 1773 training complete
Epoch 1774 training complete
Epoch 1775 training complete
Epoch 1776 training complete
Epoch 1777 training complete
Epoch 1778 training complete
Epoch 1779 training complete
Epoch 1780 training complete
Epoch 1781 training complete
Epoch 1782 training complete
Epoch 1783 training complete
Epoch 1784 training complete
Epoch 1785 training complete
Epoch 1786 training complete
Epoch 1787 training complete
Epoch 1788 training complete
Epoch 1789 training complete
Epoch 1790 training complete
Epoch 1791 training complete
Epoch 1792 training complete
Epoch 1793 training complete
Epoch 1794 training complete
Epoch 1795 training complete
Epoch 1796 training complete
Epoch 1797 training complete
Epoch 1798 training complete
Epoch 1799 training complete
Epoch 1800 training complete
Epoch 1801 training complete
Epoch 1802 training complete
Epoch 1803 training complete
Epoch 1804 training complete
Epoch 1805 training complete
Epoch 1806 training complete
Epoch 1807 training complete
Epoch 1808 training complete
Epoch 1809 training complete
Epoch 1810 training complete
Epoch 1811 training complete
Epoch 1812 training complete
Epoch 1813 training complete
Epoch 1814 training complete
Epoch 1815 training complete
Epoch 1816 training complete
Epoch 1817 training complete
Epoch 1818 training complete
Epoch 1819 training complete
Epoch 1820 training complete
Epoch 1821 training complete
Epoch 1822 training complete
Epoch 1823 training complete
Epoch 1824 training complete
Epoch 1825 training complete
Epoch 1826 training complete
Epoch 1827 training complete
Epoch 1828 training complete
Epoch 1829 training complete
Epoch 1830 training complete
Epoch 1831 training complete
Epoch 1832 training complete
Epoch 1833 training complete
Epoch 1834 training complete
Epoch 1835 training complete
Epoch 1836 training complete
Epoch 1837 training complete
Epoch 1838 training complete
Epoch 1839 training complete
Epoch 1840 training complete
Epoch 1841 training complete
Epoch 1842 training complete
Epoch 1843 training complete
Epoch 1844 training complete
Epoch 1845 training complete
Epoch 1846 training complete
Epoch 1847 training complete
Epoch 1848 training complete
Epoch 1849 training complete
Epoch 1850 training complete
Epoch 1851 training complete
Epoch 1852 training complete
Epoch 1853 training complete
Epoch 1854 training complete
Epoch 1855 training complete
Epoch 1856 training complete
Epoch 1857 training complete
Epoch 1858 training complete
Epoch 1859 training complete
Epoch 1860 training complete
Epoch 1861 training complete
Epoch 1862 training complete
Epoch 1863 training complete
Epoch 1864 training complete
Epoch 1865 training complete
Epoch 1866 training complete
Epoch 1867 training complete
Epoch 1868 training complete
Epoch 1869 training complete
Epoch 1870 training complete
Epoch 1871 training complete
Epoch 1872 training complete
Epoch 1873 training complete
Epoch 1874 training complete
Epoch 1875 training complete
Epoch 1876 training complete
Epoch 1877 training complete
Epoch 1878 training complete
Epoch 1879 training complete
Epoch 1880 training complete
Epoch 1881 training complete
Epoch 1882 training complete
Epoch 1883 training complete
Epoch 1884 training complete
Epoch 1885 training complete
Epoch 1886 training complete
Epoch 1887 training complete
Epoch 1888 training complete
Epoch 1889 training complete
Epoch 1890 training complete
Epoch 1891 training complete
Epoch 1892 training complete
Epoch 1893 training complete
Epoch 1894 training complete
Epoch 1895 training complete
Epoch 1896 training complete
Epoch 1897 training complete
Epoch 1898 training complete
Epoch 1899 training complete
Epoch 1900 training complete
Epoch 1901 training complete
Epoch 1902 training complete
Epoch 1903 training complete
Epoch 1904 training complete
Epoch 1905 training complete
Epoch 1906 training complete
Epoch 1907 training complete
Epoch 1908 training complete
Epoch 1909 training complete
Epoch 1910 training complete
Epoch 1911 training complete
Epoch 1912 training complete
Epoch 1913 training complete
Epoch 1914 training complete
Epoch 1915 training complete
Epoch 1916 training complete
Epoch 1917 training complete
Epoch 1918 training complete
Epoch 1919 training complete
Epoch 1920 training complete
Epoch 1921 training complete
Epoch 1922 training complete
Epoch 1923 training complete
Epoch 1924 training complete
Epoch 1925 training complete
Epoch 1926 training complete
Epoch 1927 training complete
Epoch 1928 training complete
Epoch 1929 training complete
Epoch 1930 training complete
Epoch 1931 training complete
Epoch 1932 training complete
Epoch 1933 training complete
Epoch 1934 training complete
Epoch 1935 training complete
Epoch 1936 training complete
Epoch 1937 training complete
Epoch 1938 training complete
Epoch 1939 training complete
Epoch 1940 training complete
Epoch 1941 training complete
Epoch 1942 training complete
Epoch 1943 training complete
Epoch 1944 training complete
Epoch 1945 training complete
Epoch 1946 training complete
Epoch 1947 training complete
Epoch 1948 training complete
Epoch 1949 training complete
Epoch 1950 training complete
Epoch 1951 training complete
Epoch 1952 training complete
Epoch 1953 training complete
Epoch 1954 training complete
Epoch 1955 training complete
Epoch 1956 training complete
Epoch 1957 training complete
Epoch 1958 training complete
Epoch 1959 training complete
Epoch 1960 training complete
Epoch 1961 training complete
Epoch 1962 training complete
Epoch 1963 training complete
Epoch 1964 training complete
Epoch 1965 training complete
Epoch 1966 training complete
Epoch 1967 training complete
Epoch 1968 training complete
Epoch 1969 training complete
Epoch 1970 training complete
Epoch 1971 training complete
Epoch 1972 training complete
Epoch 1973 training complete
Epoch 1974 training complete
Epoch 1975 training complete
Epoch 1976 training complete
Epoch 1977 training complete
Epoch 1978 training complete
Epoch 1979 training complete
Epoch 1980 training complete
Epoch 1981 training complete
Epoch 1982 training complete
Epoch 1983 training complete
Epoch 1984 training complete
Epoch 1985 training complete
Epoch 1986 training complete
Epoch 1987 training complete
Epoch 1988 training complete
Epoch 1989 training complete
Epoch 1990 training complete
Epoch 1991 training complete
Epoch 1992 training complete
Epoch 1993 training complete
Epoch 1994 training complete
Epoch 1995 training complete
Epoch 1996 training complete
Epoch 1997 training complete
Epoch 1998 training complete
Epoch 1999 training complete
Epoch 2000 training complete
Epoch 2001 training complete
Epoch 2002 training complete
Epoch 2003 training complete
Epoch 2004 training complete
Epoch 2005 training complete
Epoch 2006 training complete
Epoch 2007 training complete
Epoch 2008 training complete
Epoch 2009 training complete
Epoch 2010 training complete
Epoch 2011 training complete
Epoch 2012 training complete
Epoch 2013 training complete
Epoch 2014 training complete
Epoch 2015 training complete
Epoch 2016 training complete
Epoch 2017 training complete
Epoch 2018 training complete
Epoch 2019 training complete
Epoch 2020 training complete
Epoch 2021 training complete
Epoch 2022 training complete
Epoch 2023 training complete
Epoch 2024 training complete
Epoch 2025 training complete
Epoch 2026 training complete
Epoch 2027 training complete
Epoch 2028 training complete
Epoch 2029 training complete
Epoch 2030 training complete
Epoch 2031 training complete
Epoch 2032 training complete
Epoch 2033 training complete
Epoch 2034 training complete
Epoch 2035 training complete
Epoch 2036 training complete
Epoch 2037 training complete
Epoch 2038 training complete
Epoch 2039 training complete
Epoch 2040 training complete
Epoch 2041 training complete
Epoch 2042 training complete
Epoch 2043 training complete
Epoch 2044 training complete
Epoch 2045 training complete
Epoch 2046 training complete
Epoch 2047 training complete
Epoch 2048 training complete
Epoch 2049 training complete
Epoch 2050 training complete
Epoch 2051 training complete
Epoch 2052 training complete
Epoch 2053 training complete
Epoch 2054 training complete
Epoch 2055 training complete
Epoch 2056 training complete
Epoch 2057 training complete
Epoch 2058 training complete
Epoch 2059 training complete
Epoch 2060 training complete
Epoch 2061 training complete
Epoch 2062 training complete
Epoch 2063 training complete
Epoch 2064 training complete
Epoch 2065 training complete
Epoch 2066 training complete
Epoch 2067 training complete
Epoch 2068 training complete
Epoch 2069 training complete
Epoch 2070 training complete
Epoch 2071 training complete
Epoch 2072 training complete
Epoch 2073 training complete
Epoch 2074 training complete
Epoch 2075 training complete
Epoch 2076 training complete
Epoch 2077 training complete
Epoch 2078 training complete
Epoch 2079 training complete
Epoch 2080 training complete
Epoch 2081 training complete
Epoch 2082 training complete
Epoch 2083 training complete
Epoch 2084 training complete
Epoch 2085 training complete
Epoch 2086 training complete
Epoch 2087 training complete
Epoch 2088 training complete
Epoch 2089 training complete
Epoch 2090 training complete
Epoch 2091 training complete
Epoch 2092 training complete
Epoch 2093 training complete
Epoch 2094 training complete
Epoch 2095 training complete
Epoch 2096 training complete
Epoch 2097 training complete
Epoch 2098 training complete
Epoch 2099 training complete
Epoch 2100 training complete
Epoch 2101 training complete
Epoch 2102 training complete
Epoch 2103 training complete
Epoch 2104 training complete
Epoch 2105 training complete
Epoch 2106 training complete
Epoch 2107 training complete
Epoch 2108 training complete
Epoch 2109 training complete
Epoch 2110 training complete
Epoch 2111 training complete
Epoch 2112 training complete
Epoch 2113 training complete
Epoch 2114 training complete
Epoch 2115 training complete
Epoch 2116 training complete
Epoch 2117 training complete
Epoch 2118 training complete
Epoch 2119 training complete
Epoch 2120 training complete
Epoch 2121 training complete
Epoch 2122 training complete
Epoch 2123 training complete
Epoch 2124 training complete
Epoch 2125 training complete
Epoch 2126 training complete
Epoch 2127 training complete
Epoch 2128 training complete
Epoch 2129 training complete
Epoch 2130 training complete
Epoch 2131 training complete
Epoch 2132 training complete
Epoch 2133 training complete
Epoch 2134 training complete
Epoch 2135 training complete
Epoch 2136 training complete
Epoch 2137 training complete
Epoch 2138 training complete
Epoch 2139 training complete
Epoch 2140 training complete
Epoch 2141 training complete
Epoch 2142 training complete
Epoch 2143 training complete
Epoch 2144 training complete
Epoch 2145 training complete
Epoch 2146 training complete
Epoch 2147 training complete
Epoch 2148 training complete
Epoch 2149 training complete
Epoch 2150 training complete
Epoch 2151 training complete
Epoch 2152 training complete
Epoch 2153 training complete
Epoch 2154 training complete
Epoch 2155 training complete
Epoch 2156 training complete
Epoch 2157 training complete
Epoch 2158 training complete
Epoch 2159 training complete
Epoch 2160 training complete
Epoch 2161 training complete
Epoch 2162 training complete
Epoch 2163 training complete
Epoch 2164 training complete
Epoch 2165 training complete
Epoch 2166 training complete
Epoch 2167 training complete
Epoch 2168 training complete
Epoch 2169 training complete
Epoch 2170 training complete
Epoch 2171 training complete
Epoch 2172 training complete
Epoch 2173 training complete
Epoch 2174 training complete
Epoch 2175 training complete
Epoch 2176 training complete
Epoch 2177 training complete
Epoch 2178 training complete
Epoch 2179 training complete
Epoch 2180 training complete
Epoch 2181 training complete
Epoch 2182 training complete
Epoch 2183 training complete
Epoch 2184 training complete
Epoch 2185 training complete
Epoch 2186 training complete
Epoch 2187 training complete
Epoch 2188 training complete
Epoch 2189 training complete
Epoch 2190 training complete
Epoch 2191 training complete
Epoch 2192 training complete
Epoch 2193 training complete
Epoch 2194 training complete
Epoch 2195 training complete
Epoch 2196 training complete
Epoch 2197 training complete
Epoch 2198 training complete
Epoch 2199 training complete
Epoch 2200 training complete
Epoch 2201 training complete
Epoch 2202 training complete
Epoch 2203 training complete
Epoch 2204 training complete
Epoch 2205 training complete
Epoch 2206 training complete
Epoch 2207 training complete
Epoch 2208 training complete
Epoch 2209 training complete
Epoch 2210 training complete
Epoch 2211 training complete
Epoch 2212 training complete
Epoch 2213 training complete
Epoch 2214 training complete
Epoch 2215 training complete
Epoch 2216 training complete
Epoch 2217 training complete
Epoch 2218 training complete
Epoch 2219 training complete
Epoch 2220 training complete
Epoch 2221 training complete
Epoch 2222 training complete
Epoch 2223 training complete
Epoch 2224 training complete
Epoch 2225 training complete
Epoch 2226 training complete
Epoch 2227 training complete
Epoch 2228 training complete
Epoch 2229 training complete
Epoch 2230 training complete
Epoch 2231 training complete
Epoch 2232 training complete
Epoch 2233 training complete
Epoch 2234 training complete
Epoch 2235 training complete
Epoch 2236 training complete
Epoch 2237 training complete
Epoch 2238 training complete
Epoch 2239 training complete
Epoch 2240 training complete
Epoch 2241 training complete
Epoch 2242 training complete
Epoch 2243 training complete
Epoch 2244 training complete
Epoch 2245 training complete
Epoch 2246 training complete
Epoch 2247 training complete
Epoch 2248 training complete
Epoch 2249 training complete
Epoch 2250 training complete
Epoch 2251 training complete
Epoch 2252 training complete
Epoch 2253 training complete
Epoch 2254 training complete
Epoch 2255 training complete
Epoch 2256 training complete
Epoch 2257 training complete
Epoch 2258 training complete
Epoch 2259 training complete
Epoch 2260 training complete
Epoch 2261 training complete
Epoch 2262 training complete
Epoch 2263 training complete
Epoch 2264 training complete
Epoch 2265 training complete
Epoch 2266 training complete
Epoch 2267 training complete
Epoch 2268 training complete
Epoch 2269 training complete
Epoch 2270 training complete
Epoch 2271 training complete
Epoch 2272 training complete
Epoch 2273 training complete
Epoch 2274 training complete
Epoch 2275 training complete
Epoch 2276 training complete
Epoch 2277 training complete
Epoch 2278 training complete
Epoch 2279 training complete
Epoch 2280 training complete
Epoch 2281 training complete
Epoch 2282 training complete
Epoch 2283 training complete
Epoch 2284 training complete
Epoch 2285 training complete
Epoch 2286 training complete
Epoch 2287 training complete
Epoch 2288 training complete
Epoch 2289 training complete
Epoch 2290 training complete
Epoch 2291 training complete
Epoch 2292 training complete
Epoch 2293 training complete
Epoch 2294 training complete
Epoch 2295 training complete
Epoch 2296 training complete
Epoch 2297 training complete
Epoch 2298 training complete
Epoch 2299 training complete
Epoch 2300 training complete
Epoch 2301 training complete
Epoch 2302 training complete
Epoch 2303 training complete
Epoch 2304 training complete
Epoch 2305 training complete
Epoch 2306 training complete
Epoch 2307 training complete
Epoch 2308 training complete
Epoch 2309 training complete
Epoch 2310 training complete
Epoch 2311 training complete
Epoch 2312 training complete
Epoch 2313 training complete
Epoch 2314 training complete
Epoch 2315 training complete
Epoch 2316 training complete
Epoch 2317 training complete
Epoch 2318 training complete
Epoch 2319 training complete
Epoch 2320 training complete
Epoch 2321 training complete
Epoch 2322 training complete
Epoch 2323 training complete
Epoch 2324 training complete
Epoch 2325 training complete
Epoch 2326 training complete
Epoch 2327 training complete
Epoch 2328 training complete
Epoch 2329 training complete
Epoch 2330 training complete
Epoch 2331 training complete
Epoch 2332 training complete
Epoch 2333 training complete
Epoch 2334 training complete
Epoch 2335 training complete
Epoch 2336 training complete
Epoch 2337 training complete
Epoch 2338 training complete
Epoch 2339 training complete
Epoch 2340 training complete
Epoch 2341 training complete
Epoch 2342 training complete
Epoch 2343 training complete
Epoch 2344 training complete
Epoch 2345 training complete
Epoch 2346 training complete
Epoch 2347 training complete
Epoch 2348 training complete
Epoch 2349 training complete
Epoch 2350 training complete
Epoch 2351 training complete
Epoch 2352 training complete
Epoch 2353 training complete
Epoch 2354 training complete
Epoch 2355 training complete
Epoch 2356 training complete
Epoch 2357 training complete
Epoch 2358 training complete
Epoch 2359 training complete
Epoch 2360 training complete
Epoch 2361 training complete
Epoch 2362 training complete
Epoch 2363 training complete
Epoch 2364 training complete
Epoch 2365 training complete
Epoch 2366 training complete
Epoch 2367 training complete
Epoch 2368 training complete
Epoch 2369 training complete
Epoch 2370 training complete
Epoch 2371 training complete
Epoch 2372 training complete
Epoch 2373 training complete
Epoch 2374 training complete
Epoch 2375 training complete
Epoch 2376 training complete
Epoch 2377 training complete
Epoch 2378 training complete
Epoch 2379 training complete
Epoch 2380 training complete
Epoch 2381 training complete
Epoch 2382 training complete
Epoch 2383 training complete
Epoch 2384 training complete
Epoch 2385 training complete
Epoch 2386 training complete
Epoch 2387 training complete
Epoch 2388 training complete
Epoch 2389 training complete
Epoch 2390 training complete
Epoch 2391 training complete
Epoch 2392 training complete
Epoch 2393 training complete
Epoch 2394 training complete
Epoch 2395 training complete
Epoch 2396 training complete
Epoch 2397 training complete
Epoch 2398 training complete
Epoch 2399 training complete
Epoch 2400 training complete
Epoch 2401 training complete
Epoch 2402 training complete
Epoch 2403 training complete
Epoch 2404 training complete
Epoch 2405 training complete
Epoch 2406 training complete
Epoch 2407 training complete
Epoch 2408 training complete
Epoch 2409 training complete
Epoch 2410 training complete
Epoch 2411 training complete
Epoch 2412 training complete
Epoch 2413 training complete
Epoch 2414 training complete
Epoch 2415 training complete
Epoch 2416 training complete
Epoch 2417 training complete
Epoch 2418 training complete
Epoch 2419 training complete
Epoch 2420 training complete
Epoch 2421 training complete
Epoch 2422 training complete
Epoch 2423 training complete
Epoch 2424 training complete
Epoch 2425 training complete
Epoch 2426 training complete
Epoch 2427 training complete
Epoch 2428 training complete
Epoch 2429 training complete
Epoch 2430 training complete
Epoch 2431 training complete
Epoch 2432 training complete
Epoch 2433 training complete
Epoch 2434 training complete
Epoch 2435 training complete
Epoch 2436 training complete
Epoch 2437 training complete
Epoch 2438 training complete
Epoch 2439 training complete
Epoch 2440 training complete
Epoch 2441 training complete
Epoch 2442 training complete
Epoch 2443 training complete
Epoch 2444 training complete
Epoch 2445 training complete
Epoch 2446 training complete
Epoch 2447 training complete
Epoch 2448 training complete
Epoch 2449 training complete
Epoch 2450 training complete
Epoch 2451 training complete
Epoch 2452 training complete
Epoch 2453 training complete
Epoch 2454 training complete
Epoch 2455 training complete
Epoch 2456 training complete
Epoch 2457 training complete
Epoch 2458 training complete
Epoch 2459 training complete
Epoch 2460 training complete
Epoch 2461 training complete
Epoch 2462 training complete
Epoch 2463 training complete
Epoch 2464 training complete
Epoch 2465 training complete
Epoch 2466 training complete
Epoch 2467 training complete
Epoch 2468 training complete
Epoch 2469 training complete
Epoch 2470 training complete
Epoch 2471 training complete
Epoch 2472 training complete
Epoch 2473 training complete
Epoch 2474 training complete
Epoch 2475 training complete
Epoch 2476 training complete
Epoch 2477 training complete
Epoch 2478 training complete
Epoch 2479 training complete
Epoch 2480 training complete
Epoch 2481 training complete
Epoch 2482 training complete
Epoch 2483 training complete
Epoch 2484 training complete
Epoch 2485 training complete
Epoch 2486 training complete
Epoch 2487 training complete
Epoch 2488 training complete
Epoch 2489 training complete
Epoch 2490 training complete
Epoch 2491 training complete
Epoch 2492 training complete
Epoch 2493 training complete
Epoch 2494 training complete
Epoch 2495 training complete
Epoch 2496 training complete
Epoch 2497 training complete
Epoch 2498 training complete
Epoch 2499 training complete
Epoch 2500 training complete
Epoch 2501 training complete
Epoch 2502 training complete
Epoch 2503 training complete
Epoch 2504 training complete
Epoch 2505 training complete
Epoch 2506 training complete
Epoch 2507 training complete
Epoch 2508 training complete
Epoch 2509 training complete
Epoch 2510 training complete
Epoch 2511 training complete
Epoch 2512 training complete
Epoch 2513 training complete
Epoch 2514 training complete
Epoch 2515 training complete
Epoch 2516 training complete
Epoch 2517 training complete
Epoch 2518 training complete
Epoch 2519 training complete
Epoch 2520 training complete
Epoch 2521 training complete
Epoch 2522 training complete
Epoch 2523 training complete
Epoch 2524 training complete
Epoch 2525 training complete
Epoch 2526 training complete
Epoch 2527 training complete
Epoch 2528 training complete
Epoch 2529 training complete
Epoch 2530 training complete
Epoch 2531 training complete
Epoch 2532 training complete
Epoch 2533 training complete
Epoch 2534 training complete
Epoch 2535 training complete
Epoch 2536 training complete
Epoch 2537 training complete
Epoch 2538 training complete
Epoch 2539 training complete
Epoch 2540 training complete
Epoch 2541 training complete
Epoch 2542 training complete
Epoch 2543 training complete
Epoch 2544 training complete
Epoch 2545 training complete
Epoch 2546 training complete
Epoch 2547 training complete
Epoch 2548 training complete
Epoch 2549 training complete
Epoch 2550 training complete
Epoch 2551 training complete
Epoch 2552 training complete
Epoch 2553 training complete
Epoch 2554 training complete
Epoch 2555 training complete
Epoch 2556 training complete
Epoch 2557 training complete
Epoch 2558 training complete
Epoch 2559 training complete
Epoch 2560 training complete
Epoch 2561 training complete
Epoch 2562 training complete
Epoch 2563 training complete
Epoch 2564 training complete
Epoch 2565 training complete
Epoch 2566 training complete
Epoch 2567 training complete
Epoch 2568 training complete
Epoch 2569 training complete
Epoch 2570 training complete
Epoch 2571 training complete
Epoch 2572 training complete
Epoch 2573 training complete
Epoch 2574 training complete
Epoch 2575 training complete
Epoch 2576 training complete
Epoch 2577 training complete
Epoch 2578 training complete
Epoch 2579 training complete
Epoch 2580 training complete
Epoch 2581 training complete
Epoch 2582 training complete
Epoch 2583 training complete
Epoch 2584 training complete
Epoch 2585 training complete
Epoch 2586 training complete
Epoch 2587 training complete
Epoch 2588 training complete
Epoch 2589 training complete
Epoch 2590 training complete
Epoch 2591 training complete
Epoch 2592 training complete
Epoch 2593 training complete
Epoch 2594 training complete
Epoch 2595 training complete
Epoch 2596 training complete
Epoch 2597 training complete
Epoch 2598 training complete
Epoch 2599 training complete
Epoch 2600 training complete
Epoch 2601 training complete
Epoch 2602 training complete
Epoch 2603 training complete
Epoch 2604 training complete
Epoch 2605 training complete
Epoch 2606 training complete
Epoch 2607 training complete
Epoch 2608 training complete
Epoch 2609 training complete
Epoch 2610 training complete
Epoch 2611 training complete
Epoch 2612 training complete
Epoch 2613 training complete
Epoch 2614 training complete
Epoch 2615 training complete
Epoch 2616 training complete
Epoch 2617 training complete
Epoch 2618 training complete
Epoch 2619 training complete
Epoch 2620 training complete
Epoch 2621 training complete
Epoch 2622 training complete
Epoch 2623 training complete
Epoch 2624 training complete
Epoch 2625 training complete
Epoch 2626 training complete
Epoch 2627 training complete
Epoch 2628 training complete
Epoch 2629 training complete
Epoch 2630 training complete
Epoch 2631 training complete
Epoch 2632 training complete
Epoch 2633 training complete
Epoch 2634 training complete
Epoch 2635 training complete
Epoch 2636 training complete
Epoch 2637 training complete
Epoch 2638 training complete
Epoch 2639 training complete
Epoch 2640 training complete
Epoch 2641 training complete
Epoch 2642 training complete
Epoch 2643 training complete
Epoch 2644 training complete
Epoch 2645 training complete
Epoch 2646 training complete
Epoch 2647 training complete
Epoch 2648 training complete
Epoch 2649 training complete
Epoch 2650 training complete
Epoch 2651 training complete
Epoch 2652 training complete
Epoch 2653 training complete
Epoch 2654 training complete
Epoch 2655 training complete
Epoch 2656 training complete
Epoch 2657 training complete
Epoch 2658 training complete
Epoch 2659 training complete
Epoch 2660 training complete
Epoch 2661 training complete
Epoch 2662 training complete
Epoch 2663 training complete
Epoch 2664 training complete
Epoch 2665 training complete
Epoch 2666 training complete
Epoch 2667 training complete
Epoch 2668 training complete
Epoch 2669 training complete
Epoch 2670 training complete
Epoch 2671 training complete
Epoch 2672 training complete
Epoch 2673 training complete
Epoch 2674 training complete
Epoch 2675 training complete
Epoch 2676 training complete
Epoch 2677 training complete
Epoch 2678 training complete
Epoch 2679 training complete
Epoch 2680 training complete
Epoch 2681 training complete
Epoch 2682 training complete
Epoch 2683 training complete
Epoch 2684 training complete
Epoch 2685 training complete
Epoch 2686 training complete
Epoch 2687 training complete
Epoch 2688 training complete
Epoch 2689 training complete
Epoch 2690 training complete
Epoch 2691 training complete
Epoch 2692 training complete
Epoch 2693 training complete
Epoch 2694 training complete
Epoch 2695 training complete
Epoch 2696 training complete
Epoch 2697 training complete
Epoch 2698 training complete
Epoch 2699 training complete
Epoch 2700 training complete
Epoch 2701 training complete
Epoch 2702 training complete
Epoch 2703 training complete
Epoch 2704 training complete
Epoch 2705 training complete
Epoch 2706 training complete
Epoch 2707 training complete
Epoch 2708 training complete
Epoch 2709 training complete
Epoch 2710 training complete
Epoch 2711 training complete
Epoch 2712 training complete
Epoch 2713 training complete
Epoch 2714 training complete
Epoch 2715 training complete
Epoch 2716 training complete
Epoch 2717 training complete
Epoch 2718 training complete
Epoch 2719 training complete
Epoch 2720 training complete
Epoch 2721 training complete
Epoch 2722 training complete
Epoch 2723 training complete
Epoch 2724 training complete
Epoch 2725 training complete
Epoch 2726 training complete
Epoch 2727 training complete
Epoch 2728 training complete
Epoch 2729 training complete
Epoch 2730 training complete
Epoch 2731 training complete
Epoch 2732 training complete
Epoch 2733 training complete
Epoch 2734 training complete
Epoch 2735 training complete
Epoch 2736 training complete
Epoch 2737 training complete
Epoch 2738 training complete
Epoch 2739 training complete
Epoch 2740 training complete
Epoch 2741 training complete
Epoch 2742 training complete
Epoch 2743 training complete
Epoch 2744 training complete
Epoch 2745 training complete
Epoch 2746 training complete
Epoch 2747 training complete
Epoch 2748 training complete
Epoch 2749 training complete
Epoch 2750 training complete
Epoch 2751 training complete
Epoch 2752 training complete
Epoch 2753 training complete
Epoch 2754 training complete
Epoch 2755 training complete
Epoch 2756 training complete
Epoch 2757 training complete
Epoch 2758 training complete
Epoch 2759 training complete
Epoch 2760 training complete
Epoch 2761 training complete
Epoch 2762 training complete
Epoch 2763 training complete
Epoch 2764 training complete
Epoch 2765 training complete
Epoch 2766 training complete
Epoch 2767 training complete
Epoch 2768 training complete
Epoch 2769 training complete
Epoch 2770 training complete
Epoch 2771 training complete
Epoch 2772 training complete
Epoch 2773 training complete
Epoch 2774 training complete
Epoch 2775 training complete
Epoch 2776 training complete
Epoch 2777 training complete
Epoch 2778 training complete
Epoch 2779 training complete
Epoch 2780 training complete
Epoch 2781 training complete
Epoch 2782 training complete
Epoch 2783 training complete
Epoch 2784 training complete
Epoch 2785 training complete
Epoch 2786 training complete
Epoch 2787 training complete
Epoch 2788 training complete
Epoch 2789 training complete
Epoch 2790 training complete
Epoch 2791 training complete
Epoch 2792 training complete
Epoch 2793 training complete
Epoch 2794 training complete
Epoch 2795 training complete
Epoch 2796 training complete
Epoch 2797 training complete
Epoch 2798 training complete
Epoch 2799 training complete
Epoch 2800 training complete
Epoch 2801 training complete
Epoch 2802 training complete
Epoch 2803 training complete
Epoch 2804 training complete
Epoch 2805 training complete
Epoch 2806 training complete
Epoch 2807 training complete
Epoch 2808 training complete
Epoch 2809 training complete
Epoch 2810 training complete
Epoch 2811 training complete
Epoch 2812 training complete
Epoch 2813 training complete
Epoch 2814 training complete
Epoch 2815 training complete
Epoch 2816 training complete
Epoch 2817 training complete
Epoch 2818 training complete
Epoch 2819 training complete
Epoch 2820 training complete
Epoch 2821 training complete
Epoch 2822 training complete
Epoch 2823 training complete
Epoch 2824 training complete
Epoch 2825 training complete
Epoch 2826 training complete
Epoch 2827 training complete
Epoch 2828 training complete
Epoch 2829 training complete
Epoch 2830 training complete
Epoch 2831 training complete
Epoch 2832 training complete
Epoch 2833 training complete
Epoch 2834 training complete
Epoch 2835 training complete
Epoch 2836 training complete
Epoch 2837 training complete
Epoch 2838 training complete
Epoch 2839 training complete
Epoch 2840 training complete
Epoch 2841 training complete
Epoch 2842 training complete
Epoch 2843 training complete
Epoch 2844 training complete
Epoch 2845 training complete
Epoch 2846 training complete
Epoch 2847 training complete
Epoch 2848 training complete
Epoch 2849 training complete
Epoch 2850 training complete
Epoch 2851 training complete
Epoch 2852 training complete
Epoch 2853 training complete
Epoch 2854 training complete
Epoch 2855 training complete
Epoch 2856 training complete
Epoch 2857 training complete
Epoch 2858 training complete
Epoch 2859 training complete
Epoch 2860 training complete
Epoch 2861 training complete
Epoch 2862 training complete
Epoch 2863 training complete
Epoch 2864 training complete
Epoch 2865 training complete
Epoch 2866 training complete
Epoch 2867 training complete
Epoch 2868 training complete
Epoch 2869 training complete
Epoch 2870 training complete
Epoch 2871 training complete
Epoch 2872 training complete
Epoch 2873 training complete
Epoch 2874 training complete
Epoch 2875 training complete
Epoch 2876 training complete
Epoch 2877 training complete
Epoch 2878 training complete
Epoch 2879 training complete
Epoch 2880 training complete
Epoch 2881 training complete
Epoch 2882 training complete
Epoch 2883 training complete
Epoch 2884 training complete
Epoch 2885 training complete
Epoch 2886 training complete
Epoch 2887 training complete
Epoch 2888 training complete
Epoch 2889 training complete
Epoch 2890 training complete
Epoch 2891 training complete
Epoch 2892 training complete
Epoch 2893 training complete
Epoch 2894 training complete
Epoch 2895 training complete
Epoch 2896 training complete
Epoch 2897 training complete
Epoch 2898 training complete
Epoch 2899 training complete
Epoch 2900 training complete
Epoch 2901 training complete
Epoch 2902 training complete
Epoch 2903 training complete
Epoch 2904 training complete
Epoch 2905 training complete
Epoch 2906 training complete
Epoch 2907 training complete
Epoch 2908 training complete
Epoch 2909 training complete
Epoch 2910 training complete
Epoch 2911 training complete
Epoch 2912 training complete
Epoch 2913 training complete
Epoch 2914 training complete
Epoch 2915 training complete
Epoch 2916 training complete
Epoch 2917 training complete
Epoch 2918 training complete
Epoch 2919 training complete
Epoch 2920 training complete
Epoch 2921 training complete
Epoch 2922 training complete
Epoch 2923 training complete
Epoch 2924 training complete
Epoch 2925 training complete
Epoch 2926 training complete
Epoch 2927 training complete
Epoch 2928 training complete
Epoch 2929 training complete
Epoch 2930 training complete
Epoch 2931 training complete
Epoch 2932 training complete
Epoch 2933 training complete
Epoch 2934 training complete
Epoch 2935 training complete
Epoch 2936 training complete
Epoch 2937 training complete
Epoch 2938 training complete
Epoch 2939 training complete
Epoch 2940 training complete
Epoch 2941 training complete
Epoch 2942 training complete
Epoch 2943 training complete
Epoch 2944 training complete
Epoch 2945 training complete
Epoch 2946 training complete
Epoch 2947 training complete
Epoch 2948 training complete
Epoch 2949 training complete
Epoch 2950 training complete
Epoch 2951 training complete
Epoch 2952 training complete
Epoch 2953 training complete
Epoch 2954 training complete
Epoch 2955 training complete
Epoch 2956 training complete
Epoch 2957 training complete
Epoch 2958 training complete
Epoch 2959 training complete
Epoch 2960 training complete
Epoch 2961 training complete
Epoch 2962 training complete
Epoch 2963 training complete
Epoch 2964 training complete
Epoch 2965 training complete
Epoch 2966 training complete
Epoch 2967 training complete
Epoch 2968 training complete
Epoch 2969 training complete
Epoch 2970 training complete
Epoch 2971 training complete
Epoch 2972 training complete
Epoch 2973 training complete
Epoch 2974 training complete
Epoch 2975 training complete
Epoch 2976 training complete
Epoch 2977 training complete
Epoch 2978 training complete
Epoch 2979 training complete
Epoch 2980 training complete
Epoch 2981 training complete
Epoch 2982 training complete
Epoch 2983 training complete
Epoch 2984 training complete
Epoch 2985 training complete
Epoch 2986 training complete
Epoch 2987 training complete
Epoch 2988 training complete
Epoch 2989 training complete
Epoch 2990 training complete
Epoch 2991 training complete
Epoch 2992 training complete
Epoch 2993 training complete
Epoch 2994 training complete
Epoch 2995 training complete
Epoch 2996 training complete
Epoch 2997 training complete
Epoch 2998 training complete
Epoch 2999 training complete
Accuracy was 85.16 percent


Training network with data set size 1000
Epoch 0 training complete
Epoch 1 training complete
Epoch 2 training complete
Epoch 3 training complete
Epoch 4 training complete
Epoch 5 training complete
Epoch 6 training complete
Epoch 7 training complete
Epoch 8 training complete
Epoch 9 training complete
Epoch 10 training complete
Epoch 11 training complete
Epoch 12 training complete
Epoch 13 training complete
Epoch 14 training complete
Epoch 15 training complete
Epoch 16 training complete
Epoch 17 training complete
Epoch 18 training complete
Epoch 19 training complete
Epoch 20 training complete
Epoch 21 training complete
Epoch 22 training complete
Epoch 23 training complete
Epoch 24 training complete
Epoch 25 training complete
Epoch 26 training complete
Epoch 27 training complete
Epoch 28 training complete
Epoch 29 training complete
Epoch 30 training complete
Epoch 31 training complete
Epoch 32 training complete
Epoch 33 training complete
Epoch 34 training complete
Epoch 35 training complete
Epoch 36 training complete
Epoch 37 training complete
Epoch 38 training complete
Epoch 39 training complete
Epoch 40 training complete
Epoch 41 training complete
Epoch 42 training complete
Epoch 43 training complete
Epoch 44 training complete
Epoch 45 training complete
Epoch 46 training complete
Epoch 47 training complete
Epoch 48 training complete
Epoch 49 training complete
Epoch 50 training complete
Epoch 51 training complete
Epoch 52 training complete
Epoch 53 training complete
Epoch 54 training complete
Epoch 55 training complete
Epoch 56 training complete
Epoch 57 training complete
Epoch 58 training complete
Epoch 59 training complete
Epoch 60 training complete
Epoch 61 training complete
Epoch 62 training complete
Epoch 63 training complete
Epoch 64 training complete
Epoch 65 training complete
Epoch 66 training complete
Epoch 67 training complete
Epoch 68 training complete
Epoch 69 training complete
Epoch 70 training complete
Epoch 71 training complete
Epoch 72 training complete
Epoch 73 training complete
Epoch 74 training complete
Epoch 75 training complete
Epoch 76 training complete
Epoch 77 training complete
Epoch 78 training complete
Epoch 79 training complete
Epoch 80 training complete
Epoch 81 training complete
Epoch 82 training complete
Epoch 83 training complete
Epoch 84 training complete
Epoch 85 training complete
Epoch 86 training complete
Epoch 87 training complete
Epoch 88 training complete
Epoch 89 training complete
Epoch 90 training complete
Epoch 91 training complete
Epoch 92 training complete
Epoch 93 training complete
Epoch 94 training complete
Epoch 95 training complete
Epoch 96 training complete
Epoch 97 training complete
Epoch 98 training complete
Epoch 99 training complete
Epoch 100 training complete
Epoch 101 training complete
Epoch 102 training complete
Epoch 103 training complete
Epoch 104 training complete
Epoch 105 training complete
Epoch 106 training complete
Epoch 107 training complete
Epoch 108 training complete
Epoch 109 training complete
Epoch 110 training complete
Epoch 111 training complete
Epoch 112 training complete
Epoch 113 training complete
Epoch 114 training complete
Epoch 115 training complete
Epoch 116 training complete
Epoch 117 training complete
Epoch 118 training complete
Epoch 119 training complete
Epoch 120 training complete
Epoch 121 training complete
Epoch 122 training complete
Epoch 123 training complete
Epoch 124 training complete
Epoch 125 training complete
Epoch 126 training complete
Epoch 127 training complete
Epoch 128 training complete
Epoch 129 training complete
Epoch 130 training complete
Epoch 131 training complete
Epoch 132 training complete
Epoch 133 training complete
Epoch 134 training complete
Epoch 135 training complete
Epoch 136 training complete
Epoch 137 training complete
Epoch 138 training complete
Epoch 139 training complete
Epoch 140 training complete
Epoch 141 training complete
Epoch 142 training complete
Epoch 143 training complete
Epoch 144 training complete
Epoch 145 training complete
Epoch 146 training complete
Epoch 147 training complete
Epoch 148 training complete
Epoch 149 training complete
Epoch 150 training complete
Epoch 151 training complete
Epoch 152 training complete
Epoch 153 training complete
Epoch 154 training complete
Epoch 155 training complete
Epoch 156 training complete
Epoch 157 training complete
Epoch 158 training complete
Epoch 159 training complete
Epoch 160 training complete
Epoch 161 training complete
Epoch 162 training complete
Epoch 163 training complete
Epoch 164 training complete
Epoch 165 training complete
Epoch 166 training complete
Epoch 167 training complete
Epoch 168 training complete
Epoch 169 training complete
Epoch 170 training complete
Epoch 171 training complete
Epoch 172 training complete
Epoch 173 training complete
Epoch 174 training complete
Epoch 175 training complete
Epoch 176 training complete
Epoch 177 training complete
Epoch 178 training complete
Epoch 179 training complete
Epoch 180 training complete
Epoch 181 training complete
Epoch 182 training complete
Epoch 183 training complete
Epoch 184 training complete
Epoch 185 training complete
Epoch 186 training complete
Epoch 187 training complete
Epoch 188 training complete
Epoch 189 training complete
Epoch 190 training complete
Epoch 191 training complete
Epoch 192 training complete
Epoch 193 training complete
Epoch 194 training complete
Epoch 195 training complete
Epoch 196 training complete
Epoch 197 training complete
Epoch 198 training complete
Epoch 199 training complete
Epoch 200 training complete
Epoch 201 training complete
Epoch 202 training complete
Epoch 203 training complete
Epoch 204 training complete
Epoch 205 training complete
Epoch 206 training complete
Epoch 207 training complete
Epoch 208 training complete
Epoch 209 training complete
Epoch 210 training complete
Epoch 211 training complete
Epoch 212 training complete
Epoch 213 training complete
Epoch 214 training complete
Epoch 215 training complete
Epoch 216 training complete
Epoch 217 training complete
Epoch 218 training complete
Epoch 219 training complete
Epoch 220 training complete
Epoch 221 training complete
Epoch 222 training complete
Epoch 223 training complete
Epoch 224 training complete
Epoch 225 training complete
Epoch 226 training complete
Epoch 227 training complete
Epoch 228 training complete
Epoch 229 training complete
Epoch 230 training complete
Epoch 231 training complete
Epoch 232 training complete
Epoch 233 training complete
Epoch 234 training complete
Epoch 235 training complete
Epoch 236 training complete
Epoch 237 training complete
Epoch 238 training complete
Epoch 239 training complete
Epoch 240 training complete
Epoch 241 training complete
Epoch 242 training complete
Epoch 243 training complete
Epoch 244 training complete
Epoch 245 training complete
Epoch 246 training complete
Epoch 247 training complete
Epoch 248 training complete
Epoch 249 training complete
Epoch 250 training complete
Epoch 251 training complete
Epoch 252 training complete
Epoch 253 training complete
Epoch 254 training complete
Epoch 255 training complete
Epoch 256 training complete
Epoch 257 training complete
Epoch 258 training complete
Epoch 259 training complete
Epoch 260 training complete
Epoch 261 training complete
Epoch 262 training complete
Epoch 263 training complete
Epoch 264 training complete
Epoch 265 training complete
Epoch 266 training complete
Epoch 267 training complete
Epoch 268 training complete
Epoch 269 training complete
Epoch 270 training complete
Epoch 271 training complete
Epoch 272 training complete
Epoch 273 training complete
Epoch 274 training complete
Epoch 275 training complete
Epoch 276 training complete
Epoch 277 training complete
Epoch 278 training complete
Epoch 279 training complete
Epoch 280 training complete
Epoch 281 training complete
Epoch 282 training complete
Epoch 283 training complete
Epoch 284 training complete
Epoch 285 training complete
Epoch 286 training complete
Epoch 287 training complete
Epoch 288 training complete
Epoch 289 training complete
Epoch 290 training complete
Epoch 291 training complete
Epoch 292 training complete
Epoch 293 training complete
Epoch 294 training complete
Epoch 295 training complete
Epoch 296 training complete
Epoch 297 training complete
Epoch 298 training complete
Epoch 299 training complete
Epoch 300 training complete
Epoch 301 training complete
Epoch 302 training complete
Epoch 303 training complete
Epoch 304 training complete
Epoch 305 training complete
Epoch 306 training complete
Epoch 307 training complete
Epoch 308 training complete
Epoch 309 training complete
Epoch 310 training complete
Epoch 311 training complete
Epoch 312 training complete
Epoch 313 training complete
Epoch 314 training complete
Epoch 315 training complete
Epoch 316 training complete
Epoch 317 training complete
Epoch 318 training complete
Epoch 319 training complete
Epoch 320 training complete
Epoch 321 training complete
Epoch 322 training complete
Epoch 323 training complete
Epoch 324 training complete
Epoch 325 training complete
Epoch 326 training complete
Epoch 327 training complete
Epoch 328 training complete
Epoch 329 training complete
Epoch 330 training complete
Epoch 331 training complete
Epoch 332 training complete
Epoch 333 training complete
Epoch 334 training complete
Epoch 335 training complete
Epoch 336 training complete
Epoch 337 training complete
Epoch 338 training complete
Epoch 339 training complete
Epoch 340 training complete
Epoch 341 training complete
Epoch 342 training complete
Epoch 343 training complete
Epoch 344 training complete
Epoch 345 training complete
Epoch 346 training complete
Epoch 347 training complete
Epoch 348 training complete
Epoch 349 training complete
Epoch 350 training complete
Epoch 351 training complete
Epoch 352 training complete
Epoch 353 training complete
Epoch 354 training complete
Epoch 355 training complete
Epoch 356 training complete
Epoch 357 training complete
Epoch 358 training complete
Epoch 359 training complete
Epoch 360 training complete
Epoch 361 training complete
Epoch 362 training complete
Epoch 363 training complete
Epoch 364 training complete
Epoch 365 training complete
Epoch 366 training complete
Epoch 367 training complete
Epoch 368 training complete
Epoch 369 training complete
Epoch 370 training complete
Epoch 371 training complete
Epoch 372 training complete
Epoch 373 training complete
Epoch 374 training complete
Epoch 375 training complete
Epoch 376 training complete
Epoch 377 training complete
Epoch 378 training complete
Epoch 379 training complete
Epoch 380 training complete
Epoch 381 training complete
Epoch 382 training complete
Epoch 383 training complete
Epoch 384 training complete
Epoch 385 training complete
Epoch 386 training complete
Epoch 387 training complete
Epoch 388 training complete
Epoch 389 training complete
Epoch 390 training complete
Epoch 391 training complete
Epoch 392 training complete
Epoch 393 training complete
Epoch 394 training complete
Epoch 395 training complete
Epoch 396 training complete
Epoch 397 training complete
Epoch 398 training complete
Epoch 399 training complete
Epoch 400 training complete
Epoch 401 training complete
Epoch 402 training complete
Epoch 403 training complete
Epoch 404 training complete
Epoch 405 training complete
Epoch 406 training complete
Epoch 407 training complete
Epoch 408 training complete
Epoch 409 training complete
Epoch 410 training complete
Epoch 411 training complete
Epoch 412 training complete
Epoch 413 training complete
Epoch 414 training complete
Epoch 415 training complete
Epoch 416 training complete
Epoch 417 training complete
Epoch 418 training complete
Epoch 419 training complete
Epoch 420 training complete
Epoch 421 training complete
Epoch 422 training complete
Epoch 423 training complete
Epoch 424 training complete
Epoch 425 training complete
Epoch 426 training complete
Epoch 427 training complete
Epoch 428 training complete
Epoch 429 training complete
Epoch 430 training complete
Epoch 431 training complete
Epoch 432 training complete
Epoch 433 training complete
Epoch 434 training complete
Epoch 435 training complete
Epoch 436 training complete
Epoch 437 training complete
Epoch 438 training complete
Epoch 439 training complete
Epoch 440 training complete
Epoch 441 training complete
Epoch 442 training complete
Epoch 443 training complete
Epoch 444 training complete
Epoch 445 training complete
Epoch 446 training complete
Epoch 447 training complete
Epoch 448 training complete
Epoch 449 training complete
Epoch 450 training complete
Epoch 451 training complete
Epoch 452 training complete
Epoch 453 training complete
Epoch 454 training complete
Epoch 455 training complete
Epoch 456 training complete
Epoch 457 training complete
Epoch 458 training complete
Epoch 459 training complete
Epoch 460 training complete
Epoch 461 training complete
Epoch 462 training complete
Epoch 463 training complete
Epoch 464 training complete
Epoch 465 training complete
Epoch 466 training complete
Epoch 467 training complete
Epoch 468 training complete
Epoch 469 training complete
Epoch 470 training complete
Epoch 471 training complete
Epoch 472 training complete
Epoch 473 training complete
Epoch 474 training complete
Epoch 475 training complete
Epoch 476 training complete
Epoch 477 training complete
Epoch 478 training complete
Epoch 479 training complete
Epoch 480 training complete
Epoch 481 training complete
Epoch 482 training complete
Epoch 483 training complete
Epoch 484 training complete
Epoch 485 training complete
Epoch 486 training complete
Epoch 487 training complete
Epoch 488 training complete
Epoch 489 training complete
Epoch 490 training complete
Epoch 491 training complete
Epoch 492 training complete
Epoch 493 training complete
Epoch 494 training complete
Epoch 495 training complete
Epoch 496 training complete
Epoch 497 training complete
Epoch 498 training complete
Epoch 499 training complete
Epoch 500 training complete
Epoch 501 training complete
Epoch 502 training complete
Epoch 503 training complete
Epoch 504 training complete
Epoch 505 training complete
Epoch 506 training complete
Epoch 507 training complete
Epoch 508 training complete
Epoch 509 training complete
Epoch 510 training complete
Epoch 511 training complete
Epoch 512 training complete
Epoch 513 training complete
Epoch 514 training complete
Epoch 515 training complete
Epoch 516 training complete
Epoch 517 training complete
Epoch 518 training complete
Epoch 519 training complete
Epoch 520 training complete
Epoch 521 training complete
Epoch 522 training complete
Epoch 523 training complete
Epoch 524 training complete
Epoch 525 training complete
Epoch 526 training complete
Epoch 527 training complete
Epoch 528 training complete
Epoch 529 training complete
Epoch 530 training complete
Epoch 531 training complete
Epoch 532 training complete
Epoch 533 training complete
Epoch 534 training complete
Epoch 535 training complete
Epoch 536 training complete
Epoch 537 training complete
Epoch 538 training complete
Epoch 539 training complete
Epoch 540 training complete
Epoch 541 training complete
Epoch 542 training complete
Epoch 543 training complete
Epoch 544 training complete
Epoch 545 training complete
Epoch 546 training complete
Epoch 547 training complete
Epoch 548 training complete
Epoch 549 training complete
Epoch 550 training complete
Epoch 551 training complete
Epoch 552 training complete
Epoch 553 training complete
Epoch 554 training complete
Epoch 555 training complete
Epoch 556 training complete
Epoch 557 training complete
Epoch 558 training complete
Epoch 559 training complete
Epoch 560 training complete
Epoch 561 training complete
Epoch 562 training complete
Epoch 563 training complete
Epoch 564 training complete
Epoch 565 training complete
Epoch 566 training complete
Epoch 567 training complete
Epoch 568 training complete
Epoch 569 training complete
Epoch 570 training complete
Epoch 571 training complete
Epoch 572 training complete
Epoch 573 training complete
Epoch 574 training complete
Epoch 575 training complete
Epoch 576 training complete
Epoch 577 training complete
Epoch 578 training complete
Epoch 579 training complete
Epoch 580 training complete
Epoch 581 training complete
Epoch 582 training complete
Epoch 583 training complete
Epoch 584 training complete
Epoch 585 training complete
Epoch 586 training complete
Epoch 587 training complete
Epoch 588 training complete
Epoch 589 training complete
Epoch 590 training complete
Epoch 591 training complete
Epoch 592 training complete
Epoch 593 training complete
Epoch 594 training complete
Epoch 595 training complete
Epoch 596 training complete
Epoch 597 training complete
Epoch 598 training complete
Epoch 599 training complete
Epoch 600 training complete
Epoch 601 training complete
Epoch 602 training complete
Epoch 603 training complete
Epoch 604 training complete
Epoch 605 training complete
Epoch 606 training complete
Epoch 607 training complete
Epoch 608 training complete
Epoch 609 training complete
Epoch 610 training complete
Epoch 611 training complete
Epoch 612 training complete
Epoch 613 training complete
Epoch 614 training complete
Epoch 615 training complete
Epoch 616 training complete
Epoch 617 training complete
Epoch 618 training complete
Epoch 619 training complete
Epoch 620 training complete
Epoch 621 training complete
Epoch 622 training complete
Epoch 623 training complete
Epoch 624 training complete
Epoch 625 training complete
Epoch 626 training complete
Epoch 627 training complete
Epoch 628 training complete
Epoch 629 training complete
Epoch 630 training complete
Epoch 631 training complete
Epoch 632 training complete
Epoch 633 training complete
Epoch 634 training complete
Epoch 635 training complete
Epoch 636 training complete
Epoch 637 training complete
Epoch 638 training complete
Epoch 639 training complete
Epoch 640 training complete
Epoch 641 training complete
Epoch 642 training complete
Epoch 643 training complete
Epoch 644 training complete
Epoch 645 training complete
Epoch 646 training complete
Epoch 647 training complete
Epoch 648 training complete
Epoch 649 training complete
Epoch 650 training complete
Epoch 651 training complete
Epoch 652 training complete
Epoch 653 training complete
Epoch 654 training complete
Epoch 655 training complete
Epoch 656 training complete
Epoch 657 training complete
Epoch 658 training complete
Epoch 659 training complete
Epoch 660 training complete
Epoch 661 training complete
Epoch 662 training complete
Epoch 663 training complete
Epoch 664 training complete
Epoch 665 training complete
Epoch 666 training complete
Epoch 667 training complete
Epoch 668 training complete
Epoch 669 training complete
Epoch 670 training complete
Epoch 671 training complete
Epoch 672 training complete
Epoch 673 training complete
Epoch 674 training complete
Epoch 675 training complete
Epoch 676 training complete
Epoch 677 training complete
Epoch 678 training complete
Epoch 679 training complete
Epoch 680 training complete
Epoch 681 training complete
Epoch 682 training complete
Epoch 683 training complete
Epoch 684 training complete
Epoch 685 training complete
Epoch 686 training complete
Epoch 687 training complete
Epoch 688 training complete
Epoch 689 training complete
Epoch 690 training complete
Epoch 691 training complete
Epoch 692 training complete
Epoch 693 training complete
Epoch 694 training complete
Epoch 695 training complete
Epoch 696 training complete
Epoch 697 training complete
Epoch 698 training complete
Epoch 699 training complete
Epoch 700 training complete
Epoch 701 training complete
Epoch 702 training complete
Epoch 703 training complete
Epoch 704 training complete
Epoch 705 training complete
Epoch 706 training complete
Epoch 707 training complete
Epoch 708 training complete
Epoch 709 training complete
Epoch 710 training complete
Epoch 711 training complete
Epoch 712 training complete
Epoch 713 training complete
Epoch 714 training complete
Epoch 715 training complete
Epoch 716 training complete
Epoch 717 training complete
Epoch 718 training complete
Epoch 719 training complete
Epoch 720 training complete
Epoch 721 training complete
Epoch 722 training complete
Epoch 723 training complete
Epoch 724 training complete
Epoch 725 training complete
Epoch 726 training complete
Epoch 727 training complete
Epoch 728 training complete
Epoch 729 training complete
Epoch 730 training complete
Epoch 731 training complete
Epoch 732 training complete
Epoch 733 training complete
Epoch 734 training complete
Epoch 735 training complete
Epoch 736 training complete
Epoch 737 training complete
Epoch 738 training complete
Epoch 739 training complete
Epoch 740 training complete
Epoch 741 training complete
Epoch 742 training complete
Epoch 743 training complete
Epoch 744 training complete
Epoch 745 training complete
Epoch 746 training complete
Epoch 747 training complete
Epoch 748 training complete
Epoch 749 training complete
Epoch 750 training complete
Epoch 751 training complete
Epoch 752 training complete
Epoch 753 training complete
Epoch 754 training complete
Epoch 755 training complete
Epoch 756 training complete
Epoch 757 training complete
Epoch 758 training complete
Epoch 759 training complete
Epoch 760 training complete
Epoch 761 training complete
Epoch 762 training complete
Epoch 763 training complete
Epoch 764 training complete
Epoch 765 training complete
Epoch 766 training complete
Epoch 767 training complete
Epoch 768 training complete
Epoch 769 training complete
Epoch 770 training complete
Epoch 771 training complete
Epoch 772 training complete
Epoch 773 training complete
Epoch 774 training complete
Epoch 775 training complete
Epoch 776 training complete
Epoch 777 training complete
Epoch 778 training complete
Epoch 779 training complete
Epoch 780 training complete
Epoch 781 training complete
Epoch 782 training complete
Epoch 783 training complete
Epoch 784 training complete
Epoch 785 training complete
Epoch 786 training complete
Epoch 787 training complete
Epoch 788 training complete
Epoch 789 training complete
Epoch 790 training complete
Epoch 791 training complete
Epoch 792 training complete
Epoch 793 training complete
Epoch 794 training complete
Epoch 795 training complete
Epoch 796 training complete
Epoch 797 training complete
Epoch 798 training complete
Epoch 799 training complete
Epoch 800 training complete
Epoch 801 training complete
Epoch 802 training complete
Epoch 803 training complete
Epoch 804 training complete
Epoch 805 training complete
Epoch 806 training complete
Epoch 807 training complete
Epoch 808 training complete
Epoch 809 training complete
Epoch 810 training complete
Epoch 811 training complete
Epoch 812 training complete
Epoch 813 training complete
Epoch 814 training complete
Epoch 815 training complete
Epoch 816 training complete
Epoch 817 training complete
Epoch 818 training complete
Epoch 819 training complete
Epoch 820 training complete
Epoch 821 training complete
Epoch 822 training complete
Epoch 823 training complete
Epoch 824 training complete
Epoch 825 training complete
Epoch 826 training complete
Epoch 827 training complete
Epoch 828 training complete
Epoch 829 training complete
Epoch 830 training complete
Epoch 831 training complete
Epoch 832 training complete
Epoch 833 training complete
Epoch 834 training complete
Epoch 835 training complete
Epoch 836 training complete
Epoch 837 training complete
Epoch 838 training complete
Epoch 839 training complete
Epoch 840 training complete
Epoch 841 training complete
Epoch 842 training complete
Epoch 843 training complete
Epoch 844 training complete
Epoch 845 training complete
Epoch 846 training complete
Epoch 847 training complete
Epoch 848 training complete
Epoch 849 training complete
Epoch 850 training complete
Epoch 851 training complete
Epoch 852 training complete
Epoch 853 training complete
Epoch 854 training complete
Epoch 855 training complete
Epoch 856 training complete
Epoch 857 training complete
Epoch 858 training complete
Epoch 859 training complete
Epoch 860 training complete
Epoch 861 training complete
Epoch 862 training complete
Epoch 863 training complete
Epoch 864 training complete
Epoch 865 training complete
Epoch 866 training complete
Epoch 867 training complete
Epoch 868 training complete
Epoch 869 training complete
Epoch 870 training complete
Epoch 871 training complete
Epoch 872 training complete
Epoch 873 training complete
Epoch 874 training complete
Epoch 875 training complete
Epoch 876 training complete
Epoch 877 training complete
Epoch 878 training complete
Epoch 879 training complete
Epoch 880 training complete
Epoch 881 training complete
Epoch 882 training complete
Epoch 883 training complete
Epoch 884 training complete
Epoch 885 training complete
Epoch 886 training complete
Epoch 887 training complete
Epoch 888 training complete
Epoch 889 training complete
Epoch 890 training complete
Epoch 891 training complete
Epoch 892 training complete
Epoch 893 training complete
Epoch 894 training complete
Epoch 895 training complete
Epoch 896 training complete
Epoch 897 training complete
Epoch 898 training complete
Epoch 899 training complete
Epoch 900 training complete
Epoch 901 training complete
Epoch 902 training complete
Epoch 903 training complete
Epoch 904 training complete
Epoch 905 training complete
Epoch 906 training complete
Epoch 907 training complete
Epoch 908 training complete
Epoch 909 training complete
Epoch 910 training complete
Epoch 911 training complete
Epoch 912 training complete
Epoch 913 training complete
Epoch 914 training complete
Epoch 915 training complete
Epoch 916 training complete
Epoch 917 training complete
Epoch 918 training complete
Epoch 919 training complete
Epoch 920 training complete
Epoch 921 training complete
Epoch 922 training complete
Epoch 923 training complete
Epoch 924 training complete
Epoch 925 training complete
Epoch 926 training complete
Epoch 927 training complete
Epoch 928 training complete
Epoch 929 training complete
Epoch 930 training complete
Epoch 931 training complete
Epoch 932 training complete
Epoch 933 training complete
Epoch 934 training complete
Epoch 935 training complete
Epoch 936 training complete
Epoch 937 training complete
Epoch 938 training complete
Epoch 939 training complete
Epoch 940 training complete
Epoch 941 training complete
Epoch 942 training complete
Epoch 943 training complete
Epoch 944 training complete
Epoch 945 training complete
Epoch 946 training complete
Epoch 947 training complete
Epoch 948 training complete
Epoch 949 training complete
Epoch 950 training complete
Epoch 951 training complete
Epoch 952 training complete
Epoch 953 training complete
Epoch 954 training complete
Epoch 955 training complete
Epoch 956 training complete
Epoch 957 training complete
Epoch 958 training complete
Epoch 959 training complete
Epoch 960 training complete
Epoch 961 training complete
Epoch 962 training complete
Epoch 963 training complete
Epoch 964 training complete
Epoch 965 training complete
Epoch 966 training complete
Epoch 967 training complete
Epoch 968 training complete
Epoch 969 training complete
Epoch 970 training complete
Epoch 971 training complete
Epoch 972 training complete
Epoch 973 training complete
Epoch 974 training complete
Epoch 975 training complete
Epoch 976 training complete
Epoch 977 training complete
Epoch 978 training complete
Epoch 979 training complete
Epoch 980 training complete
Epoch 981 training complete
Epoch 982 training complete
Epoch 983 training complete
Epoch 984 training complete
Epoch 985 training complete
Epoch 986 training complete
Epoch 987 training complete
Epoch 988 training complete
Epoch 989 training complete
Epoch 990 training complete
Epoch 991 training complete
Epoch 992 training complete
Epoch 993 training complete
Epoch 994 training complete
Epoch 995 training complete
Epoch 996 training complete
Epoch 997 training complete
Epoch 998 training complete
Epoch 999 training complete
Epoch 1000 training complete
Epoch 1001 training complete
Epoch 1002 training complete
Epoch 1003 training complete
Epoch 1004 training complete
Epoch 1005 training complete
Epoch 1006 training complete
Epoch 1007 training complete
Epoch 1008 training complete
Epoch 1009 training complete
Epoch 1010 training complete
Epoch 1011 training complete
Epoch 1012 training complete
Epoch 1013 training complete
Epoch 1014 training complete
Epoch 1015 training complete
Epoch 1016 training complete
Epoch 1017 training complete
Epoch 1018 training complete
Epoch 1019 training complete
Epoch 1020 training complete
Epoch 1021 training complete
Epoch 1022 training complete
Epoch 1023 training complete
Epoch 1024 training complete
Epoch 1025 training complete
Epoch 1026 training complete
Epoch 1027 training complete
Epoch 1028 training complete
Epoch 1029 training complete
Epoch 1030 training complete
Epoch 1031 training complete
Epoch 1032 training complete
Epoch 1033 training complete
Epoch 1034 training complete
Epoch 1035 training complete
Epoch 1036 training complete
Epoch 1037 training complete
Epoch 1038 training complete
Epoch 1039 training complete
Epoch 1040 training complete
Epoch 1041 training complete
Epoch 1042 training complete
Epoch 1043 training complete
Epoch 1044 training complete
Epoch 1045 training complete
Epoch 1046 training complete
Epoch 1047 training complete
Epoch 1048 training complete
Epoch 1049 training complete
Epoch 1050 training complete
Epoch 1051 training complete
Epoch 1052 training complete
Epoch 1053 training complete
Epoch 1054 training complete
Epoch 1055 training complete
Epoch 1056 training complete
Epoch 1057 training complete
Epoch 1058 training complete
Epoch 1059 training complete
Epoch 1060 training complete
Epoch 1061 training complete
Epoch 1062 training complete
Epoch 1063 training complete
Epoch 1064 training complete
Epoch 1065 training complete
Epoch 1066 training complete
Epoch 1067 training complete
Epoch 1068 training complete
Epoch 1069 training complete
Epoch 1070 training complete
Epoch 1071 training complete
Epoch 1072 training complete
Epoch 1073 training complete
Epoch 1074 training complete
Epoch 1075 training complete
Epoch 1076 training complete
Epoch 1077 training complete
Epoch 1078 training complete
Epoch 1079 training complete
Epoch 1080 training complete
Epoch 1081 training complete
Epoch 1082 training complete
Epoch 1083 training complete
Epoch 1084 training complete
Epoch 1085 training complete
Epoch 1086 training complete
Epoch 1087 training complete
Epoch 1088 training complete
Epoch 1089 training complete
Epoch 1090 training complete
Epoch 1091 training complete
Epoch 1092 training complete
Epoch 1093 training complete
Epoch 1094 training complete
Epoch 1095 training complete
Epoch 1096 training complete
Epoch 1097 training complete
Epoch 1098 training complete
Epoch 1099 training complete
Epoch 1100 training complete
Epoch 1101 training complete
Epoch 1102 training complete
Epoch 1103 training complete
Epoch 1104 training complete
Epoch 1105 training complete
Epoch 1106 training complete
Epoch 1107 training complete
Epoch 1108 training complete
Epoch 1109 training complete
Epoch 1110 training complete
Epoch 1111 training complete
Epoch 1112 training complete
Epoch 1113 training complete
Epoch 1114 training complete
Epoch 1115 training complete
Epoch 1116 training complete
Epoch 1117 training complete
Epoch 1118 training complete
Epoch 1119 training complete
Epoch 1120 training complete
Epoch 1121 training complete
Epoch 1122 training complete
Epoch 1123 training complete
Epoch 1124 training complete
Epoch 1125 training complete
Epoch 1126 training complete
Epoch 1127 training complete
Epoch 1128 training complete
Epoch 1129 training complete
Epoch 1130 training complete
Epoch 1131 training complete
Epoch 1132 training complete
Epoch 1133 training complete
Epoch 1134 training complete
Epoch 1135 training complete
Epoch 1136 training complete
Epoch 1137 training complete
Epoch 1138 training complete
Epoch 1139 training complete
Epoch 1140 training complete
Epoch 1141 training complete
Epoch 1142 training complete
Epoch 1143 training complete
Epoch 1144 training complete
Epoch 1145 training complete
Epoch 1146 training complete
Epoch 1147 training complete
Epoch 1148 training complete
Epoch 1149 training complete
Epoch 1150 training complete
Epoch 1151 training complete
Epoch 1152 training complete
Epoch 1153 training complete
Epoch 1154 training complete
Epoch 1155 training complete
Epoch 1156 training complete
Epoch 1157 training complete
Epoch 1158 training complete
Epoch 1159 training complete
Epoch 1160 training complete
Epoch 1161 training complete
Epoch 1162 training complete
Epoch 1163 training complete
Epoch 1164 training complete
Epoch 1165 training complete
Epoch 1166 training complete
Epoch 1167 training complete
Epoch 1168 training complete
Epoch 1169 training complete
Epoch 1170 training complete
Epoch 1171 training complete
Epoch 1172 training complete
Epoch 1173 training complete
Epoch 1174 training complete
Epoch 1175 training complete
Epoch 1176 training complete
Epoch 1177 training complete
Epoch 1178 training complete
Epoch 1179 training complete
Epoch 1180 training complete
Epoch 1181 training complete
Epoch 1182 training complete
Epoch 1183 training complete
Epoch 1184 training complete
Epoch 1185 training complete
Epoch 1186 training complete
Epoch 1187 training complete
Epoch 1188 training complete
Epoch 1189 training complete
Epoch 1190 training complete
Epoch 1191 training complete
Epoch 1192 training complete
Epoch 1193 training complete
Epoch 1194 training complete
Epoch 1195 training complete
Epoch 1196 training complete
Epoch 1197 training complete
Epoch 1198 training complete
Epoch 1199 training complete
Epoch 1200 training complete
Epoch 1201 training complete
Epoch 1202 training complete
Epoch 1203 training complete
Epoch 1204 training complete
Epoch 1205 training complete
Epoch 1206 training complete
Epoch 1207 training complete
Epoch 1208 training complete
Epoch 1209 training complete
Epoch 1210 training complete
Epoch 1211 training complete
Epoch 1212 training complete
Epoch 1213 training complete
Epoch 1214 training complete
Epoch 1215 training complete
Epoch 1216 training complete
Epoch 1217 training complete
Epoch 1218 training complete
Epoch 1219 training complete
Epoch 1220 training complete
Epoch 1221 training complete
Epoch 1222 training complete
Epoch 1223 training complete
Epoch 1224 training complete
Epoch 1225 training complete
Epoch 1226 training complete
Epoch 1227 training complete
Epoch 1228 training complete
Epoch 1229 training complete
Epoch 1230 training complete
Epoch 1231 training complete
Epoch 1232 training complete
Epoch 1233 training complete
Epoch 1234 training complete
Epoch 1235 training complete
Epoch 1236 training complete
Epoch 1237 training complete
Epoch 1238 training complete
Epoch 1239 training complete
Epoch 1240 training complete
Epoch 1241 training complete
Epoch 1242 training complete
Epoch 1243 training complete
Epoch 1244 training complete
Epoch 1245 training complete
Epoch 1246 training complete
Epoch 1247 training complete
Epoch 1248 training complete
Epoch 1249 training complete
Epoch 1250 training complete
Epoch 1251 training complete
Epoch 1252 training complete
Epoch 1253 training complete
Epoch 1254 training complete
Epoch 1255 training complete
Epoch 1256 training complete
Epoch 1257 training complete
Epoch 1258 training complete
Epoch 1259 training complete
Epoch 1260 training complete
Epoch 1261 training complete
Epoch 1262 training complete
Epoch 1263 training complete
Epoch 1264 training complete
Epoch 1265 training complete
Epoch 1266 training complete
Epoch 1267 training complete
Epoch 1268 training complete
Epoch 1269 training complete
Epoch 1270 training complete
Epoch 1271 training complete
Epoch 1272 training complete
Epoch 1273 training complete
Epoch 1274 training complete
Epoch 1275 training complete
Epoch 1276 training complete
Epoch 1277 training complete
Epoch 1278 training complete
Epoch 1279 training complete
Epoch 1280 training complete
Epoch 1281 training complete
Epoch 1282 training complete
Epoch 1283 training complete
Epoch 1284 training complete
Epoch 1285 training complete
Epoch 1286 training complete
Epoch 1287 training complete
Epoch 1288 training complete
Epoch 1289 training complete
Epoch 1290 training complete
Epoch 1291 training complete
Epoch 1292 training complete
Epoch 1293 training complete
Epoch 1294 training complete
Epoch 1295 training complete
Epoch 1296 training complete
Epoch 1297 training complete
Epoch 1298 training complete
Epoch 1299 training complete
Epoch 1300 training complete
Epoch 1301 training complete
Epoch 1302 training complete
Epoch 1303 training complete
Epoch 1304 training complete
Epoch 1305 training complete
Epoch 1306 training complete
Epoch 1307 training complete
Epoch 1308 training complete
Epoch 1309 training complete
Epoch 1310 training complete
Epoch 1311 training complete
Epoch 1312 training complete
Epoch 1313 training complete
Epoch 1314 training complete
Epoch 1315 training complete
Epoch 1316 training complete
Epoch 1317 training complete
Epoch 1318 training complete
Epoch 1319 training complete
Epoch 1320 training complete
Epoch 1321 training complete
Epoch 1322 training complete
Epoch 1323 training complete
Epoch 1324 training complete
Epoch 1325 training complete
Epoch 1326 training complete
Epoch 1327 training complete
Epoch 1328 training complete
Epoch 1329 training complete
Epoch 1330 training complete
Epoch 1331 training complete
Epoch 1332 training complete
Epoch 1333 training complete
Epoch 1334 training complete
Epoch 1335 training complete
Epoch 1336 training complete
Epoch 1337 training complete
Epoch 1338 training complete
Epoch 1339 training complete
Epoch 1340 training complete
Epoch 1341 training complete
Epoch 1342 training complete
Epoch 1343 training complete
Epoch 1344 training complete
Epoch 1345 training complete
Epoch 1346 training complete
Epoch 1347 training complete
Epoch 1348 training complete
Epoch 1349 training complete
Epoch 1350 training complete
Epoch 1351 training complete
Epoch 1352 training complete
Epoch 1353 training complete
Epoch 1354 training complete
Epoch 1355 training complete
Epoch 1356 training complete
Epoch 1357 training complete
Epoch 1358 training complete
Epoch 1359 training complete
Epoch 1360 training complete
Epoch 1361 training complete
Epoch 1362 training complete
Epoch 1363 training complete
Epoch 1364 training complete
Epoch 1365 training complete
Epoch 1366 training complete
Epoch 1367 training complete
Epoch 1368 training complete
Epoch 1369 training complete
Epoch 1370 training complete
Epoch 1371 training complete
Epoch 1372 training complete
Epoch 1373 training complete
Epoch 1374 training complete
Epoch 1375 training complete
Epoch 1376 training complete
Epoch 1377 training complete
Epoch 1378 training complete
Epoch 1379 training complete
Epoch 1380 training complete
Epoch 1381 training complete
Epoch 1382 training complete
Epoch 1383 training complete
Epoch 1384 training complete
Epoch 1385 training complete
Epoch 1386 training complete
Epoch 1387 training complete
Epoch 1388 training complete
Epoch 1389 training complete
Epoch 1390 training complete
Epoch 1391 training complete
Epoch 1392 training complete
Epoch 1393 training complete
Epoch 1394 training complete
Epoch 1395 training complete
Epoch 1396 training complete
Epoch 1397 training complete
Epoch 1398 training complete
Epoch 1399 training complete
Epoch 1400 training complete
Epoch 1401 training complete
Epoch 1402 training complete
Epoch 1403 training complete
Epoch 1404 training complete
Epoch 1405 training complete
Epoch 1406 training complete
Epoch 1407 training complete
Epoch 1408 training complete
Epoch 1409 training complete
Epoch 1410 training complete
Epoch 1411 training complete
Epoch 1412 training complete
Epoch 1413 training complete
Epoch 1414 training complete
Epoch 1415 training complete
Epoch 1416 training complete
Epoch 1417 training complete
Epoch 1418 training complete
Epoch 1419 training complete
Epoch 1420 training complete
Epoch 1421 training complete
Epoch 1422 training complete
Epoch 1423 training complete
Epoch 1424 training complete
Epoch 1425 training complete
Epoch 1426 training complete
Epoch 1427 training complete
Epoch 1428 training complete
Epoch 1429 training complete
Epoch 1430 training complete
Epoch 1431 training complete
Epoch 1432 training complete
Epoch 1433 training complete
Epoch 1434 training complete
Epoch 1435 training complete
Epoch 1436 training complete
Epoch 1437 training complete
Epoch 1438 training complete
Epoch 1439 training complete
Epoch 1440 training complete
Epoch 1441 training complete
Epoch 1442 training complete
Epoch 1443 training complete
Epoch 1444 training complete
Epoch 1445 training complete
Epoch 1446 training complete
Epoch 1447 training complete
Epoch 1448 training complete
Epoch 1449 training complete
Epoch 1450 training complete
Epoch 1451 training complete
Epoch 1452 training complete
Epoch 1453 training complete
Epoch 1454 training complete
Epoch 1455 training complete
Epoch 1456 training complete
Epoch 1457 training complete
Epoch 1458 training complete
Epoch 1459 training complete
Epoch 1460 training complete
Epoch 1461 training complete
Epoch 1462 training complete
Epoch 1463 training complete
Epoch 1464 training complete
Epoch 1465 training complete
Epoch 1466 training complete
Epoch 1467 training complete
Epoch 1468 training complete
Epoch 1469 training complete
Epoch 1470 training complete
Epoch 1471 training complete
Epoch 1472 training complete
Epoch 1473 training complete
Epoch 1474 training complete
Epoch 1475 training complete
Epoch 1476 training complete
Epoch 1477 training complete
Epoch 1478 training complete
Epoch 1479 training complete
Epoch 1480 training complete
Epoch 1481 training complete
Epoch 1482 training complete
Epoch 1483 training complete
Epoch 1484 training complete
Epoch 1485 training complete
Epoch 1486 training complete
Epoch 1487 training complete
Epoch 1488 training complete
Epoch 1489 training complete
Epoch 1490 training complete
Epoch 1491 training complete
Epoch 1492 training complete
Epoch 1493 training complete
Epoch 1494 training complete
Epoch 1495 training complete
Epoch 1496 training complete
Epoch 1497 training complete
Epoch 1498 training complete
Epoch 1499 training complete
Accuracy was 89.05 percent


Training network with data set size 2000
Epoch 0 training complete
Epoch 1 training complete
Epoch 2 training complete
Epoch 3 training complete
Epoch 4 training complete
Epoch 5 training complete
Epoch 6 training complete
Epoch 7 training complete
Epoch 8 training complete
Epoch 9 training complete
Epoch 10 training complete
Epoch 11 training complete
Epoch 12 training complete
Epoch 13 training complete
Epoch 14 training complete
Epoch 15 training complete
Epoch 16 training complete
Epoch 17 training complete
Epoch 18 training complete
Epoch 19 training complete
Epoch 20 training complete
Epoch 21 training complete
Epoch 22 training complete
Epoch 23 training complete
Epoch 24 training complete
Epoch 25 training complete
Epoch 26 training complete
Epoch 27 training complete
Epoch 28 training complete
Epoch 29 training complete
Epoch 30 training complete
Epoch 31 training complete
Epoch 32 training complete
Epoch 33 training complete
Epoch 34 training complete
Epoch 35 training complete
Epoch 36 training complete
Epoch 37 training complete
Epoch 38 training complete
Epoch 39 training complete
Epoch 40 training complete
Epoch 41 training complete
Epoch 42 training complete
Epoch 43 training complete
Epoch 44 training complete
Epoch 45 training complete
Epoch 46 training complete
Epoch 47 training complete
Epoch 48 training complete
Epoch 49 training complete
Epoch 50 training complete
Epoch 51 training complete
Epoch 52 training complete
Epoch 53 training complete
Epoch 54 training complete
Epoch 55 training complete
Epoch 56 training complete
Epoch 57 training complete
Epoch 58 training complete
Epoch 59 training complete
Epoch 60 training complete
Epoch 61 training complete
Epoch 62 training complete
Epoch 63 training complete
Epoch 64 training complete
Epoch 65 training complete
Epoch 66 training complete
Epoch 67 training complete
Epoch 68 training complete
Epoch 69 training complete
Epoch 70 training complete
Epoch 71 training complete
Epoch 72 training complete
Epoch 73 training complete
Epoch 74 training complete
Epoch 75 training complete
Epoch 76 training complete
Epoch 77 training complete
Epoch 78 training complete
Epoch 79 training complete
Epoch 80 training complete
Epoch 81 training complete
Epoch 82 training complete
Epoch 83 training complete
Epoch 84 training complete
Epoch 85 training complete
Epoch 86 training complete
Epoch 87 training complete
Epoch 88 training complete
Epoch 89 training complete
Epoch 90 training complete
Epoch 91 training complete
Epoch 92 training complete
Epoch 93 training complete
Epoch 94 training complete
Epoch 95 training complete
Epoch 96 training complete
Epoch 97 training complete
Epoch 98 training complete
Epoch 99 training complete
Epoch 100 training complete
Epoch 101 training complete
Epoch 102 training complete
Epoch 103 training complete
Epoch 104 training complete
Epoch 105 training complete
Epoch 106 training complete
Epoch 107 training complete
Epoch 108 training complete
Epoch 109 training complete
Epoch 110 training complete
Epoch 111 training complete
Epoch 112 training complete
Epoch 113 training complete
Epoch 114 training complete
Epoch 115 training complete
Epoch 116 training complete
Epoch 117 training complete
Epoch 118 training complete
Epoch 119 training complete
Epoch 120 training complete
Epoch 121 training complete
Epoch 122 training complete
Epoch 123 training complete
Epoch 124 training complete
Epoch 125 training complete
Epoch 126 training complete
Epoch 127 training complete
Epoch 128 training complete
Epoch 129 training complete
Epoch 130 training complete
Epoch 131 training complete
Epoch 132 training complete
Epoch 133 training complete
Epoch 134 training complete
Epoch 135 training complete
Epoch 136 training complete
Epoch 137 training complete
Epoch 138 training complete
Epoch 139 training complete
Epoch 140 training complete
Epoch 141 training complete
Epoch 142 training complete
Epoch 143 training complete
Epoch 144 training complete
Epoch 145 training complete
Epoch 146 training complete
Epoch 147 training complete
Epoch 148 training complete
Epoch 149 training complete
Epoch 150 training complete
Epoch 151 training complete
Epoch 152 training complete
Epoch 153 training complete
Epoch 154 training complete
Epoch 155 training complete
Epoch 156 training complete
Epoch 157 training complete
Epoch 158 training complete
Epoch 159 training complete
Epoch 160 training complete
Epoch 161 training complete
Epoch 162 training complete
Epoch 163 training complete
Epoch 164 training complete
Epoch 165 training complete
Epoch 166 training complete
Epoch 167 training complete
Epoch 168 training complete
Epoch 169 training complete
Epoch 170 training complete
Epoch 171 training complete
Epoch 172 training complete
Epoch 173 training complete
Epoch 174 training complete
Epoch 175 training complete
Epoch 176 training complete
Epoch 177 training complete
Epoch 178 training complete
Epoch 179 training complete
Epoch 180 training complete
Epoch 181 training complete
Epoch 182 training complete
Epoch 183 training complete
Epoch 184 training complete
Epoch 185 training complete
Epoch 186 training complete
Epoch 187 training complete
Epoch 188 training complete
Epoch 189 training complete
Epoch 190 training complete
Epoch 191 training complete
Epoch 192 training complete
Epoch 193 training complete
Epoch 194 training complete
Epoch 195 training complete
Epoch 196 training complete
Epoch 197 training complete
Epoch 198 training complete
Epoch 199 training complete
Epoch 200 training complete
Epoch 201 training complete
Epoch 202 training complete
Epoch 203 training complete
Epoch 204 training complete
Epoch 205 training complete
Epoch 206 training complete
Epoch 207 training complete
Epoch 208 training complete
Epoch 209 training complete
Epoch 210 training complete
Epoch 211 training complete
Epoch 212 training complete
Epoch 213 training complete
Epoch 214 training complete
Epoch 215 training complete
Epoch 216 training complete
Epoch 217 training complete
Epoch 218 training complete
Epoch 219 training complete
Epoch 220 training complete
Epoch 221 training complete
Epoch 222 training complete
Epoch 223 training complete
Epoch 224 training complete
Epoch 225 training complete
Epoch 226 training complete
Epoch 227 training complete
Epoch 228 training complete
Epoch 229 training complete
Epoch 230 training complete
Epoch 231 training complete
Epoch 232 training complete
Epoch 233 training complete
Epoch 234 training complete
Epoch 235 training complete
Epoch 236 training complete
Epoch 237 training complete
Epoch 238 training complete
Epoch 239 training complete
Epoch 240 training complete
Epoch 241 training complete
Epoch 242 training complete
Epoch 243 training complete
Epoch 244 training complete
Epoch 245 training complete
Epoch 246 training complete
Epoch 247 training complete
Epoch 248 training complete
Epoch 249 training complete
Epoch 250 training complete
Epoch 251 training complete
Epoch 252 training complete
Epoch 253 training complete
Epoch 254 training complete
Epoch 255 training complete
Epoch 256 training complete
Epoch 257 training complete
Epoch 258 training complete
Epoch 259 training complete
Epoch 260 training complete
Epoch 261 training complete
Epoch 262 training complete
Epoch 263 training complete
Epoch 264 training complete
Epoch 265 training complete
Epoch 266 training complete
Epoch 267 training complete
Epoch 268 training complete
Epoch 269 training complete
Epoch 270 training complete
Epoch 271 training complete
Epoch 272 training complete
Epoch 273 training complete
Epoch 274 training complete
Epoch 275 training complete
Epoch 276 training complete
Epoch 277 training complete
Epoch 278 training complete
Epoch 279 training complete
Epoch 280 training complete
Epoch 281 training complete
Epoch 282 training complete
Epoch 283 training complete
Epoch 284 training complete
Epoch 285 training complete
Epoch 286 training complete
Epoch 287 training complete
Epoch 288 training complete
Epoch 289 training complete
Epoch 290 training complete
Epoch 291 training complete
Epoch 292 training complete
Epoch 293 training complete
Epoch 294 training complete
Epoch 295 training complete
Epoch 296 training complete
Epoch 297 training complete
Epoch 298 training complete
Epoch 299 training complete
Epoch 300 training complete
Epoch 301 training complete
Epoch 302 training complete
Epoch 303 training complete
Epoch 304 training complete
Epoch 305 training complete
Epoch 306 training complete
Epoch 307 training complete
Epoch 308 training complete
Epoch 309 training complete
Epoch 310 training complete
Epoch 311 training complete
Epoch 312 training complete
Epoch 313 training complete
Epoch 314 training complete
Epoch 315 training complete
Epoch 316 training complete
Epoch 317 training complete
Epoch 318 training complete
Epoch 319 training complete
Epoch 320 training complete
Epoch 321 training complete
Epoch 322 training complete
Epoch 323 training complete
Epoch 324 training complete
Epoch 325 training complete
Epoch 326 training complete
Epoch 327 training complete
Epoch 328 training complete
Epoch 329 training complete
Epoch 330 training complete
Epoch 331 training complete
Epoch 332 training complete
Epoch 333 training complete
Epoch 334 training complete
Epoch 335 training complete
Epoch 336 training complete
Epoch 337 training complete
Epoch 338 training complete
Epoch 339 training complete
Epoch 340 training complete
Epoch 341 training complete
Epoch 342 training complete
Epoch 343 training complete
Epoch 344 training complete
Epoch 345 training complete
Epoch 346 training complete
Epoch 347 training complete
Epoch 348 training complete
Epoch 349 training complete
Epoch 350 training complete
Epoch 351 training complete
Epoch 352 training complete
Epoch 353 training complete
Epoch 354 training complete
Epoch 355 training complete
Epoch 356 training complete
Epoch 357 training complete
Epoch 358 training complete
Epoch 359 training complete
Epoch 360 training complete
Epoch 361 training complete
Epoch 362 training complete
Epoch 363 training complete
Epoch 364 training complete
Epoch 365 training complete
Epoch 366 training complete
Epoch 367 training complete
Epoch 368 training complete
Epoch 369 training complete
Epoch 370 training complete
Epoch 371 training complete
Epoch 372 training complete
Epoch 373 training complete
Epoch 374 training complete
Epoch 375 training complete
Epoch 376 training complete
Epoch 377 training complete
Epoch 378 training complete
Epoch 379 training complete
Epoch 380 training complete
Epoch 381 training complete
Epoch 382 training complete
Epoch 383 training complete
Epoch 384 training complete
Epoch 385 training complete
Epoch 386 training complete
Epoch 387 training complete
Epoch 388 training complete
Epoch 389 training complete
Epoch 390 training complete
Epoch 391 training complete
Epoch 392 training complete
Epoch 393 training complete
Epoch 394 training complete
Epoch 395 training complete
Epoch 396 training complete
Epoch 397 training complete
Epoch 398 training complete
Epoch 399 training complete
Epoch 400 training complete
Epoch 401 training complete
Epoch 402 training complete
Epoch 403 training complete
Epoch 404 training complete
Epoch 405 training complete
Epoch 406 training complete
Epoch 407 training complete
Epoch 408 training complete
Epoch 409 training complete
Epoch 410 training complete
Epoch 411 training complete
Epoch 412 training complete
Epoch 413 training complete
Epoch 414 training complete
Epoch 415 training complete
Epoch 416 training complete
Epoch 417 training complete
Epoch 418 training complete
Epoch 419 training complete
Epoch 420 training complete
Epoch 421 training complete
Epoch 422 training complete
Epoch 423 training complete
Epoch 424 training complete
Epoch 425 training complete
Epoch 426 training complete
Epoch 427 training complete
Epoch 428 training complete
Epoch 429 training complete
Epoch 430 training complete
Epoch 431 training complete
Epoch 432 training complete
Epoch 433 training complete
Epoch 434 training complete
Epoch 435 training complete
Epoch 436 training complete
Epoch 437 training complete
Epoch 438 training complete
Epoch 439 training complete
Epoch 440 training complete
Epoch 441 training complete
Epoch 442 training complete
Epoch 443 training complete
Epoch 444 training complete
Epoch 445 training complete
Epoch 446 training complete
Epoch 447 training complete
Epoch 448 training complete
Epoch 449 training complete
Epoch 450 training complete
Epoch 451 training complete
Epoch 452 training complete
Epoch 453 training complete
Epoch 454 training complete
Epoch 455 training complete
Epoch 456 training complete
Epoch 457 training complete
Epoch 458 training complete
Epoch 459 training complete
Epoch 460 training complete
Epoch 461 training complete
Epoch 462 training complete
Epoch 463 training complete
Epoch 464 training complete
Epoch 465 training complete
Epoch 466 training complete
Epoch 467 training complete
Epoch 468 training complete
Epoch 469 training complete
Epoch 470 training complete
Epoch 471 training complete
Epoch 472 training complete
Epoch 473 training complete
Epoch 474 training complete
Epoch 475 training complete
Epoch 476 training complete
Epoch 477 training complete
Epoch 478 training complete
Epoch 479 training complete
Epoch 480 training complete
Epoch 481 training complete
Epoch 482 training complete
Epoch 483 training complete
Epoch 484 training complete
Epoch 485 training complete
Epoch 486 training complete
Epoch 487 training complete
Epoch 488 training complete
Epoch 489 training complete
Epoch 490 training complete
Epoch 491 training complete
Epoch 492 training complete
Epoch 493 training complete
Epoch 494 training complete
Epoch 495 training complete
Epoch 496 training complete
Epoch 497 training complete
Epoch 498 training complete
Epoch 499 training complete
Epoch 500 training complete
Epoch 501 training complete
Epoch 502 training complete
Epoch 503 training complete
Epoch 504 training complete
Epoch 505 training complete
Epoch 506 training complete
Epoch 507 training complete
Epoch 508 training complete
Epoch 509 training complete
Epoch 510 training complete
Epoch 511 training complete
Epoch 512 training complete
Epoch 513 training complete
Epoch 514 training complete
Epoch 515 training complete
Epoch 516 training complete
Epoch 517 training complete
Epoch 518 training complete
Epoch 519 training complete
Epoch 520 training complete
Epoch 521 training complete
Epoch 522 training complete
Epoch 523 training complete
Epoch 524 training complete
Epoch 525 training complete
Epoch 526 training complete
Epoch 527 training complete
Epoch 528 training complete
Epoch 529 training complete
Epoch 530 training complete
Epoch 531 training complete
Epoch 532 training complete
Epoch 533 training complete
Epoch 534 training complete
Epoch 535 training complete
Epoch 536 training complete
Epoch 537 training complete
Epoch 538 training complete
Epoch 539 training complete
Epoch 540 training complete
Epoch 541 training complete
Epoch 542 training complete
Epoch 543 training complete
Epoch 544 training complete
Epoch 545 training complete
Epoch 546 training complete
Epoch 547 training complete
Epoch 548 training complete
Epoch 549 training complete
Epoch 550 training complete
Epoch 551 training complete
Epoch 552 training complete
Epoch 553 training complete
Epoch 554 training complete
Epoch 555 training complete
Epoch 556 training complete
Epoch 557 training complete
Epoch 558 training complete
Epoch 559 training complete
Epoch 560 training complete
Epoch 561 training complete
Epoch 562 training complete
Epoch 563 training complete
Epoch 564 training complete
Epoch 565 training complete
Epoch 566 training complete
Epoch 567 training complete
Epoch 568 training complete
Epoch 569 training complete
Epoch 570 training complete
Epoch 571 training complete
Epoch 572 training complete
Epoch 573 training complete
Epoch 574 training complete
Epoch 575 training complete
Epoch 576 training complete
Epoch 577 training complete
Epoch 578 training complete
Epoch 579 training complete
Epoch 580 training complete
Epoch 581 training complete
Epoch 582 training complete
Epoch 583 training complete
Epoch 584 training complete
Epoch 585 training complete
Epoch 586 training complete
Epoch 587 training complete
Epoch 588 training complete
Epoch 589 training complete
Epoch 590 training complete
Epoch 591 training complete
Epoch 592 training complete
Epoch 593 training complete
Epoch 594 training complete
Epoch 595 training complete
Epoch 596 training complete
Epoch 597 training complete
Epoch 598 training complete
Epoch 599 training complete
Epoch 600 training complete
Epoch 601 training complete
Epoch 602 training complete
Epoch 603 training complete
Epoch 604 training complete
Epoch 605 training complete
Epoch 606 training complete
Epoch 607 training complete
Epoch 608 training complete
Epoch 609 training complete
Epoch 610 training complete
Epoch 611 training complete
Epoch 612 training complete
Epoch 613 training complete
Epoch 614 training complete
Epoch 615 training complete
Epoch 616 training complete
Epoch 617 training complete
Epoch 618 training complete
Epoch 619 training complete
Epoch 620 training complete
Epoch 621 training complete
Epoch 622 training complete
Epoch 623 training complete
Epoch 624 training complete
Epoch 625 training complete
Epoch 626 training complete
Epoch 627 training complete
Epoch 628 training complete
Epoch 629 training complete
Epoch 630 training complete
Epoch 631 training complete
Epoch 632 training complete
Epoch 633 training complete
Epoch 634 training complete
Epoch 635 training complete
Epoch 636 training complete
Epoch 637 training complete
Epoch 638 training complete
Epoch 639 training complete
Epoch 640 training complete
Epoch 641 training complete
Epoch 642 training complete
Epoch 643 training complete
Epoch 644 training complete
Epoch 645 training complete
Epoch 646 training complete
Epoch 647 training complete
Epoch 648 training complete
Epoch 649 training complete
Epoch 650 training complete
Epoch 651 training complete
Epoch 652 training complete
Epoch 653 training complete
Epoch 654 training complete
Epoch 655 training complete
Epoch 656 training complete
Epoch 657 training complete
Epoch 658 training complete
Epoch 659 training complete
Epoch 660 training complete
Epoch 661 training complete
Epoch 662 training complete
Epoch 663 training complete
Epoch 664 training complete
Epoch 665 training complete
Epoch 666 training complete
Epoch 667 training complete
Epoch 668 training complete
Epoch 669 training complete
Epoch 670 training complete
Epoch 671 training complete
Epoch 672 training complete
Epoch 673 training complete
Epoch 674 training complete
Epoch 675 training complete
Epoch 676 training complete
Epoch 677 training complete
Epoch 678 training complete
Epoch 679 training complete
Epoch 680 training complete
Epoch 681 training complete
Epoch 682 training complete
Epoch 683 training complete
Epoch 684 training complete
Epoch 685 training complete
Epoch 686 training complete
Epoch 687 training complete
Epoch 688 training complete
Epoch 689 training complete
Epoch 690 training complete
Epoch 691 training complete
Epoch 692 training complete
Epoch 693 training complete
Epoch 694 training complete
Epoch 695 training complete
Epoch 696 training complete
Epoch 697 training complete
Epoch 698 training complete
Epoch 699 training complete
Epoch 700 training complete
Epoch 701 training complete
Epoch 702 training complete
Epoch 703 training complete
Epoch 704 training complete
Epoch 705 training complete
Epoch 706 training complete
Epoch 707 training complete
Epoch 708 training complete
Epoch 709 training complete
Epoch 710 training complete
Epoch 711 training complete
Epoch 712 training complete
Epoch 713 training complete
Epoch 714 training complete
Epoch 715 training complete
Epoch 716 training complete
Epoch 717 training complete
Epoch 718 training complete
Epoch 719 training complete
Epoch 720 training complete
Epoch 721 training complete
Epoch 722 training complete
Epoch 723 training complete
Epoch 724 training complete
Epoch 725 training complete
Epoch 726 training complete
Epoch 727 training complete
Epoch 728 training complete
Epoch 729 training complete
Epoch 730 training complete
Epoch 731 training complete
Epoch 732 training complete
Epoch 733 training complete
Epoch 734 training complete
Epoch 735 training complete
Epoch 736 training complete
Epoch 737 training complete
Epoch 738 training complete
Epoch 739 training complete
Epoch 740 training complete
Epoch 741 training complete
Epoch 742 training complete
Epoch 743 training complete
Epoch 744 training complete
Epoch 745 training complete
Epoch 746 training complete
Epoch 747 training complete
Epoch 748 training complete
Epoch 749 training complete
Accuracy was 91.05 percent


Training network with data set size 5000
Epoch 0 training complete
Epoch 1 training complete
Epoch 2 training complete
Epoch 3 training complete
Epoch 4 training complete
Epoch 5 training complete
Epoch 6 training complete
Epoch 7 training complete
Epoch 8 training complete
Epoch 9 training complete
Epoch 10 training complete
Epoch 11 training complete
Epoch 12 training complete
Epoch 13 training complete
Epoch 14 training complete
Epoch 15 training complete
Epoch 16 training complete
Epoch 17 training complete
Epoch 18 training complete
Epoch 19 training complete
Epoch 20 training complete
Epoch 21 training complete
Epoch 22 training complete
Epoch 23 training complete
Epoch 24 training complete
Epoch 25 training complete
Epoch 26 training complete
Epoch 27 training complete
Epoch 28 training complete
Epoch 29 training complete
Epoch 30 training complete
Epoch 31 training complete
Epoch 32 training complete
Epoch 33 training complete
Epoch 34 training complete
Epoch 35 training complete
Epoch 36 training complete
Epoch 37 training complete
Epoch 38 training complete
Epoch 39 training complete
Epoch 40 training complete
Epoch 41 training complete
Epoch 42 training complete
Epoch 43 training complete
Epoch 44 training complete
Epoch 45 training complete
Epoch 46 training complete
Epoch 47 training complete
Epoch 48 training complete
Epoch 49 training complete
Epoch 50 training complete
Epoch 51 training complete
Epoch 52 training complete
Epoch 53 training complete
Epoch 54 training complete
Epoch 55 training complete
Epoch 56 training complete
Epoch 57 training complete
Epoch 58 training complete
Epoch 59 training complete
Epoch 60 training complete
Epoch 61 training complete
Epoch 62 training complete
Epoch 63 training complete
Epoch 64 training complete
Epoch 65 training complete
Epoch 66 training complete
Epoch 67 training complete
Epoch 68 training complete
Epoch 69 training complete
Epoch 70 training complete
Epoch 71 training complete
Epoch 72 training complete
Epoch 73 training complete
Epoch 74 training complete
Epoch 75 training complete
Epoch 76 training complete
Epoch 77 training complete
Epoch 78 training complete
Epoch 79 training complete
Epoch 80 training complete
Epoch 81 training complete
Epoch 82 training complete
Epoch 83 training complete
Epoch 84 training complete
Epoch 85 training complete
Epoch 86 training complete
Epoch 87 training complete
Epoch 88 training complete
Epoch 89 training complete
Epoch 90 training complete
Epoch 91 training complete
Epoch 92 training complete
Epoch 93 training complete
Epoch 94 training complete
Epoch 95 training complete
Epoch 96 training complete
Epoch 97 training complete
Epoch 98 training complete
Epoch 99 training complete
Epoch 100 training complete
Epoch 101 training complete
Epoch 102 training complete
Epoch 103 training complete
Epoch 104 training complete
Epoch 105 training complete
Epoch 106 training complete
Epoch 107 training complete
Epoch 108 training complete
Epoch 109 training complete
Epoch 110 training complete
Epoch 111 training complete
Epoch 112 training complete
Epoch 113 training complete
Epoch 114 training complete
Epoch 115 training complete
Epoch 116 training complete
Epoch 117 training complete
Epoch 118 training complete
Epoch 119 training complete
Epoch 120 training complete
Epoch 121 training complete
Epoch 122 training complete
Epoch 123 training complete
Epoch 124 training complete
Epoch 125 training complete
Epoch 126 training complete
Epoch 127 training complete
Epoch 128 training complete
Epoch 129 training complete
Epoch 130 training complete
Epoch 131 training complete
Epoch 132 training complete
Epoch 133 training complete
Epoch 134 training complete
Epoch 135 training complete
Epoch 136 training complete
Epoch 137 training complete
Epoch 138 training complete
Epoch 139 training complete
Epoch 140 training complete
Epoch 141 training complete
Epoch 142 training complete
Epoch 143 training complete
Epoch 144 training complete
Epoch 145 training complete
Epoch 146 training complete
Epoch 147 training complete
Epoch 148 training complete
Epoch 149 training complete
Epoch 150 training complete
Epoch 151 training complete
Epoch 152 training complete
Epoch 153 training complete
Epoch 154 training complete
Epoch 155 training complete
Epoch 156 training complete
Epoch 157 training complete
Epoch 158 training complete
Epoch 159 training complete
Epoch 160 training complete
Epoch 161 training complete
Epoch 162 training complete
Epoch 163 training complete
Epoch 164 training complete
Epoch 165 training complete
Epoch 166 training complete
Epoch 167 training complete
Epoch 168 training complete
Epoch 169 training complete
Epoch 170 training complete
Epoch 171 training complete
Epoch 172 training complete
Epoch 173 training complete
Epoch 174 training complete
Epoch 175 training complete
Epoch 176 training complete
Epoch 177 training complete
Epoch 178 training complete
Epoch 179 training complete
Epoch 180 training complete
Epoch 181 training complete
Epoch 182 training complete
Epoch 183 training complete
Epoch 184 training complete
Epoch 185 training complete
Epoch 186 training complete
Epoch 187 training complete
Epoch 188 training complete
Epoch 189 training complete
Epoch 190 training complete
Epoch 191 training complete
Epoch 192 training complete
Epoch 193 training complete
Epoch 194 training complete
Epoch 195 training complete
Epoch 196 training complete
Epoch 197 training complete
Epoch 198 training complete
Epoch 199 training complete
Epoch 200 training complete
Epoch 201 training complete
Epoch 202 training complete
Epoch 203 training complete
Epoch 204 training complete
Epoch 205 training complete
Epoch 206 training complete
Epoch 207 training complete
Epoch 208 training complete
Epoch 209 training complete
Epoch 210 training complete
Epoch 211 training complete
Epoch 212 training complete
Epoch 213 training complete
Epoch 214 training complete
Epoch 215 training complete
Epoch 216 training complete
Epoch 217 training complete
Epoch 218 training complete
Epoch 219 training complete
Epoch 220 training complete
Epoch 221 training complete
Epoch 222 training complete
Epoch 223 training complete
Epoch 224 training complete
Epoch 225 training complete
Epoch 226 training complete
Epoch 227 training complete
Epoch 228 training complete
Epoch 229 training complete
Epoch 230 training complete
Epoch 231 training complete
Epoch 232 training complete
Epoch 233 training complete
Epoch 234 training complete
Epoch 235 training complete
Epoch 236 training complete
Epoch 237 training complete
Epoch 238 training complete
Epoch 239 training complete
Epoch 240 training complete
Epoch 241 training complete
Epoch 242 training complete
Epoch 243 training complete
Epoch 244 training complete
Epoch 245 training complete
Epoch 246 training complete
Epoch 247 training complete
Epoch 248 training complete
Epoch 249 training complete
Epoch 250 training complete
Epoch 251 training complete
Epoch 252 training complete
Epoch 253 training complete
Epoch 254 training complete
Epoch 255 training complete
Epoch 256 training complete
Epoch 257 training complete
Epoch 258 training complete
Epoch 259 training complete
Epoch 260 training complete
Epoch 261 training complete
Epoch 262 training complete
Epoch 263 training complete
Epoch 264 training complete
Epoch 265 training complete
Epoch 266 training complete
Epoch 267 training complete
Epoch 268 training complete
Epoch 269 training complete
Epoch 270 training complete
Epoch 271 training complete
Epoch 272 training complete
Epoch 273 training complete
Epoch 274 training complete
Epoch 275 training complete
Epoch 276 training complete
Epoch 277 training complete
Epoch 278 training complete
Epoch 279 training complete
Epoch 280 training complete
Epoch 281 training complete
Epoch 282 training complete
Epoch 283 training complete
Epoch 284 training complete
Epoch 285 training complete
Epoch 286 training complete
Epoch 287 training complete
Epoch 288 training complete
Epoch 289 training complete
Epoch 290 training complete
Epoch 291 training complete
Epoch 292 training complete
Epoch 293 training complete
Epoch 294 training complete
Epoch 295 training complete
Epoch 296 training complete
Epoch 297 training complete
Epoch 298 training complete
Epoch 299 training complete
Accuracy was 93.25 percent


Training network with data set size 10000
Epoch 0 training complete
Epoch 1 training complete
Epoch 2 training complete
Epoch 3 training complete
Epoch 4 training complete
Epoch 5 training complete
Epoch 6 training complete
Epoch 7 training complete
Epoch 8 training complete
Epoch 9 training complete
Epoch 10 training complete
Epoch 11 training complete
Epoch 12 training complete
Epoch 13 training complete
Epoch 14 training complete
Epoch 15 training complete
Epoch 16 training complete
Epoch 17 training complete
Epoch 18 training complete
Epoch 19 training complete
Epoch 20 training complete
Epoch 21 training complete
Epoch 22 training complete
Epoch 23 training complete
Epoch 24 training complete
Epoch 25 training complete
Epoch 26 training complete
Epoch 27 training complete
Epoch 28 training complete
Epoch 29 training complete
Epoch 30 training complete
Epoch 31 training complete
Epoch 32 training complete
Epoch 33 training complete
Epoch 34 training complete
Epoch 35 training complete
Epoch 36 training complete
Epoch 37 training complete
Epoch 38 training complete
Epoch 39 training complete
Epoch 40 training complete
Epoch 41 training complete
Epoch 42 training complete
Epoch 43 training complete
Epoch 44 training complete
Epoch 45 training complete
Epoch 46 training complete
Epoch 47 training complete
Epoch 48 training complete
Epoch 49 training complete
Epoch 50 training complete
Epoch 51 training complete
Epoch 52 training complete
Epoch 53 training complete
Epoch 54 training complete
Epoch 55 training complete
Epoch 56 training complete
Epoch 57 training complete
Epoch 58 training complete
Epoch 59 training complete
Epoch 60 training complete
Epoch 61 training complete
Epoch 62 training complete
Epoch 63 training complete
Epoch 64 training complete
Epoch 65 training complete
Epoch 66 training complete
Epoch 67 training complete
Epoch 68 training complete
Epoch 69 training complete
Epoch 70 training complete
Epoch 71 training complete
Epoch 72 training complete
Epoch 73 training complete
Epoch 74 training complete
Epoch 75 training complete
Epoch 76 training complete
Epoch 77 training complete
Epoch 78 training complete
Epoch 79 training complete
Epoch 80 training complete
Epoch 81 training complete
Epoch 82 training complete
Epoch 83 training complete
Epoch 84 training complete
Epoch 85 training complete
Epoch 86 training complete
Epoch 87 training complete
Epoch 88 training complete
Epoch 89 training complete
Epoch 90 training complete
Epoch 91 training complete
Epoch 92 training complete
Epoch 93 training complete
Epoch 94 training complete
Epoch 95 training complete
Epoch 96 training complete
Epoch 97 training complete
Epoch 98 training complete
Epoch 99 training complete
Epoch 100 training complete
Epoch 101 training complete
Epoch 102 training complete
Epoch 103 training complete
Epoch 104 training complete
Epoch 105 training complete
Epoch 106 training complete
Epoch 107 training complete
Epoch 108 training complete
Epoch 109 training complete
Epoch 110 training complete
Epoch 111 training complete
Epoch 112 training complete
Epoch 113 training complete
Epoch 114 training complete
Epoch 115 training complete
Epoch 116 training complete
Epoch 117 training complete
Epoch 118 training complete
Epoch 119 training complete
Epoch 120 training complete
Epoch 121 training complete
Epoch 122 training complete
Epoch 123 training complete
Epoch 124 training complete
Epoch 125 training complete
Epoch 126 training complete
Epoch 127 training complete
Epoch 128 training complete
Epoch 129 training complete
Epoch 130 training complete
Epoch 131 training complete
Epoch 132 training complete
Epoch 133 training complete
Epoch 134 training complete
Epoch 135 training complete
Epoch 136 training complete
Epoch 137 training complete
Epoch 138 training complete
Epoch 139 training complete
Epoch 140 training complete
Epoch 141 training complete
Epoch 142 training complete
Epoch 143 training complete
Epoch 144 training complete
Epoch 145 training complete
Epoch 146 training complete
Epoch 147 training complete
Epoch 148 training complete
Epoch 149 training complete
Accuracy was 94.65 percent


Training network with data set size 20000
Epoch 0 training complete
Epoch 1 training complete
Epoch 2 training complete
Epoch 3 training complete
Epoch 4 training complete
Epoch 5 training complete
Epoch 6 training complete
Epoch 7 training complete
Epoch 8 training complete
Epoch 9 training complete
Epoch 10 training complete
Epoch 11 training complete
Epoch 12 training complete
Epoch 13 training complete
Epoch 14 training complete
Epoch 15 training complete
Epoch 16 training complete
Epoch 17 training complete
Epoch 18 training complete
Epoch 19 training complete
Epoch 20 training complete
Epoch 21 training complete
Epoch 22 training complete
Epoch 23 training complete
Epoch 24 training complete
Epoch 25 training complete
Epoch 26 training complete
Epoch 27 training complete
Epoch 28 training complete
Epoch 29 training complete
Epoch 30 training complete
Epoch 31 training complete
Epoch 32 training complete
Epoch 33 training complete
Epoch 34 training complete
Epoch 35 training complete
Epoch 36 training complete
Epoch 37 training complete
Epoch 38 training complete
Epoch 39 training complete
Epoch 40 training complete
Epoch 41 training complete
Epoch 42 training complete
Epoch 43 training complete
Epoch 44 training complete
Epoch 45 training complete
Epoch 46 training complete
Epoch 47 training complete
Epoch 48 training complete
Epoch 49 training complete
Epoch 50 training complete
Epoch 51 training complete
Epoch 52 training complete
Epoch 53 training complete
Epoch 54 training complete
Epoch 55 training complete
Epoch 56 training complete
Epoch 57 training complete
Epoch 58 training complete
Epoch 59 training complete
Epoch 60 training complete
Epoch 61 training complete
Epoch 62 training complete
Epoch 63 training complete
Epoch 64 training complete
Epoch 65 training complete
Epoch 66 training complete
Epoch 67 training complete
Epoch 68 training complete
Epoch 69 training complete
Epoch 70 training complete
Epoch 71 training complete
Epoch 72 training complete
Epoch 73 training complete
Epoch 74 training complete
Accuracy was 95.52 percent


Training network with data set size 50000
Epoch 0 training complete
Epoch 1 training complete
Epoch 2 training complete
Epoch 3 training complete
Epoch 4 training complete
Epoch 5 training complete
Epoch 6 training complete
Epoch 7 training complete
Epoch 8 training complete
Epoch 9 training complete
Epoch 10 training complete
Epoch 11 training complete
Epoch 12 training complete
Epoch 13 training complete
Epoch 14 training complete
Epoch 15 training complete
Epoch 16 training complete
Epoch 17 training complete
Epoch 18 training complete
Epoch 19 training complete
Epoch 20 training complete
Epoch 21 training complete
Epoch 22 training complete
Epoch 23 training complete
Epoch 24 training complete
Epoch 25 training complete
Epoch 26 training complete
Epoch 27 training complete
Epoch 28 training complete
Epoch 29 training complete
Accuracy was 96.25 percent


Training SVM with data set size 100
Accuracy was 66.77 percent


Training SVM with data set size 200
Accuracy was 77.6 percent


Training SVM with data set size 500
Accuracy was 87.64 percent


Training SVM with data set size 1000
Accuracy was 91.38 percent


Training SVM with data set size 2000
Accuracy was 93.42 percent


Training SVM with data set size 5000
Accuracy was 95.65 percent


Training SVM with data set size 10000
Accuracy was 96.6 percent


Training SVM with data set size 20000
Accuracy was 97.25 percent


Training SVM with data set size 50000
Accuracy was 98.02 percent

make_plots()

# @title
"""weight_initialization
~~~~~~~~~~~~~~~~~~~~~~~~

This program shows how weight initialization affects training.  In
particular, we'll plot out how the classification accuracies improve
using either large starting weights, whose standard deviation is 1, or
the default starting weights, whose standard deviation is 1 over the
square root of the number of input neurons.

"""

# Standard library
import json
import random
import sys

# My library
sys.path.append('../src/')
import mnist_loader
# import network2

# Third-party libraries
import matplotlib.pyplot as plt
import numpy as np

def main(filename, n, eta):
    run_network(filename, n, eta)
    make_plot(filename)

def run_network(filename, n, eta):
    """Train the network using both the default and the large starting
    weights.  Store the results in the file with name ``filename``,
    where they can later be used by ``make_plots``.

    """
    # Make results more easily reproducible
    random.seed(12345678)
    np.random.seed(12345678)
    training_data, validation_data, test_data = mnist_loader.load_data_wrapper()

    # Convert data to lists
    training_data = list(training_data)
    validation_data = list(validation_data)
    test_data = list(test_data)

    net = Network2mat([784, n, 10], cost=CrossEntropyCost)
    print("Train the network using the default starting weights.")
    default_vc, default_va, default_tc, default_ta \
        = net.SGD(training_data, 30, 10, eta, lmbda=5.0,
                  evaluation_data=validation_data,
                  monitor_evaluation_accuracy=True)
    print("Train the network using the large starting weights.")
    net.large_weight_initializer()
    large_vc, large_va, large_tc, large_ta \
        = net.SGD(training_data, 30, 10, eta, lmbda=5.0,
                  evaluation_data=validation_data,
                  monitor_evaluation_accuracy=True)
    f = open(filename, "w")
    json.dump({"default_weight_initialization":
               [default_vc, default_va, default_tc, default_ta],
               "large_weight_initialization":
               [large_vc, large_va, large_tc, large_ta]},
              f)
    f.close()

def make_plot(filename):
    """Load the results from the file ``filename``, and generate the
    corresponding plot.

    """
    f = open(filename, "r")
    results = json.load(f)
    f.close()
    default_vc, default_va, default_tc, default_ta = results[
        "default_weight_initialization"]
    large_vc, large_va, large_tc, large_ta = results[
        "large_weight_initialization"]
    # Convert raw classification numbers to percentages, for plotting
    default_va = [x/100.0 for x in default_va]
    large_va = [x/100.0 for x in large_va]
    fig = plt.figure()
    ax = fig.add_subplot(111)
    ax.plot(np.arange(0, 30, 1), large_va, color='#2A6EA6',
            label="Old approach to weight initialization")
    ax.plot(np.arange(0, 30, 1), default_va, color='#FFA933',
            label="New approach to weight initialization")
    ax.set_xlim([0, 30])
    ax.set_xlabel('Epoch')
    ax.set_ylim([85, 100])
    ax.set_title('Classification accuracy')
    plt.legend(loc="lower right")
    plt.show()

if __name__ == "__main__":
    main("weight_initialization_30.json", 30, 0.1)
    main("weight_initialization_100.json", 100, 0.1)

Train the network using the default starting weights.
Epoch 0 training complete
Accuracy on evaluation data: 9267 / 10000
Epoch 1 training complete
Accuracy on evaluation data: 9391 / 10000
Epoch 2 training complete
Accuracy on evaluation data: 9437 / 10000
Epoch 3 training complete
Accuracy on evaluation data: 9514 / 10000
Epoch 4 training complete
Accuracy on evaluation data: 9538 / 10000
Epoch 5 training complete
Accuracy on evaluation data: 9528 / 10000
Epoch 6 training complete
Accuracy on evaluation data: 9550 / 10000
Epoch 7 training complete
Accuracy on evaluation data: 9558 / 10000
Epoch 8 training complete
Accuracy on evaluation data: 9561 / 10000
Epoch 9 training complete
Accuracy on evaluation data: 9578 / 10000
Epoch 10 training complete
Accuracy on evaluation data: 9595 / 10000
Epoch 11 training complete
Accuracy on evaluation data: 9606 / 10000
Epoch 12 training complete
Accuracy on evaluation data: 9608 / 10000
Epoch 13 training complete
Accuracy on evaluation data: 9607 / 10000
Epoch 14 training complete
Accuracy on evaluation data: 9610 / 10000
Epoch 15 training complete
Accuracy on evaluation data: 9622 / 10000
Epoch 16 training complete
Accuracy on evaluation data: 9611 / 10000
Epoch 17 training complete
Accuracy on evaluation data: 9622 / 10000
Epoch 18 training complete
Accuracy on evaluation data: 9639 / 10000
Epoch 19 training complete
Accuracy on evaluation data: 9623 / 10000
Epoch 20 training complete
Accuracy on evaluation data: 9621 / 10000
Epoch 21 training complete
Accuracy on evaluation data: 9638 / 10000
Epoch 22 training complete
Accuracy on evaluation data: 9637 / 10000
Epoch 23 training complete
Accuracy on evaluation data: 9634 / 10000
Epoch 24 training complete
Accuracy on evaluation data: 9640 / 10000
Epoch 25 training complete
Accuracy on evaluation data: 9644 / 10000
Epoch 26 training complete
Accuracy on evaluation data: 9644 / 10000
Epoch 27 training complete
Accuracy on evaluation data: 9637 / 10000
Epoch 28 training complete
Accuracy on evaluation data: 9638 / 10000
Epoch 29 training complete
Accuracy on evaluation data: 9649 / 10000
Train the network using the large starting weights.
Epoch 0 training complete
Accuracy on evaluation data: 8740 / 10000
Epoch 1 training complete
Accuracy on evaluation data: 9039 / 10000
Epoch 2 training complete
Accuracy on evaluation data: 9172 / 10000
Epoch 3 training complete
Accuracy on evaluation data: 9255 / 10000
Epoch 4 training complete
Accuracy on evaluation data: 9325 / 10000
Epoch 5 training complete
Accuracy on evaluation data: 9392 / 10000
Epoch 6 training complete
Accuracy on evaluation data: 9399 / 10000
Epoch 7 training complete
Accuracy on evaluation data: 9422 / 10000
Epoch 8 training complete
Accuracy on evaluation data: 9440 / 10000
Epoch 9 training complete
Accuracy on evaluation data: 9471 / 10000
Epoch 10 training complete
Accuracy on evaluation data: 9490 / 10000
Epoch 11 training complete
Accuracy on evaluation data: 9520 / 10000
Epoch 12 training complete
Accuracy on evaluation data: 9533 / 10000
Epoch 13 training complete
Accuracy on evaluation data: 9526 / 10000
Epoch 14 training complete
Accuracy on evaluation data: 9555 / 10000
Epoch 15 training complete
Accuracy on evaluation data: 9542 / 10000
Epoch 16 training complete
Accuracy on evaluation data: 9569 / 10000
Epoch 17 training complete
Accuracy on evaluation data: 9591 / 10000
Epoch 18 training complete
Accuracy on evaluation data: 9584 / 10000
Epoch 19 training complete
Accuracy on evaluation data: 9568 / 10000
Epoch 20 training complete
Accuracy on evaluation data: 9586 / 10000
Epoch 21 training complete
Accuracy on evaluation data: 9601 / 10000
Epoch 22 training complete
Accuracy on evaluation data: 9574 / 10000
Epoch 23 training complete
Accuracy on evaluation data: 9598 / 10000
Epoch 24 training complete
Accuracy on evaluation data: 9613 / 10000
Epoch 25 training complete
Accuracy on evaluation data: 9587 / 10000
Epoch 26 training complete
Accuracy on evaluation data: 9604 / 10000
Epoch 27 training complete
Accuracy on evaluation data: 9619 / 10000
Epoch 28 training complete
Accuracy on evaluation data: 9630 / 10000
Epoch 29 training complete
Accuracy on evaluation data: 9630 / 10000

Train the network using the default starting weights.
Epoch 0 training complete
Accuracy on evaluation data: 9305 / 10000
Epoch 1 training complete
Accuracy on evaluation data: 9479 / 10000
Epoch 2 training complete
Accuracy on evaluation data: 9538 / 10000
Epoch 3 training complete
Accuracy on evaluation data: 9631 / 10000
Epoch 4 training complete
Accuracy on evaluation data: 9662 / 10000
Epoch 5 training complete
Accuracy on evaluation data: 9672 / 10000
Epoch 6 training complete
Accuracy on evaluation data: 9694 / 10000
Epoch 7 training complete
Accuracy on evaluation data: 9705 / 10000
Epoch 8 training complete
Accuracy on evaluation data: 9713 / 10000
Epoch 9 training complete
Accuracy on evaluation data: 9724 / 10000
Epoch 10 training complete
Accuracy on evaluation data: 9737 / 10000
Epoch 11 training complete
Accuracy on evaluation data: 9743 / 10000
Epoch 12 training complete
Accuracy on evaluation data: 9746 / 10000
Epoch 13 training complete
Accuracy on evaluation data: 9748 / 10000
Epoch 14 training complete
Accuracy on evaluation data: 9767 / 10000
Epoch 15 training complete
Accuracy on evaluation data: 9763 / 10000
Epoch 16 training complete
Accuracy on evaluation data: 9763 / 10000
Epoch 17 training complete
Accuracy on evaluation data: 9757 / 10000
Epoch 18 training complete
Accuracy on evaluation data: 9767 / 10000
Epoch 19 training complete
Accuracy on evaluation data: 9771 / 10000
Epoch 20 training complete
Accuracy on evaluation data: 9774 / 10000
Epoch 21 training complete
Accuracy on evaluation data: 9774 / 10000
Epoch 22 training complete
Accuracy on evaluation data: 9774 / 10000
Epoch 23 training complete
Accuracy on evaluation data: 9776 / 10000
Epoch 24 training complete
Accuracy on evaluation data: 9781 / 10000
Epoch 25 training complete
Accuracy on evaluation data: 9775 / 10000
Epoch 26 training complete
Accuracy on evaluation data: 9779 / 10000
Epoch 27 training complete
Accuracy on evaluation data: 9763 / 10000
Epoch 28 training complete
Accuracy on evaluation data: 9778 / 10000
Epoch 29 training complete
Accuracy on evaluation data: 9784 / 10000
Train the network using the large starting weights.
Epoch 0 training complete
Accuracy on evaluation data: 9001 / 10000
Epoch 1 training complete
Accuracy on evaluation data: 9167 / 10000
Epoch 2 training complete
Accuracy on evaluation data: 9319 / 10000
Epoch 3 training complete
Accuracy on evaluation data: 9347 / 10000
Epoch 4 training complete
Accuracy on evaluation data: 9423 / 10000
Epoch 5 training complete
Accuracy on evaluation data: 9469 / 10000
Epoch 6 training complete
Accuracy on evaluation data: 9520 / 10000
Epoch 7 training complete
Accuracy on evaluation data: 9536 / 10000
Epoch 8 training complete
Accuracy on evaluation data: 9556 / 10000
Epoch 9 training complete
Accuracy on evaluation data: 9589 / 10000
Epoch 10 training complete
Accuracy on evaluation data: 9596 / 10000
Epoch 11 training complete
Accuracy on evaluation data: 9623 / 10000
Epoch 12 training complete
Accuracy on evaluation data: 9624 / 10000
Epoch 13 training complete
Accuracy on evaluation data: 9649 / 10000
Epoch 14 training complete
Accuracy on evaluation data: 9666 / 10000
Epoch 15 training complete
Accuracy on evaluation data: 9675 / 10000
Epoch 16 training complete
Accuracy on evaluation data: 9673 / 10000
Epoch 17 training complete
Accuracy on evaluation data: 9689 / 10000
Epoch 18 training complete
Accuracy on evaluation data: 9695 / 10000
Epoch 19 training complete
Accuracy on evaluation data: 9688 / 10000
Epoch 20 training complete
Accuracy on evaluation data: 9687 / 10000
Epoch 21 training complete
Accuracy on evaluation data: 9716 / 10000
Epoch 22 training complete
Accuracy on evaluation data: 9702 / 10000
Epoch 23 training complete
Accuracy on evaluation data: 9706 / 10000
Epoch 24 training complete
Accuracy on evaluation data: 9719 / 10000
Epoch 25 training complete
Accuracy on evaluation data: 9728 / 10000
Epoch 26 training complete
Accuracy on evaluation data: 9732 / 10000
Epoch 27 training complete
Accuracy on evaluation data: 9734 / 10000
Epoch 28 training complete
Accuracy on evaluation data: 9743 / 10000
Epoch 29 training complete
Accuracy on evaluation data: 9734 / 10000

make_plot("/content/weight_initialization_30.json")

make_plot("/content/weight_initialization_100.json")

%%time
# Load data
training_data, validation_data, test_data = mnist_loader.load_data_wrapper()

# Convert data to lists
training_data = list(training_data)
test_data = list(test_data)

# Initialize and train network
net = Network2mat([784, 30, 10], cost=CrossEntropyCost, reg='l1')
net.SGD(training_data, 30, 10, 0.5, evaluation_data=test_data, monitor_evaluation_accuracy=True)

Epoch 0 training complete
Accuracy on evaluation data: 9433 / 10000
Epoch 1 training complete
Accuracy on evaluation data: 9478 / 10000
Epoch 2 training complete
Accuracy on evaluation data: 9468 / 10000
Epoch 3 training complete
Accuracy on evaluation data: 9506 / 10000
Epoch 4 training complete
Accuracy on evaluation data: 9495 / 10000
Epoch 5 training complete
Accuracy on evaluation data: 9526 / 10000
Epoch 6 training complete
Accuracy on evaluation data: 9523 / 10000
Epoch 7 training complete
Accuracy on evaluation data: 9530 / 10000
Epoch 8 training complete
Accuracy on evaluation data: 9572 / 10000
Epoch 9 training complete
Accuracy on evaluation data: 9531 / 10000
Epoch 10 training complete
Accuracy on evaluation data: 9597 / 10000
Epoch 11 training complete
Accuracy on evaluation data: 9565 / 10000
Epoch 12 training complete
Accuracy on evaluation data: 9568 / 10000
Epoch 13 training complete
Accuracy on evaluation data: 9576 / 10000
Epoch 14 training complete
Accuracy on evaluation data: 9553 / 10000
Epoch 15 training complete
Accuracy on evaluation data: 9570 / 10000
Epoch 16 training complete
Accuracy on evaluation data: 9548 / 10000
Epoch 17 training complete
Accuracy on evaluation data: 9542 / 10000
Epoch 18 training complete
Accuracy on evaluation data: 9535 / 10000
Epoch 19 training complete
Accuracy on evaluation data: 9575 / 10000
Epoch 20 training complete
Accuracy on evaluation data: 9560 / 10000
Epoch 21 training complete
Accuracy on evaluation data: 9600 / 10000
Epoch 22 training complete
Accuracy on evaluation data: 9546 / 10000
Epoch 23 training complete
Accuracy on evaluation data: 9586 / 10000
Epoch 24 training complete
Accuracy on evaluation data: 9545 / 10000
Epoch 25 training complete
Accuracy on evaluation data: 9540 / 10000
Epoch 26 training complete
Accuracy on evaluation data: 9574 / 10000
Epoch 27 training complete
Accuracy on evaluation data: 9563 / 10000
Epoch 28 training complete
Accuracy on evaluation data: 9585 / 10000
Epoch 29 training complete
Accuracy on evaluation data: 9576 / 10000
CPU times: user 1min 52s, sys: 28.7 s, total: 2min 21s
Wall time: 2min 2s

([],
 [9433,
  9478,
  9468,
  9506,
  9495,
  9526,
  9523,
  9530,
  9572,
  9531,
  9597,
  9565,
  9568,
  9576,
  9553,
  9570,
  9548,
  9542,
  9535,
  9575,
  9560,
  9600,
  9546,
  9586,
  9545,
  9540,
  9574,
  9563,
  9585,
  9576],
 [],
 [])

"""multiple_eta
~~~~~~~~~~~~~~~

This program shows how different values for the learning rate affect
training.  In particular, we'll plot out how the cost changes using
three different values for eta.

"""

# Standard library
import json
import random
import sys

# My library
sys.path.append('../src/')
import mnist_loader
# import network2

# Third-party libraries
import matplotlib.pyplot as plt
import numpy as np

# Constants
LEARNING_RATES = [0.025, 0.25, 2.5]
COLORS = ['#2A6EA6', '#FFCD33', '#FF7033']
NUM_EPOCHS = 30

def main():
    run_networks()
    make_plot()

def run_networks():
    """Train networks using three different values for the learning rate,
    and store the cost curves in the file ``multiple_eta.json``, where
    they can later be used by ``make_plot``.

    """
    # Make results more easily reproducible
    random.seed(12345678)
    np.random.seed(12345678)
    training_data, validation_data, test_data = mnist_loader.load_data_wrapper()

    # Convert data to lists
    training_data = list(training_data)
    validation_data = list(validation_data)
    test_data = list(test_data)

    results = []
    for eta in LEARNING_RATES:
        print("\nTrain a network using eta = "+str(eta))
        net = Network2mat([784, 30, 10])
        results.append(
            net.SGD(training_data, NUM_EPOCHS, 10, eta, lmbda=5.0,
                    evaluation_data=validation_data,
                    monitor_training_cost=True))
    f = open("multiple_eta.json", "w")
    json.dump(results, f)
    f.close()

def make_plot():
    f = open("multiple_eta.json", "r")
    results = json.load(f)
    f.close()
    fig = plt.figure()
    ax = fig.add_subplot(111)
    for eta, result, color in zip(LEARNING_RATES, results, COLORS):
        _, _, training_cost, _ = result
        ax.plot(np.arange(NUM_EPOCHS), training_cost, "o-",
                label="$\eta$ = "+str(eta),
                color=color)
    ax.set_xlim([0, NUM_EPOCHS])
    ax.set_xlabel('Epoch')
    ax.set_ylabel('Cost')
    plt.legend(loc='upper right')
    plt.show()

if __name__ == "__main__":
    main()

Train a network using eta = 0.025
Epoch 0 training complete
Cost on training data: 0.8770331897602134
Epoch 1 training complete
Cost on training data: 0.6800202379930531
Epoch 2 training complete
Cost on training data: 0.6013079806061424
Epoch 3 training complete
Cost on training data: 0.5527053739105007
Epoch 4 training complete
Cost on training data: 0.5179543849384399
Epoch 5 training complete
Cost on training data: 0.4955397040294193
Epoch 6 training complete
Cost on training data: 0.4745554137439946
Epoch 7 training complete
Cost on training data: 0.4587896115016323
Epoch 8 training complete
Cost on training data: 0.4444038098998699
Epoch 9 training complete
Cost on training data: 0.4317001615800511
Epoch 10 training complete
Cost on training data: 0.42276045930292666
Epoch 11 training complete
Cost on training data: 0.41366345768044344
Epoch 12 training complete
Cost on training data: 0.4052121487359579
Epoch 13 training complete
Cost on training data: 0.4011762531301635
Epoch 14 training complete
Cost on training data: 0.39208573758375054
Epoch 15 training complete
Cost on training data: 0.38782297744898403
Epoch 16 training complete
Cost on training data: 0.38282400824016694
Epoch 17 training complete
Cost on training data: 0.3761222504114732
Epoch 18 training complete
Cost on training data: 0.373326247673892
Epoch 19 training complete
Cost on training data: 0.3707198539121999
Epoch 20 training complete
Cost on training data: 0.3676077879858481
Epoch 21 training complete
Cost on training data: 0.3627197104918224
Epoch 22 training complete
Cost on training data: 0.359770388374952
Epoch 23 training complete
Cost on training data: 0.36024706957392433
Epoch 24 training complete
Cost on training data: 0.3546841872005115
Epoch 25 training complete
Cost on training data: 0.3505525396970845
Epoch 26 training complete
Cost on training data: 0.34908251185091005
Epoch 27 training complete
Cost on training data: 0.34976387165142
Epoch 28 training complete
Cost on training data: 0.34525207515323214
Epoch 29 training complete
Cost on training data: 0.3434612543680221

Train a network using eta = 0.25
Epoch 0 training complete
Cost on training data: 0.4837516024444846
Epoch 1 training complete
Cost on training data: 0.4425375562096384
Epoch 2 training complete
Cost on training data: 0.38821105445071585
Epoch 3 training complete
Cost on training data: 0.3848554310323614
Epoch 4 training complete
Cost on training data: 0.3689079931901275
Epoch 5 training complete
Cost on training data: 0.35680467532721527
Epoch 6 training complete
Cost on training data: 0.3531395937702067
Epoch 7 training complete
Cost on training data: 0.3621633254668532
Epoch 8 training complete
Cost on training data: 0.3422049808557279
Epoch 9 training complete
Cost on training data: 0.3510130610748645
Epoch 10 training complete
Cost on training data: 0.34592115896265
Epoch 11 training complete
Cost on training data: 0.3516460058684411
Epoch 12 training complete
Cost on training data: 0.3425555289441641
Epoch 13 training complete
Cost on training data: 0.33214138873660676
Epoch 14 training complete
Cost on training data: 0.3303001208079793
Epoch 15 training complete
Cost on training data: 0.33926913160117556
Epoch 16 training complete
Cost on training data: 0.3470145604521435
Epoch 17 training complete
Cost on training data: 0.33410043985957644
Epoch 18 training complete
Cost on training data: 0.32379859633231783
Epoch 19 training complete
Cost on training data: 0.33143473734499795
Epoch 20 training complete
Cost on training data: 0.33939467389087663
Epoch 21 training complete
Cost on training data: 0.3315227602386052
Epoch 22 training complete
Cost on training data: 0.33941168857525206
Epoch 23 training complete
Cost on training data: 0.3244202623356884
Epoch 24 training complete
Cost on training data: 0.32137496383030506
Epoch 25 training complete
Cost on training data: 0.3444015102464981
Epoch 26 training complete
Cost on training data: 0.3189434147710474
Epoch 27 training complete
Cost on training data: 0.3265885067082064
Epoch 28 training complete
Cost on training data: 0.3243196459654086
Epoch 29 training complete
Cost on training data: 0.33682509182487597

Train a network using eta = 2.5
Epoch 0 training complete
Cost on training data: 0.8704860427795891
Epoch 1 training complete
Cost on training data: 0.8378201807565109
Epoch 2 training complete
Cost on training data: 0.8357597543081385
Epoch 3 training complete
Cost on training data: 0.8030169683974442
Epoch 4 training complete
Cost on training data: 0.7609997844869193
Epoch 5 training complete
Cost on training data: 0.7810435180893391
Epoch 6 training complete
Cost on training data: 0.7843999962934751
Epoch 7 training complete
Cost on training data: 0.8073463329326707
Epoch 8 training complete
Cost on training data: 0.7596941141589468
Epoch 9 training complete
Cost on training data: 0.7729554594197052
Epoch 10 training complete
Cost on training data: 0.8264344798208973
Epoch 11 training complete
Cost on training data: 0.8280576013224166
Epoch 12 training complete
Cost on training data: 0.7363904309525556
Epoch 13 training complete
Cost on training data: 0.7776189718823248
Epoch 14 training complete
Cost on training data: 0.8829167668135711
Epoch 15 training complete
Cost on training data: 0.8101941700673816
Epoch 16 training complete
Cost on training data: 0.7489392673140054
Epoch 17 training complete
Cost on training data: 0.8679774972931549
Epoch 18 training complete
Cost on training data: 0.7424915984894183
Epoch 19 training complete
Cost on training data: 0.9238272263101148
Epoch 20 training complete
Cost on training data: 0.8416209003383086
Epoch 21 training complete
Cost on training data: 0.7787935992188466
Epoch 22 training complete
Cost on training data: 0.803045572771393
Epoch 23 training complete
Cost on training data: 0.7687473380737608
Epoch 24 training complete
Cost on training data: 0.848359845940096
Epoch 25 training complete
Cost on training data: 0.861285966451833
Epoch 26 training complete
Cost on training data: 0.8173613562148566
Epoch 27 training complete
Cost on training data: 0.7514922662388354
Epoch 28 training complete
Cost on training data: 0.8522702949192857
Epoch 29 training complete
Cost on training data: 0.8207342717199184

make_plot()

# @title
"""network2mat2.py
~~~~~~~~~~~~~~

An improved version of network2mat.py, that adds early stopping.

"""

#### Main Network class
class Network2mat2(object):

    def __init__(self, sizes, cost=CrossEntropyCost, reg='l2'):
        self.num_layers = len(sizes)
        self.sizes = sizes
        self.default_weight_initializer()
        self.cost = cost
        self.reg = reg

    def default_weight_initializer(self):
        self.biases = [np.random.randn(y, 1) for y in self.sizes[1:]]
        self.weights = [np.random.randn(y, x)/np.sqrt(x)
                        for x, y in zip(self.sizes[:-1], self.sizes[1:])]

    def large_weight_initializer(self):
        self.biases = [np.random.randn(y, 1) for y in self.sizes[1:]]
        self.weights = [np.random.randn(y, x)
                        for x, y in zip(self.sizes[:-1], self.sizes[1:])]

    def feedforward(self, a):
        """Return the output of the network if ``a`` is input."""
        for b, w in zip(self.biases, self.weights):
            a = sigmoid(np.dot(w, a)+b)
        return a

    def SGD(self, training_data, epochs, mini_batch_size, eta,
            lmbda=0.0, evaluation_data=None,
            monitor_evaluation_cost=False,
            monitor_evaluation_accuracy=False,
            monitor_training_cost=False,
            monitor_training_accuracy=False,
            early_stopping_n=10):
        """
        Train the neural network using mini-batch stochastic gradient descent with
        early stopping. Early stopping is triggered if the performance on the
        evaluation data does not improve for 'early_stopping_n' consecutive epochs.

        Parameters:
        - early_stopping_n: The number of epochs to wait for improvement before stopping
        """
        if evaluation_data:
            n_data = len(list(evaluation_data))
        n = len(list(training_data))

        evaluation_cost, evaluation_accuracy = [], []
        training_cost, training_accuracy = [], []

        best_evaluation_accuracy = 0
        best_evaluation_cost = float('inf')
        epochs_without_improvement = 0

        for j in range(epochs):
            # Shuffle and create mini-batches
            random.shuffle(training_data)
            mini_batches = [
                training_data[k:k+mini_batch_size]
                for k in range(0, n, mini_batch_size)]

            # Update weights and biases using each mini-batch
            for mini_batch in mini_batches:
                self.update_mini_batch(mini_batch, eta, lmbda, n)

            print(f"Epoch {j} training complete")

            # Monitor cost/accuracy if required
            if monitor_training_cost:
                cost = self.total_cost(training_data, lmbda)
                training_cost.append(cost)
                print(f"Cost on training data: {cost}")
            if monitor_training_accuracy:
                accuracy = self.accuracy(training_data, convert=True)
                training_accuracy.append(accuracy)
                print(f"Accuracy on training data: {accuracy} / {n}")
            if monitor_evaluation_cost:
                cost = self.total_cost(evaluation_data, lmbda, convert=True)
                evaluation_cost.append(cost)
                print(f"Cost on evaluation data: {cost}")
            if monitor_evaluation_accuracy:
                accuracy = self.accuracy(evaluation_data)
                evaluation_accuracy.append(accuracy)
                print(f"Accuracy on evaluation data: {accuracy} / {n_data}")

            # Early stopping logic based on evaluation accuracy or cost
            if monitor_evaluation_accuracy:
                if accuracy > best_evaluation_accuracy:
                    best_evaluation_accuracy = accuracy
                    epochs_without_improvement = 0  # Reset counter if we see improvement
                else:
                    epochs_without_improvement += 1

            elif monitor_evaluation_cost:
                if cost < best_evaluation_cost:
                    best_evaluation_cost = cost
                    epochs_without_improvement = 0  # Reset counter if we see improvement
                else:
                    epochs_without_improvement += 1

            # Check for early stopping condition
            if epochs_without_improvement >= early_stopping_n:
                print(f"Early stopping after {j+1} epochs due to no improvement for {early_stopping_n} epochs.")
                break  # Exit the training loop

        return evaluation_cost, evaluation_accuracy, training_cost, training_accuracy

    def update_mini_batch(self, mini_batch, eta, lmbda, n):
        """Update the network's weights and biases by applying gradient descent
        using backpropagation to a single mini-batch."""

        # Convert the mini-batch into matrix form
        X = np.column_stack([x for x, y in mini_batch])
        Y = np.column_stack([y for x, y in mini_batch])

        # Perform forward and backward passes
        nabla_b, nabla_w = self.backprop_matrix(X, Y)

        # Update weights and biases using gradient descent and L1 or L2 regularization
        if self.reg == 'l2':
          self.weights = [(1 - eta * (lmbda/n)) * w - (eta/len(mini_batch)) * nw
                          for w, nw in zip(self.weights, nabla_w)]
        elif self.reg == 'l1':
          self.weights = [w - (eta * (lmbda/n)) * np.sign(w) - (eta / len(mini_batch)) * nw
                          for w, nw in zip(self.weights, nabla_w)]
        self.biases = [b - (eta / len(mini_batch)) * nb
                      for b, nb in zip(self.biases, nabla_b)]


    def backprop_matrix(self, X, Y):
        """Matrix-based backpropagation for a mini-batch."""

        # Initialize gradients
        nabla_b = [np.zeros(b.shape) for b in self.biases]
        nabla_w = [np.zeros(w.shape) for w in self.weights]

        # Step 1: Forward pass
        activation = X
        activations = [X]  # Store all activations layer by layer
        zs = []  # Store all z vectors layer by layer

        # Feedforward step through all layers
        for b, w in zip(self.biases, self.weights):
            z = np.dot(w, activation) + b
            zs.append(z)
            activation = sigmoid(z)
            activations.append(activation)

        # Step 2: Backward pass

        # Compute delta at the output layer
        delta = self.cost.delta(zs[-1], activations[-1], Y)

        # Gradients for the output layer
        nabla_b[-1] = np.sum(delta, axis=1, keepdims=True)
        nabla_w[-1] = np.dot(delta, activations[-2].T)

        # Backpropagate through previous layers
        for l in range(2, self.num_layers):
            z = zs[-l]
            sp = sigmoid_prime(z)
            delta = np.dot(self.weights[-l+1].T, delta) * sp
            nabla_b[-l] = np.sum(delta, axis=1, keepdims=True)
            nabla_w[-l] = np.dot(delta, activations[-l-1].T)

        return (nabla_b, nabla_w)

    def accuracy(self, data, convert=False):
        if convert:
            results = [(np.argmax(self.feedforward(x)), np.argmax(y))
                       for (x, y) in data]
        else:
            results = [(np.argmax(self.feedforward(x)), y)
                        for (x, y) in data]
        return sum(int(x == y) for (x, y) in results)

    def total_cost(self, data, lmbda, convert=False):
        cost = 0.0
        for x, y in data:
            a = self.feedforward(x)
            if convert: y = vectorized_result(y)
            cost += self.cost.fn(a, y)/len(data)
        if self.reg == 'l2':
          cost += 0.5*(lmbda/len(data))*sum(np.linalg.norm(w)**2 for w in self.weights)
        elif self.reg == 'l1':
          cost += (lmbda/len(data))*sum(np.linalg.norm(w, ord=1) for w in self.weights)
        return cost

    def save(self, filename):
        """Save the neural network to the file ``filename``."""
        data = {"sizes": self.sizes,
                "weights": [w.tolist() for w in self.weights],
                "biases": [b.tolist() for b in self.biases],
                "cost": str(self.cost.__name__)}
        f = open(filename, "w")
        json.dump(data, f)
        f.close()

# @title
"""network2mat3.py
~~~~~~~~~~~~~~

An improved version of network2mat2.py, that adds
New Early Stopping Rule: Relative Improvement Below Threshold.

"""

#### Main Network class
class Network2mat3(object):

    def __init__(self, sizes, cost=CrossEntropyCost, reg='l2'):
        self.num_layers = len(sizes)
        self.sizes = sizes
        self.default_weight_initializer()
        self.cost = cost
        self.reg = reg

    def default_weight_initializer(self):
        self.biases = [np.random.randn(y, 1) for y in self.sizes[1:]]
        self.weights = [np.random.randn(y, x)/np.sqrt(x)
                        for x, y in zip(self.sizes[:-1], self.sizes[1:])]

    def large_weight_initializer(self):
        self.biases = [np.random.randn(y, 1) for y in self.sizes[1:]]
        self.weights = [np.random.randn(y, x)
                        for x, y in zip(self.sizes[:-1], self.sizes[1:])]

    def feedforward(self, a):
        """Return the output of the network if ``a`` is input."""
        for b, w in zip(self.biases, self.weights):
            a = sigmoid(np.dot(w, a)+b)
        return a


    def SGD(self, training_data, epochs, mini_batch_size, eta,
            lmbda=0.0, evaluation_data=None,
            monitor_evaluation_cost=False,
            monitor_evaluation_accuracy=False,
            monitor_training_cost=False,
            monitor_training_accuracy=False,
            early_stopping_n=10, relative_improvement_threshold=0.01):
        """
        Train the neural network using mini-batch stochastic gradient descent with
        early stopping. There are two stopping criteria:
        - No-improvement-in-n: Stop if there is no improvement in performance for `n` epochs.
        - Relative improvement threshold: Stop if the improvement is smaller than the threshold.

        Parameters:
        - early_stopping_n: The number of epochs to wait for improvement before stopping.
        - relative_improvement_threshold: The minimum relative improvement required between epochs.
        """
        if evaluation_data:
            n_data = len(list(evaluation_data))
        n = len(list(training_data))

        evaluation_cost, evaluation_accuracy = [], []
        training_cost, training_accuracy = [], []

        best_evaluation_accuracy = 0
        best_evaluation_cost = float('inf')
        epochs_without_improvement = 0
        previous_accuracy = 0
        previous_cost = float('inf')

        for j in range(epochs):
            # Shuffle and create mini-batches
            random.shuffle(training_data)
            mini_batches = [
                training_data[k:k+mini_batch_size]
                for k in range(0, n, mini_batch_size)]

            # Update weights and biases using each mini-batch
            for mini_batch in mini_batches:
                self.update_mini_batch(mini_batch, eta, lmbda, n)

            print(f"Epoch {j} training complete")

            # Monitor cost/accuracy if required
            if monitor_training_cost:
                cost = self.total_cost(training_data, lmbda)
                training_cost.append(cost)
                print(f"Cost on training data: {cost}")
            if monitor_training_accuracy:
                accuracy = self.accuracy(training_data, convert=True)
                training_accuracy.append(accuracy)
                print(f"Accuracy on training data: {accuracy} / {n}")
            if monitor_evaluation_cost:
                cost = self.total_cost(evaluation_data, lmbda, convert=True)
                evaluation_cost.append(cost)
                print(f"Cost on evaluation data: {cost}")
            if monitor_evaluation_accuracy:
                accuracy = self.accuracy(evaluation_data)
                evaluation_accuracy.append(accuracy)
                print(f"Accuracy on evaluation data: {accuracy} / {n_data}")

            # Early stopping logic based on no-improvement-in-n or relative improvement
            if monitor_evaluation_accuracy:
                relative_improvement = abs((accuracy - previous_accuracy) / (previous_accuracy + 1e-8))  # avoid division by zero

                if accuracy > best_evaluation_accuracy:
                    best_evaluation_accuracy = accuracy
                    epochs_without_improvement = 0  # Reset counter if we see improvement
                else:
                    epochs_without_improvement += 1

                # Relative improvement check
                if relative_improvement < relative_improvement_threshold:
                    print(f"Early stopping after {j+1} epochs due to relative improvement threshold on eval. accuracy: \n{relative_improvement:.6f} < {relative_improvement_threshold}")
                    break  # Stop if relative improvement is small

                previous_accuracy = accuracy  # Update the previous accuracy for next iteration

            elif monitor_evaluation_cost:
                relative_improvement = abs((cost - previous_cost) / (previous_cost + 1e-8))  # avoid division by zero

                if cost < best_evaluation_cost:
                    best_evaluation_cost = cost
                    epochs_without_improvement = 0  # Reset counter if we see improvement
                else:
                    epochs_without_improvement += 1

                # Relative improvement check
                if relative_improvement < relative_improvement_threshold:
                    print(f"Early stopping after {j+1} epochs due to relative improvement threshold on eval. cost: \n{relative_improvement:.6f} < {relative_improvement_threshold}")
                    break  # Stop if relative improvement is small

                previous_cost = cost  # Update the previous cost for next iteration

            # No-improvement-in-n epochs check
            if epochs_without_improvement >= early_stopping_n:
                print(f"Early stopping after {j+1} epochs due to no improvement for {early_stopping_n} epochs.")
                break  # Exit the training loop

        return evaluation_cost, evaluation_accuracy, training_cost, training_accuracy

    def update_mini_batch(self, mini_batch, eta, lmbda, n):
        """Update the network's weights and biases by applying gradient descent
        using backpropagation to a single mini-batch."""

        # Convert the mini-batch into matrix form
        X = np.column_stack([x for x, y in mini_batch])
        Y = np.column_stack([y for x, y in mini_batch])

        # Perform forward and backward passes
        nabla_b, nabla_w = self.backprop_matrix(X, Y)

        # Update weights and biases using gradient descent and L1 or L2 regularization
        if self.reg == 'l2':
          self.weights = [(1 - eta * (lmbda/n)) * w - (eta/len(mini_batch)) * nw
                          for w, nw in zip(self.weights, nabla_w)]
        elif self.reg == 'l1':
          self.weights = [w - (eta * (lmbda/n)) * np.sign(w) - (eta / len(mini_batch)) * nw
                          for w, nw in zip(self.weights, nabla_w)]
        self.biases = [b - (eta / len(mini_batch)) * nb
                      for b, nb in zip(self.biases, nabla_b)]


    def backprop_matrix(self, X, Y):
        """Matrix-based backpropagation for a mini-batch."""

        # Initialize gradients
        nabla_b = [np.zeros(b.shape) for b in self.biases]
        nabla_w = [np.zeros(w.shape) for w in self.weights]

        # Step 1: Forward pass
        activation = X
        activations = [X]  # Store all activations layer by layer
        zs = []  # Store all z vectors layer by layer

        # Feedforward step through all layers
        for b, w in zip(self.biases, self.weights):
            z = np.dot(w, activation) + b
            zs.append(z)
            activation = sigmoid(z)
            activations.append(activation)

        # Step 2: Backward pass

        # Compute delta at the output layer
        delta = self.cost.delta(zs[-1], activations[-1], Y)

        # Gradients for the output layer
        nabla_b[-1] = np.sum(delta, axis=1, keepdims=True)
        nabla_w[-1] = np.dot(delta, activations[-2].T)

        # Backpropagate through previous layers
        for l in range(2, self.num_layers):
            z = zs[-l]
            sp = sigmoid_prime(z)
            delta = np.dot(self.weights[-l+1].T, delta) * sp
            nabla_b[-l] = np.sum(delta, axis=1, keepdims=True)
            nabla_w[-l] = np.dot(delta, activations[-l-1].T)

        return (nabla_b, nabla_w)

    def accuracy(self, data, convert=False):
        if convert:
            results = [(np.argmax(self.feedforward(x)), np.argmax(y))
                       for (x, y) in data]
        else:
            results = [(np.argmax(self.feedforward(x)), y)
                        for (x, y) in data]
        return sum(int(x == y) for (x, y) in results)

    def total_cost(self, data, lmbda, convert=False):
        cost = 0.0
        for x, y in data:
            a = self.feedforward(x)
            if convert: y = vectorized_result(y)
            cost += self.cost.fn(a, y)/len(data)
        if self.reg == 'l2':
          cost += 0.5*(lmbda/len(data))*sum(np.linalg.norm(w)**2 for w in self.weights)
        elif self.reg == 'l1':
          cost += (lmbda/len(data))*sum(np.linalg.norm(w, ord=1) for w in self.weights)
        return cost

    def save(self, filename):
        """Save the neural network to the file ``filename``."""
        data = {"sizes": self.sizes,
                "weights": [w.tolist() for w in self.weights],
                "biases": [b.tolist() for b in self.biases],
                "cost": str(self.cost.__name__)}
        f = open(filename, "w")
        json.dump(data, f)
        f.close()

%%time
# Initialize and train network
net = Network2mat2([784, 30, 10], cost=CrossEntropyCost)
net.large_weight_initializer()
eval_cost_n10, eval_acc_n10, train_cost_n10, train_acc_n10 = net.SGD(
        training_data, 30, 10, 0.5, 5, evaluation_data=test_data,
        monitor_evaluation_cost=True, monitor_evaluation_accuracy=True,
        early_stopping_n=10)

Epoch 0 training complete
Cost on evaluation data: 4.370084705653962
Accuracy on evaluation data: 9170 / 10000
Epoch 1 training complete
Cost on evaluation data: 2.952264492793333
Accuracy on evaluation data: 9315 / 10000
Epoch 2 training complete
Cost on evaluation data: 2.11608194188085
Accuracy on evaluation data: 9407 / 10000
Epoch 3 training complete
Cost on evaluation data: 1.6206092381096153
Accuracy on evaluation data: 9483 / 10000
Epoch 4 training complete
Cost on evaluation data: 1.338428694958743
Accuracy on evaluation data: 9517 / 10000
Epoch 5 training complete
Cost on evaluation data: 1.1792030229196118
Accuracy on evaluation data: 9502 / 10000
Epoch 6 training complete
Cost on evaluation data: 1.1349722093919958
Accuracy on evaluation data: 9434 / 10000
Epoch 7 training complete
Cost on evaluation data: 1.0295785208869956
Accuracy on evaluation data: 9550 / 10000
Epoch 8 training complete
Cost on evaluation data: 1.0377775762636152
Accuracy on evaluation data: 9487 / 10000
Epoch 9 training complete
Cost on evaluation data: 1.0068097471168036
Accuracy on evaluation data: 9520 / 10000
Epoch 10 training complete
Cost on evaluation data: 0.9794982764219005
Accuracy on evaluation data: 9544 / 10000
Epoch 11 training complete
Cost on evaluation data: 0.9885904138495364
Accuracy on evaluation data: 9522 / 10000
Epoch 12 training complete
Cost on evaluation data: 0.9722102017908278
Accuracy on evaluation data: 9545 / 10000
Epoch 13 training complete
Cost on evaluation data: 0.9621591912962708
Accuracy on evaluation data: 9563 / 10000
Epoch 14 training complete
Cost on evaluation data: 0.9476244973855212
Accuracy on evaluation data: 9579 / 10000
Epoch 15 training complete
Cost on evaluation data: 0.9530991555732815
Accuracy on evaluation data: 9586 / 10000
Epoch 16 training complete
Cost on evaluation data: 0.9874105395241157
Accuracy on evaluation data: 9529 / 10000
Epoch 17 training complete
Cost on evaluation data: 0.9635348223965305
Accuracy on evaluation data: 9574 / 10000
Epoch 18 training complete
Cost on evaluation data: 0.9587894234915423
Accuracy on evaluation data: 9566 / 10000
Epoch 19 training complete
Cost on evaluation data: 0.9871476367587257
Accuracy on evaluation data: 9540 / 10000
Epoch 20 training complete
Cost on evaluation data: 0.9651627872692474
Accuracy on evaluation data: 9565 / 10000
Epoch 21 training complete
Cost on evaluation data: 0.9679960108134145
Accuracy on evaluation data: 9581 / 10000
Epoch 22 training complete
Cost on evaluation data: 0.9948548867476981
Accuracy on evaluation data: 9527 / 10000
Epoch 23 training complete
Cost on evaluation data: 0.9526303429514341
Accuracy on evaluation data: 9567 / 10000
Epoch 24 training complete
Cost on evaluation data: 0.9680736098211036
Accuracy on evaluation data: 9561 / 10000
Epoch 25 training complete
Cost on evaluation data: 0.9766240326746376
Accuracy on evaluation data: 9553 / 10000
Early stopping after 26 epochs due to no improvement for 10 epochs.
CPU times: user 1min 30s, sys: 37.3 s, total: 2min 7s
Wall time: 1min 34s

%%time
# Initialize and train network --> Relative Improvement Threshold (1%)
net = Network2mat3([784, 30, 10], cost=CrossEntropyCost)
net.large_weight_initializer()
eval_cost_res1p, eval_acc_res1p, train_cost_res1p, train_acc_res1p = net.SGD(
        training_data, 30, 10, 0.5, 5, evaluation_data=test_data,
        monitor_evaluation_cost=True, monitor_evaluation_accuracy=True,
        early_stopping_n=10, relative_improvement_threshold=0.01)

Epoch 0 training complete
Cost on evaluation data: 4.440950566645229
Accuracy on evaluation data: 9146 / 10000
Epoch 1 training complete
Cost on evaluation data: 2.986621201620257
Accuracy on evaluation data: 9310 / 10000
Epoch 2 training complete
Cost on evaluation data: 2.1155207226509374
Accuracy on evaluation data: 9447 / 10000
Epoch 3 training complete
Cost on evaluation data: 1.6304008007274098
Accuracy on evaluation data: 9466 / 10000
Early stopping after 4 epochs due to relative improvement threshold on eval. accuracy: 
0.002011 < 0.01
CPU times: user 15.6 s, sys: 6.39 s, total: 22 s
Wall time: 17.6 s

%%time
# Initialize and train network --> Relative Improvement Threshold (0.05%)
net = Network2mat3([784, 30, 10], cost=CrossEntropyCost)
net.large_weight_initializer()
eval_cost_res2, eval_acc_res2, train_cost_res2, train_acc_res2 = net.SGD(
        training_data, 30, 10, 0.5, 5, evaluation_data=test_data,
        monitor_evaluation_cost=True, monitor_evaluation_accuracy=True,
        early_stopping_n=10, relative_improvement_threshold=0.0005)

Epoch 0 training complete
Cost on evaluation data: 4.308369597153946
Accuracy on evaluation data: 9208 / 10000
Epoch 1 training complete
Cost on evaluation data: 2.899905351508916
Accuracy on evaluation data: 9365 / 10000
Epoch 2 training complete
Cost on evaluation data: 2.099951867450372
Accuracy on evaluation data: 9418 / 10000
Epoch 3 training complete
Cost on evaluation data: 1.6281577096451232
Accuracy on evaluation data: 9452 / 10000
Epoch 4 training complete
Cost on evaluation data: 1.376248076408555
Accuracy on evaluation data: 9434 / 10000
Epoch 5 training complete
Cost on evaluation data: 1.1803890136870558
Accuracy on evaluation data: 9543 / 10000
Epoch 6 training complete
Cost on evaluation data: 1.1531647646034178
Accuracy on evaluation data: 9447 / 10000
Epoch 7 training complete
Cost on evaluation data: 1.0244347291990943
Accuracy on evaluation data: 9581 / 10000
Epoch 8 training complete
Cost on evaluation data: 1.0269805140377313
Accuracy on evaluation data: 9542 / 10000
Epoch 9 training complete
Cost on evaluation data: 0.9781595710679946
Accuracy on evaluation data: 9565 / 10000
Epoch 10 training complete
Cost on evaluation data: 0.9888852968668128
Accuracy on evaluation data: 9568 / 10000
Early stopping after 11 epochs due to relative improvement threshold on eval. accuracy: 
0.000314 < 0.0005
CPU times: user 39.8 s, sys: 16.9 s, total: 56.7 s
Wall time: 42.3 s

%%time
# Initialize and train network --> Relative Improvement Threshold (0.007%)
net = Network2mat3([784, 30, 10], cost=CrossEntropyCost)
net.large_weight_initializer()
eval_cost_res3, eval_acc_res3, train_cost_res3, train_acc_res3 = net.SGD(
        training_data, 30, 10, 0.5, 5, evaluation_data=test_data,
        monitor_evaluation_cost=True, monitor_evaluation_accuracy=True,
        early_stopping_n=10, relative_improvement_threshold=0.00007)

Epoch 0 training complete
Cost on evaluation data: 4.391974338541706
Accuracy on evaluation data: 9122 / 10000
Epoch 1 training complete
Cost on evaluation data: 2.9557418621005933
Accuracy on evaluation data: 9314 / 10000
Epoch 2 training complete
Cost on evaluation data: 2.150090473942308
Accuracy on evaluation data: 9366 / 10000
Epoch 3 training complete
Cost on evaluation data: 1.6292553781691939
Accuracy on evaluation data: 9466 / 10000
Epoch 4 training complete
Cost on evaluation data: 1.3431368103123131
Accuracy on evaluation data: 9520 / 10000
Epoch 5 training complete
Cost on evaluation data: 1.2545291666155989
Accuracy on evaluation data: 9420 / 10000
Epoch 6 training complete
Cost on evaluation data: 1.087030835721143
Accuracy on evaluation data: 9549 / 10000
Epoch 7 training complete
Cost on evaluation data: 1.0558654702260977
Accuracy on evaluation data: 9513 / 10000
Epoch 8 training complete
Cost on evaluation data: 0.9914709204307675
Accuracy on evaluation data: 9580 / 10000
Epoch 9 training complete
Cost on evaluation data: 0.984261582010804
Accuracy on evaluation data: 9577 / 10000
Epoch 10 training complete
Cost on evaluation data: 1.0152498009758495
Accuracy on evaluation data: 9519 / 10000
Epoch 11 training complete
Cost on evaluation data: 1.0011863209370127
Accuracy on evaluation data: 9491 / 10000
Epoch 12 training complete
Cost on evaluation data: 0.9529485729762095
Accuracy on evaluation data: 9597 / 10000
Epoch 13 training complete
Cost on evaluation data: 1.0226212711870284
Accuracy on evaluation data: 9471 / 10000
Epoch 14 training complete
Cost on evaluation data: 0.9652110157070748
Accuracy on evaluation data: 9560 / 10000
Epoch 15 training complete
Cost on evaluation data: 0.9512892198687124
Accuracy on evaluation data: 9582 / 10000
Epoch 16 training complete
Cost on evaluation data: 0.9724898688241087
Accuracy on evaluation data: 9570 / 10000
Epoch 17 training complete
Cost on evaluation data: 0.952412282689242
Accuracy on evaluation data: 9577 / 10000
Epoch 18 training complete
Cost on evaluation data: 0.9508221671184987
Accuracy on evaluation data: 9585 / 10000
Epoch 19 training complete
Cost on evaluation data: 0.9813257400747502
Accuracy on evaluation data: 9550 / 10000
Epoch 20 training complete
Cost on evaluation data: 0.9750242439804422
Accuracy on evaluation data: 9561 / 10000
Epoch 21 training complete
Cost on evaluation data: 0.9766279040201181
Accuracy on evaluation data: 9537 / 10000
Epoch 22 training complete
Cost on evaluation data: 0.9717542415689987
Accuracy on evaluation data: 9559 / 10000
Early stopping after 23 epochs due to no improvement for 10 epochs.
CPU times: user 1min 23s, sys: 34.4 s, total: 1min 57s
Wall time: 1min 34s

epochs_n10 = list(range(1, len(eval_acc_n10) + 1))
epochs_res1p = list(range(1, len(eval_acc_res1p) + 1))
epochs_res2 = list(range(1, len(eval_acc_res2) + 1))
epochs_res3 = list(range(1, len(eval_acc_res3) + 1))

# Plotting the accuracies
plt.figure(figsize=(10, 6))
plt.plot(epochs_n10, eval_acc_n10, label="No Improvement in 10 Epochs", marker='o')
plt.plot(epochs_res1p, eval_acc_res1p, label="Relative Improvement Threshold (1%)", marker='s')
plt.plot(epochs_res2, eval_acc_res2, label="Relative Improvement Threshold (0.05%)", marker='s')
plt.plot(epochs_res3, eval_acc_res3, label="Relative Improvement Threshold (0.007%)", marker='s')

# Labels and title
plt.xlabel("Epochs")
plt.ylabel("Validation Accuracy")
plt.title("Validation Accuracies: No Improvement in 10 vs. Relative Improvement Threshold (0.007%, 0.05%, 1%)")
plt.legend()

# Show the plot
plt.grid(True)
plt.show()

import numpy as np
import matplotlib.pyplot as plt

z = np.arange(-2, 2, .1)
zero = np.zeros(len(z))
y = np.max([zero, z], axis=0)

fig = plt.figure()
ax = fig.add_subplot(111)
ax.plot(z, y)
ax.set_ylim([-2.0, 2.0])
ax.set_xlim([-2.0, 2.0])
ax.grid(True)
ax.set_xlabel('z')
ax.set_title('Rectified linear unit')

plt.show()

z = np.arange(-5, 5, .1)
sigma_fn = np.vectorize(lambda z: 1/(1+np.exp(-z)))
sigma = sigma_fn(z)

fig = plt.figure()
ax = fig.add_subplot(111)
ax.plot(z, sigma)
ax.set_ylim([-0.5, 1.5])
ax.set_xlim([-5,5])
ax.grid(True)
ax.set_xlabel('z')
ax.set_title('sigmoid function')

plt.show()

z = np.arange(-5, 5, .02)
step_fn = np.vectorize(lambda z: 1.0 if z >= 0.0 else 0.0)
step = step_fn(z)

fig = plt.figure()
ax = fig.add_subplot(111)
ax.plot(z, step)
ax.set_ylim([-0.5, 1.5])
ax.set_xlim([-5,5])
ax.grid(True)
ax.set_xlabel('z')
ax.set_title('step function')

plt.show()

z = np.arange(-5, 5, .1)
t = np.tanh(z)

fig = plt.figure()
ax = fig.add_subplot(111)
ax.plot(z, t)
ax.set_ylim([-1.0, 1.0])
ax.set_xlim([-5,5])
ax.grid(True)
ax.set_xlabel('z')
ax.set_title('tanh function')

plt.show()

----------------------------------------------------------------------------¶

Baseline Backprop (n_hidden = 30): 94.96% Accuracy¶

Fully Matrix-Based Backprop (n_hidden = 30): 95.50% Accuracy¶

Baseline Backprop (n_hidden = 100): 96.51% Accuracy¶

Fully Matrix-Based Backprop (n_hidden = 100): 96.79% Accuracy¶

Overfitting¶

Overfitting: 84.64% Accuracy¶

Regularization¶

Why does regularization help reduce overfitting?¶

Other Regularization Techniques: L1 Reg., Dropout, Artificial Expansion of Training Data¶

Weight Initialization¶

Fully Matrix-Based Backprop (n_hidden = 30) with L1 regularization: 95.76% Accuracy¶

How to choose a NN's hyperparameters?¶

Guide to Choosing Neural Network Hyper-parameters¶

Introduction¶

Broad Strategy¶

Experimental Approach¶

Key Hyper-parameters to Focus On¶

Tips¶

Conclusion¶

Other techniques to improve NNs¶

Variations On SGD¶

Other models of artificial neuron¶

On stories in neural networks¶

Problems¶

Problem 1:¶

Answer 1:¶

(a) Step Functions in Arbitrary Directions¶

(b) Circular Approximation Using Step Functions¶

(c) Approximation of Arbitrary Functions Using Circular Towers¶

Conclusion¶

Answer 2:¶

Proof Sketch: Universal Approximation with Single Hidden Layer¶

(a) Step Functions in Arbitrary Directions¶

(b) Approximating Circular Tower Functions¶

(c) Approximating Arbitrary Functions¶

Problem 2:¶

Answer:¶

Why Linear Neurons Don't Satisfy Universality¶

Proof: Linear Networks Are Equivalent to a Single Linear Layer¶

Single Layer of Linear Neurons¶

Stacking Multiple Layers of Linear Neurons¶

Consequence: No Increase in Expressive Power¶

Conclusion: No Universal Computation with Linear Neurons¶

Problem 3:¶

Answer 1:¶

ReLU Universality in Neural Networks¶

Part 1: Why ReLUs Don’t Satisfy Previous Conditions¶

Part 2: Proof of Universality for ReLUs¶

Answer 2:¶

Step 1: Define the ReLU Function¶

Step 4: Error Bound and Universality¶

Step 5: Generalization to Higher Dimensions¶

Conclusion¶

Baseline Backprop (`n_hidden = 30`): `94.96%` Accuracy¶

Fully Matrix-Based Backprop (`n_hidden = 30`): `95.50%` Accuracy¶

Baseline Backprop (`n_hidden = 100`): `96.51%` Accuracy¶

Fully Matrix-Based Backprop (`n_hidden` = 100): `96.79%` Accuracy¶

Overfitting: `84.64%` Accuracy¶

Fully Matrix-Based Backprop (`n_hidden` = 30) with L1 regularization: `95.76%` Accuracy¶