dlnd_image_classification


Image Classification

In this project, you'll classify images from the CIFAR-10 dataset. The dataset consists of airplanes, dogs, cats, and other objects. You'll preprocess the images, then train a convolutional neural network on all the samples. The images need to be normalized and the labels need to be one-hot encoded. You'll get to apply what you learned and build a convolutional, max pooling, dropout, and fully connected layers. At the end, you'll get to see your neural network's predictions on the sample images.

Get the Data

Run the following cell to download the CIFAR-10 dataset for python.


In [3]:
"""
DON'T MODIFY ANYTHING IN THIS CELL THAT IS BELOW THIS LINE
"""
from urllib.request import urlretrieve
from os.path import isfile, isdir
from tqdm import tqdm
import problem_unittests as tests
import helper
import tarfile

cifar10_dataset_folder_path = 'cifar-10-batches-py'

# Use Floyd's cifar-10 dataset if present
floyd_cifar10_location = '/input/cifar-10/python.tar.gz'
if isfile(floyd_cifar10_location):
    tar_gz_path = floyd_cifar10_location
else:
    tar_gz_path = 'cifar-10-python.tar.gz'

class DLProgress(tqdm):
    last_block = 0

    def hook(self, block_num=1, block_size=1, total_size=None):
        self.total = total_size
        self.update((block_num - self.last_block) * block_size)
        self.last_block = block_num

if not isfile(tar_gz_path):
    with DLProgress(unit='B', unit_scale=True, miniters=1, desc='CIFAR-10 Dataset') as pbar:
        urlretrieve(
            'https://www.cs.toronto.edu/~kriz/cifar-10-python.tar.gz',
            tar_gz_path,
            pbar.hook)

if not isdir(cifar10_dataset_folder_path):
    with tarfile.open(tar_gz_path) as tar:
        tar.extractall()
        tar.close()


tests.test_folder_path(cifar10_dataset_folder_path)


All files found!

Explore the Data

The dataset is broken into batches to prevent your machine from running out of memory. The CIFAR-10 dataset consists of 5 batches, named data_batch_1, data_batch_2, etc.. Each batch contains the labels and images that are one of the following:

  • airplane
  • automobile
  • bird
  • cat
  • deer
  • dog
  • frog
  • horse
  • ship
  • truck

Understanding a dataset is part of making predictions on the data. Play around with the code cell below by changing the batch_id and sample_id. The batch_id is the id for a batch (1-5). The sample_id is the id for a image and label pair in the batch.

Ask yourself "What are all possible labels?", "What is the range of values for the image data?", "Are the labels in order or random?". Answers to questions like these will help you preprocess the data and end up with better predictions.


In [4]:
%matplotlib inline
%config InlineBackend.figure_format = 'retina'

import helper
import numpy as np

# Explore the dataset
batch_id = 1
sample_id = 5
helper.display_stats(cifar10_dataset_folder_path, batch_id, sample_id)


Stats of batch 1:
Samples: 10000
Label Counts: {0: 1005, 1: 974, 2: 1032, 3: 1016, 4: 999, 5: 937, 6: 1030, 7: 1001, 8: 1025, 9: 981}
First 20 Labels: [6, 9, 9, 4, 1, 1, 2, 7, 8, 3, 4, 7, 7, 2, 9, 9, 9, 3, 2, 6]

Example of Image 5:
Image - Min Value: 0 Max Value: 252
Image - Shape: (32, 32, 3)
Label - Label Id: 1 Name: automobile

Implement Preprocess Functions

Normalize

In the cell below, implement the normalize function to take in image data, x, and return it as a normalized Numpy array. The values should be in the range of 0 to 1, inclusive. The return object should be the same shape as x.


In [6]:
def normalize(x):
    """
    Normalize a list of sample image data in the range of 0 to 1
    : x: List of image data.  The image shape is (32, 32, 3)
    : return: Numpy array of normalize data
    """
    import numpy as np
    array = np.array(x) # Transform list of lists x into Numpy ndarray array
    array_min = array.min()
    return (array - array_min) / (array.max() - array_min) # Normalize array data (alternatively use sklearn's Normalizer)

"""
DON'T MODIFY ANYTHING IN THIS CELL THAT IS BELOW THIS LINE
"""
tests.test_normalize(normalize)


Tests Passed

One-hot encode

Just like the previous code cell, you'll be implementing a function for preprocessing. This time, you'll implement the one_hot_encode function. The input, x, are a list of labels. Implement the function to return the list of labels as One-Hot encoded Numpy array. The possible values for labels are 0 to 9. The one-hot encoding function should return the same encoding for each value between each call to one_hot_encode. Make sure to save the map of encodings outside the function.

Hint: Don't reinvent the wheel.


In [197]:
def one_hot_encode(x):
    """
    One hot encode a list of sample labels. Return a one-hot encoded vector for each label.
    : x: List of sample Labels
    : return: Numpy array of one-hot encoded labels
    """
    import numpy as np
    n_classes = 10 # Number of possible label values given fixed as 0 to 9
    return np.eye(n_classes)[x] # Return array of rows of identity matrix as given by x

"""
DON'T MODIFY ANYTHING IN THIS CELL THAT IS BELOW THIS LINE
"""
tests.test_one_hot_encode(one_hot_encode)


Tests Passed

Randomize Data

As you saw from exploring the data above, the order of the samples are randomized. It doesn't hurt to randomize it again, but you don't need to for this dataset.

Preprocess all the data and save it

Running the code cell below will preprocess all the CIFAR-10 data and save it to file. The code below also uses 10% of the training data for validation.


In [198]:
"""
DON'T MODIFY ANYTHING IN THIS CELL
"""
# Preprocess Training, Validation, and Testing Data
helper.preprocess_and_save_data(cifar10_dataset_folder_path, normalize, one_hot_encode)

Check Point

This is your first checkpoint. If you ever decide to come back to this notebook or have to restart the notebook, you can start from here. The preprocessed data has been saved to disk.


In [199]:
"""
DON'T MODIFY ANYTHING IN THIS CELL
"""
import pickle
import problem_unittests as tests
import helper

# Load the Preprocessed Validation data
valid_features, valid_labels = pickle.load(open('preprocess_validation.p', mode='rb'))

Build the network

For the neural network, you'll build each layer into a function. Most of the code you've seen has been outside of functions. To test your code more thoroughly, we require that you put each layer in a function. This allows us to give you better feedback and test for simple mistakes using our unittests before you submit your project.

Note: If you're finding it hard to dedicate enough time for this course each week, we've provided a small shortcut to this part of the project. In the next couple of problems, you'll have the option to use classes from the TensorFlow Layers or TensorFlow Layers (contrib) packages to build each layer, except the layers you build in the "Convolutional and Max Pooling Layer" section. TF Layers is similar to Keras's and TFLearn's abstraction to layers, so it's easy to pickup.

However, if you would like to get the most out of this course, try to solve all the problems without using anything from the TF Layers packages. You can still use classes from other packages that happen to have the same name as ones you find in TF Layers! For example, instead of using the TF Layers version of the conv2d class, tf.layers.conv2d, you would want to use the TF Neural Network version of conv2d, tf.nn.conv2d.

Let's begin!

Input

The neural network needs to read the image data, one-hot encoded labels, and dropout keep probability. Implement the following functions

  • Implement neural_net_image_input
    • Return a TF Placeholder
    • Set the shape using image_shape with batch size set to None.
    • Name the TensorFlow placeholder "x" using the TensorFlow name parameter in the TF Placeholder.
  • Implement neural_net_label_input
    • Return a TF Placeholder
    • Set the shape using n_classes with batch size set to None.
    • Name the TensorFlow placeholder "y" using the TensorFlow name parameter in the TF Placeholder.
  • Implement neural_net_keep_prob_input
    • Return a TF Placeholder for dropout keep probability.
    • Name the TensorFlow placeholder "keep_prob" using the TensorFlow name parameter in the TF Placeholder.

These names will be used at the end of the project to load your saved model.

Note: None for shapes in TensorFlow allow for a dynamic size.


In [8]:
import tensorflow as tf
tf.reset_default_graph()


def neural_net_image_input(image_shape):
    """
    Return a Tensor for a batch of image input
    : image_shape: Shape of the images
    : return: Tensor for image input.
    """
    return tf.placeholder(tf.float32, shape=(None, *image_shape), name='x')


def neural_net_label_input(n_classes):
    """
    Return a Tensor for a batch of label input
    : n_classes: Number of classes
    : return: Tensor for label input.
    """
    return tf.placeholder(tf.int32, shape=(None, n_classes), name='y')


def neural_net_keep_prob_input():
    """
    Return a Tensor for keep probability
    : return: Tensor for keep probability.
    """
    return tf.placeholder(tf.float32, name='keep_prob')


"""
DON'T MODIFY ANYTHING IN THIS CELL THAT IS BELOW THIS LINE
"""
tests.test_nn_image_inputs(neural_net_image_input)
tests.test_nn_label_inputs(neural_net_label_input)
tests.test_nn_keep_prob_inputs(neural_net_keep_prob_input)


Image Input Tests Passed.
Label Input Tests Passed.
Keep Prob Tests Passed.

Convolution and Max Pooling Layer

Convolution layers have a lot of success with images. For this code cell, you should implement the function conv2d_maxpool to apply convolution then max pooling:

  • Create the weight and bias using conv_ksize, conv_num_outputs and the shape of x_tensor.
  • Apply a convolution to x_tensor using weight and conv_strides.
    • We recommend you use same padding, but you're welcome to use any padding.
  • Add bias
  • Add a nonlinear activation to the convolution.
  • Apply Max Pooling using pool_ksize and pool_strides.
    • We recommend you use same padding, but you're welcome to use any padding.

Note: You can't use TensorFlow Layers or TensorFlow Layers (contrib) for this layer, but you can still use TensorFlow's Neural Network package. You may still use the shortcut option for all the other layers.


In [9]:
def conv2d_maxpool(x_tensor, conv_num_outputs, conv_ksize, conv_strides, pool_ksize, pool_strides):
    """
    Apply convolution then max pooling to x_tensor
    :param x_tensor: TensorFlow Tensor
    :param conv_num_outputs: Number of outputs for the convolutional layer
    :param conv_ksize: kernel size 2-D Tuple for the convolutional layer
    :param conv_strides: Stride 2-D Tuple for convolution
    :param pool_ksize: kernel size 2-D Tuple for pool
    :param pool_strides: Stride 2-D Tuple for pool
    : return: A tensor that represents convolution and max pooling of x_tensor
    """
    conv_filter = tf.Variable(tf.truncated_normal([*conv_ksize, x_tensor.get_shape().as_list()[-1], conv_num_outputs],  mean=0.0, stddev=0.1, dtype=tf.float32))
    conv_bias = tf.Variable(tf.constant(0.1, shape=[conv_num_outputs]))
    conv_strides = [1, *conv_strides, 1]
    mpool_ksize = [1, *pool_ksize, 1]
    mpool_strides = [1, *pool_strides, 1]
    x_tensor = tf.nn.conv2d(x_tensor, conv_filter, conv_strides, padding='SAME')
    x_tensor = tf.nn.bias_add(x_tensor, conv_bias)
    x_tensor = tf.nn.relu(x_tensor)
    x_tensor = tf.nn.max_pool(x_tensor, mpool_ksize, mpool_strides, padding='SAME')
    return x_tensor


"""
DON'T MODIFY ANYTHING IN THIS CELL THAT IS BELOW THIS LINE
"""
tests.test_con_pool(conv2d_maxpool)


Tests Passed

Flatten Layer

Implement the flatten function to change the dimension of x_tensor from a 4-D tensor to a 2-D tensor. The output should be the shape (Batch Size, Flattened Image Size). Shortcut option: you can use classes from the TensorFlow Layers or TensorFlow Layers (contrib) packages for this layer. For more of a challenge, only use other TensorFlow packages.


In [202]:
def flatten(x_tensor):
    """
    Flatten x_tensor to (Batch Size, Flattened Image Size)
    : x_tensor: A tensor of size (Batch Size, ...), where ... are the image dimensions.
    : return: A tensor of size (Batch Size, Flattened Image Size).
    """
    flat_size = 1
    for i in x_tensor.get_shape().as_list()[1:] : flat_size *= i 
    x_tensor = tf.reshape(x_tensor, [-1, flat_size])
    return x_tensor


"""
DON'T MODIFY ANYTHING IN THIS CELL THAT IS BELOW THIS LINE
"""
tests.test_flatten(flatten)


Tests Passed

Fully-Connected Layer

Implement the fully_conn function to apply a fully connected layer to x_tensor with the shape (Batch Size, num_outputs). Shortcut option: you can use classes from the TensorFlow Layers or TensorFlow Layers (contrib) packages for this layer. For more of a challenge, only use other TensorFlow packages.


In [13]:
def fully_conn(x_tensor, num_outputs):
    """
    Apply a fully connected layer to x_tensor using weight and bias
    : x_tensor: A 2-D tensor where the first dimension is batch size.
    : num_outputs: The number of output that the new tensor should be.
    : return: A 2-D tensor where the second dimension is num_outputs.
    """
    weights = tf.Variable(tf.truncated_normal([x_tensor.get_shape().as_list()[1], num_outputs], mean=0.0, stddev=0.1, dtype=tf.float32))
    biases = tf.Variable(tf.ones([num_outputs]) / 10)
    x_tensor = tf.add(tf.matmul(x_tensor, weights), biases)
    return x_tensor


"""
DON'T MODIFY ANYTHING IN THIS CELL THAT IS BELOW THIS LINE
"""
tests.test_fully_conn(fully_conn)


Tests Passed

Output Layer

Implement the output function to apply a fully connected layer to x_tensor with the shape (Batch Size, num_outputs). Shortcut option: you can use classes from the TensorFlow Layers or TensorFlow Layers (contrib) packages for this layer. For more of a challenge, only use other TensorFlow packages.

Note: Activation, softmax, or cross entropy should not be applied to this.


In [11]:
def output(x_tensor, num_outputs):
    """
    Apply a output layer to x_tensor using weight and bias
    : x_tensor: A 2-D tensor where the first dimension is batch size.
    : num_outputs: The number of output that the new tensor should be.
    : return: A 2-D tensor where the second dimension is num_outputs.
    """
    weights = tf.Variable(tf.truncated_normal([x_tensor.get_shape().as_list()[1], num_outputs], mean=0.0, stddev=0.1, dtype=tf.float32))
    biases = tf.Variable(tf.ones([num_outputs]) / 10)
    y = tf.add(tf.matmul(x_tensor, weights), biases)
    return y


"""
DON'T MODIFY ANYTHING IN THIS CELL THAT IS BELOW THIS LINE
"""
tests.test_output(output)


Tests Passed

Create Convolutional Model

Implement the function conv_net to create a convolutional neural network model. The function takes in a batch of images, x, and outputs logits. Use the layers you created above to create this model:

  • Apply 1, 2, or 3 Convolution and Max Pool layers
  • Apply a Flatten Layer
  • Apply 1, 2, or 3 Fully Connected Layers
  • Apply an Output Layer
  • Return the output
  • Apply TensorFlow's Dropout to one or more layers in the model using keep_prob.

In [205]:
def conv_net(x, keep_prob):
    """
    Create a convolutional neural network model
    : x: Placeholder tensor that holds image data.
    : keep_prob: Placeholder tensor that hold dropout keep probability.
    : return: Tensor that represents logits
    """
    # TODO: Apply 1, 2, or 3 Convolution and Max Pool layers
    #    Play around with different number of outputs, kernel size and stride
    # Function Definition from Above:
    #    conv2d_maxpool(x_tensor, conv_num_outputs, conv_ksize, conv_strides, pool_ksize, pool_strides)
    
    x = conv2d_maxpool(x,  32, (5, 5), (1, 1), (2, 2), (2, 2))
    x = conv2d_maxpool(x,  64, (5, 5), (1, 1), (2, 2), (2, 2))
    x = conv2d_maxpool(x, 128, (3, 3), (1, 1), (2, 2), (2, 2))

    # TODO: Apply a Flatten Layer
    # Function Definition from Above:
    #   flatten(x_tensor)
    
    x = flatten(x)

    # TODO: Apply 1, 2, or 3 Fully Connected Layers
    #    Play around with different number of outputs
    # Function Definition from Above:
    #   fully_conn(x_tensor, num_outputs)
    
    x = fully_conn(x, 1024)
    x = tf.nn.dropout(x, keep_prob)
    x = fully_conn(x, 256)
    x = tf.nn.dropout(x, keep_prob)
    x = fully_conn(x, 64)
    x = tf.nn.dropout(x, keep_prob)

    
    # TODO: Apply an Output Layer
    #    Set this to the number of classes
    # Function Definition from Above:
    #   output(x_tensor, num_outputs)
    
    y = output(x, 10)
    
    # TODO: return output
    return y


"""
DON'T MODIFY ANYTHING IN THIS CELL THAT IS BELOW THIS LINE
"""

##############################
## Build the Neural Network ##
##############################

# Remove previous weights, bias, inputs, etc..
tf.reset_default_graph()

# Inputs
x = neural_net_image_input((32, 32, 3))
y = neural_net_label_input(10)
keep_prob = neural_net_keep_prob_input()

# Model
logits = conv_net(x, keep_prob)

# Name logits Tensor, so that is can be loaded from disk after training
logits = tf.identity(logits, name='logits')

# Loss and Optimizer
cost = tf.reduce_mean(tf.nn.softmax_cross_entropy_with_logits(logits=logits, labels=y))
optimizer = tf.train.AdamOptimizer().minimize(cost)

# Accuracy
correct_pred = tf.equal(tf.argmax(logits, 1), tf.argmax(y, 1))
accuracy = tf.reduce_mean(tf.cast(correct_pred, tf.float32), name='accuracy')

tests.test_conv_net(conv_net)


Neural Network Built!

Train the Neural Network

Single Optimization

Implement the function train_neural_network to do a single optimization. The optimization should use optimizer to optimize in session with a feed_dict of the following:

  • x for image input
  • y for labels
  • keep_prob for keep probability for dropout

This function will be called for each batch, so tf.global_variables_initializer() has already been called.

Note: Nothing needs to be returned. This function is only optimizing the neural network.


In [206]:
def train_neural_network(session, optimizer, keep_probability, feature_batch, label_batch):
    """
    Optimize the session on a batch of images and labels
    : session: Current TensorFlow session
    : optimizer: TensorFlow optimizer function
    : keep_probability: keep probability
    : feature_batch: Batch of Numpy image data
    : label_batch: Batch of Numpy label data
    """
    session.run(optimizer, feed_dict={x: feature_batch, y: label_batch, keep_prob: keep_probability})
    
"""
DON'T MODIFY ANYTHING IN THIS CELL THAT IS BELOW THIS LINE
"""
tests.test_train_nn(train_neural_network)


Tests Passed

Show Stats

Implement the function print_stats to print loss and validation accuracy. Use the global variables valid_features and valid_labels to calculate validation accuracy. Use a keep probability of 1.0 to calculate the loss and validation accuracy.


In [207]:
def print_stats(session, feature_batch, label_batch, cost, accuracy):
    """
    Print information about loss and validation accuracy
    : session: Current TensorFlow session
    : feature_batch: Batch of Numpy image data
    : label_batch: Batch of Numpy label data
    : cost: TensorFlow cost function
    : accuracy: TensorFlow accuracy function
    """
    # TODO: Implement Function
    train_accur, loss = session.run([accuracy, cost], feed_dict={x: feature_batch, y: label_batch, keep_prob: 1.})
    print("Training Accuracy= " + "{:.4f}".format(train_accur) + ", Batch Loss= " + "{:.2f}".format(loss))
    valid_accur = session.run(accuracy, feed_dict={x: valid_features, y: valid_labels, keep_prob: 1.})
    print("Total Validation Accuracy= " + "{:.4f}".format(valid_accur))

Hyperparameters

Tune the following parameters:

  • Set epochs to the number of iterations until the network stops learning or start overfitting
  • Set batch_size to the highest number that your machine has memory for. Most people set them to common sizes of memory:
    • 64
    • 128
    • 256
    • ...
  • Set keep_probability to the probability of keeping a node using dropout

In [208]:
# TODO: Tune Parameters
epochs = 30
batch_size = 256
keep_probability = .6

Train on a Single CIFAR-10 Batch

Instead of training the neural network on all the CIFAR-10 batches of data, let's use a single batch. This should save time while you iterate on the model to get a better accuracy. Once the final validation accuracy is 50% or greater, run the model on all the data in the next section.


In [209]:
"""
DON'T MODIFY ANYTHING IN THIS CELL
"""
print('Checking the Training on a Single Batch...')
with tf.Session() as sess:
    # Initializing the variables
    sess.run(tf.global_variables_initializer())
    
    # Training cycle
    for epoch in range(epochs):
        batch_i = 1
        for batch_features, batch_labels in helper.load_preprocess_training_batch(batch_i, batch_size):
            train_neural_network(sess, optimizer, keep_probability, batch_features, batch_labels)
        print('Epoch {:>2}, CIFAR-10 Batch {}:  '.format(epoch + 1, batch_i), end='')
        print_stats(sess, batch_features, batch_labels, cost, accuracy)


Checking the Training on a Single Batch...
Epoch  1, CIFAR-10 Batch 1:  Training Accuracy= 0.1000, Batch Loss= 2.29
Total Validation Accuracy= 0.1716
Epoch  2, CIFAR-10 Batch 1:  Training Accuracy= 0.1750, Batch Loss= 2.18
Total Validation Accuracy= 0.1886
Epoch  3, CIFAR-10 Batch 1:  Training Accuracy= 0.4250, Batch Loss= 1.98
Total Validation Accuracy= 0.3066
Epoch  4, CIFAR-10 Batch 1:  Training Accuracy= 0.4000, Batch Loss= 1.87
Total Validation Accuracy= 0.3446
Epoch  5, CIFAR-10 Batch 1:  Training Accuracy= 0.4250, Batch Loss= 1.80
Total Validation Accuracy= 0.3642
Epoch  6, CIFAR-10 Batch 1:  Training Accuracy= 0.4500, Batch Loss= 1.66
Total Validation Accuracy= 0.3966
Epoch  7, CIFAR-10 Batch 1:  Training Accuracy= 0.5250, Batch Loss= 1.46
Total Validation Accuracy= 0.4242
Epoch  8, CIFAR-10 Batch 1:  Training Accuracy= 0.5750, Batch Loss= 1.30
Total Validation Accuracy= 0.4580
Epoch  9, CIFAR-10 Batch 1:  Training Accuracy= 0.6000, Batch Loss= 1.18
Total Validation Accuracy= 0.4702
Epoch 10, CIFAR-10 Batch 1:  Training Accuracy= 0.7000, Batch Loss= 1.06
Total Validation Accuracy= 0.4632
Epoch 11, CIFAR-10 Batch 1:  Training Accuracy= 0.7000, Batch Loss= 0.92
Total Validation Accuracy= 0.4776
Epoch 12, CIFAR-10 Batch 1:  Training Accuracy= 0.7750, Batch Loss= 0.76
Total Validation Accuracy= 0.5086
Epoch 13, CIFAR-10 Batch 1:  Training Accuracy= 0.8250, Batch Loss= 0.68
Total Validation Accuracy= 0.4970
Epoch 14, CIFAR-10 Batch 1:  Training Accuracy= 0.8250, Batch Loss= 0.59
Total Validation Accuracy= 0.5192
Epoch 15, CIFAR-10 Batch 1:  Training Accuracy= 0.8250, Batch Loss= 0.50
Total Validation Accuracy= 0.5358
Epoch 16, CIFAR-10 Batch 1:  Training Accuracy= 0.9000, Batch Loss= 0.41
Total Validation Accuracy= 0.5302
Epoch 17, CIFAR-10 Batch 1:  Training Accuracy= 0.8750, Batch Loss= 0.43
Total Validation Accuracy= 0.5160
Epoch 18, CIFAR-10 Batch 1:  Training Accuracy= 0.9000, Batch Loss= 0.28
Total Validation Accuracy= 0.5400
Epoch 19, CIFAR-10 Batch 1:  Training Accuracy= 0.9250, Batch Loss= 0.29
Total Validation Accuracy= 0.5344
Epoch 20, CIFAR-10 Batch 1:  Training Accuracy= 0.9500, Batch Loss= 0.25
Total Validation Accuracy= 0.5680
Epoch 21, CIFAR-10 Batch 1:  Training Accuracy= 0.9500, Batch Loss= 0.22
Total Validation Accuracy= 0.5534
Epoch 22, CIFAR-10 Batch 1:  Training Accuracy= 0.9500, Batch Loss= 0.20
Total Validation Accuracy= 0.5564
Epoch 23, CIFAR-10 Batch 1:  Training Accuracy= 0.9250, Batch Loss= 0.18
Total Validation Accuracy= 0.5612
Epoch 24, CIFAR-10 Batch 1:  Training Accuracy= 0.9750, Batch Loss= 0.13
Total Validation Accuracy= 0.5518
Epoch 25, CIFAR-10 Batch 1:  Training Accuracy= 0.9750, Batch Loss= 0.12
Total Validation Accuracy= 0.5432
Epoch 26, CIFAR-10 Batch 1:  Training Accuracy= 0.9750, Batch Loss= 0.09
Total Validation Accuracy= 0.5546
Epoch 27, CIFAR-10 Batch 1:  Training Accuracy= 1.0000, Batch Loss= 0.05
Total Validation Accuracy= 0.5570
Epoch 28, CIFAR-10 Batch 1:  Training Accuracy= 1.0000, Batch Loss= 0.04
Total Validation Accuracy= 0.5540
Epoch 29, CIFAR-10 Batch 1:  Training Accuracy= 1.0000, Batch Loss= 0.05
Total Validation Accuracy= 0.5466
Epoch 30, CIFAR-10 Batch 1:  Training Accuracy= 1.0000, Batch Loss= 0.04
Total Validation Accuracy= 0.5702

Fully Train the Model

Now that you got a good accuracy with a single CIFAR-10 batch, try it with all five batches.


In [210]:
"""
DON'T MODIFY ANYTHING IN THIS CELL
"""
save_model_path = './image_classification'

print('Training...')
with tf.Session() as sess:
    # Initializing the variables
    sess.run(tf.global_variables_initializer())
    
    # Training cycle
    for epoch in range(epochs):
        # Loop over all batches
        n_batches = 5
        for batch_i in range(1, n_batches + 1):
            for batch_features, batch_labels in helper.load_preprocess_training_batch(batch_i, batch_size):
                train_neural_network(sess, optimizer, keep_probability, batch_features, batch_labels)
            print('Epoch {:>2}, CIFAR-10 Batch {}:  '.format(epoch + 1, batch_i), end='')
            print_stats(sess, batch_features, batch_labels, cost, accuracy)
            
    # Save Model
    saver = tf.train.Saver()
    save_path = saver.save(sess, save_model_path)


Training...
Epoch  1, CIFAR-10 Batch 1:  Training Accuracy= 0.1250, Batch Loss= 2.23
Total Validation Accuracy= 0.1500
Epoch  1, CIFAR-10 Batch 2:  Training Accuracy= 0.2250, Batch Loss= 2.20
Total Validation Accuracy= 0.2050
Epoch  1, CIFAR-10 Batch 3:  Training Accuracy= 0.3250, Batch Loss= 1.95
Total Validation Accuracy= 0.3052
Epoch  1, CIFAR-10 Batch 4:  Training Accuracy= 0.4000, Batch Loss= 1.77
Total Validation Accuracy= 0.3192
Epoch  1, CIFAR-10 Batch 5:  Training Accuracy= 0.3750, Batch Loss= 1.79
Total Validation Accuracy= 0.3754
Epoch  2, CIFAR-10 Batch 1:  Training Accuracy= 0.5250, Batch Loss= 1.75
Total Validation Accuracy= 0.4082
Epoch  2, CIFAR-10 Batch 2:  Training Accuracy= 0.4250, Batch Loss= 1.73
Total Validation Accuracy= 0.4154
Epoch  2, CIFAR-10 Batch 3:  Training Accuracy= 0.5250, Batch Loss= 1.25
Total Validation Accuracy= 0.4350
Epoch  2, CIFAR-10 Batch 4:  Training Accuracy= 0.4250, Batch Loss= 1.45
Total Validation Accuracy= 0.4572
Epoch  2, CIFAR-10 Batch 5:  Training Accuracy= 0.4250, Batch Loss= 1.50
Total Validation Accuracy= 0.4736
Epoch  3, CIFAR-10 Batch 1:  Training Accuracy= 0.4250, Batch Loss= 1.46
Total Validation Accuracy= 0.4784
Epoch  3, CIFAR-10 Batch 2:  Training Accuracy= 0.4500, Batch Loss= 1.35
Total Validation Accuracy= 0.4774
Epoch  3, CIFAR-10 Batch 3:  Training Accuracy= 0.5250, Batch Loss= 1.12
Total Validation Accuracy= 0.4600
Epoch  3, CIFAR-10 Batch 4:  Training Accuracy= 0.6250, Batch Loss= 1.30
Total Validation Accuracy= 0.5056
Epoch  3, CIFAR-10 Batch 5:  Training Accuracy= 0.5750, Batch Loss= 1.25
Total Validation Accuracy= 0.5206
Epoch  4, CIFAR-10 Batch 1:  Training Accuracy= 0.6000, Batch Loss= 1.19
Total Validation Accuracy= 0.5246
Epoch  4, CIFAR-10 Batch 2:  Training Accuracy= 0.6000, Batch Loss= 0.98
Total Validation Accuracy= 0.5444
Epoch  4, CIFAR-10 Batch 3:  Training Accuracy= 0.7000, Batch Loss= 0.89
Total Validation Accuracy= 0.5310
Epoch  4, CIFAR-10 Batch 4:  Training Accuracy= 0.6000, Batch Loss= 1.04
Total Validation Accuracy= 0.5670
Epoch  4, CIFAR-10 Batch 5:  Training Accuracy= 0.6250, Batch Loss= 1.06
Total Validation Accuracy= 0.5636
Epoch  5, CIFAR-10 Batch 1:  Training Accuracy= 0.6750, Batch Loss= 0.97
Total Validation Accuracy= 0.5690
Epoch  5, CIFAR-10 Batch 2:  Training Accuracy= 0.6250, Batch Loss= 0.83
Total Validation Accuracy= 0.5494
Epoch  5, CIFAR-10 Batch 3:  Training Accuracy= 0.8000, Batch Loss= 0.67
Total Validation Accuracy= 0.5758
Epoch  5, CIFAR-10 Batch 4:  Training Accuracy= 0.7000, Batch Loss= 0.84
Total Validation Accuracy= 0.5908
Epoch  5, CIFAR-10 Batch 5:  Training Accuracy= 0.7500, Batch Loss= 0.84
Total Validation Accuracy= 0.5986
Epoch  6, CIFAR-10 Batch 1:  Training Accuracy= 0.7000, Batch Loss= 0.76
Total Validation Accuracy= 0.6106
Epoch  6, CIFAR-10 Batch 2:  Training Accuracy= 0.7000, Batch Loss= 0.74
Total Validation Accuracy= 0.5966
Epoch  6, CIFAR-10 Batch 3:  Training Accuracy= 0.8500, Batch Loss= 0.59
Total Validation Accuracy= 0.5972
Epoch  6, CIFAR-10 Batch 4:  Training Accuracy= 0.7750, Batch Loss= 0.72
Total Validation Accuracy= 0.6108
Epoch  6, CIFAR-10 Batch 5:  Training Accuracy= 0.8750, Batch Loss= 0.61
Total Validation Accuracy= 0.6128
Epoch  7, CIFAR-10 Batch 1:  Training Accuracy= 0.7000, Batch Loss= 0.67
Total Validation Accuracy= 0.6092
Epoch  7, CIFAR-10 Batch 2:  Training Accuracy= 0.8000, Batch Loss= 0.60
Total Validation Accuracy= 0.6088
Epoch  7, CIFAR-10 Batch 3:  Training Accuracy= 0.9000, Batch Loss= 0.45
Total Validation Accuracy= 0.6288
Epoch  7, CIFAR-10 Batch 4:  Training Accuracy= 0.8250, Batch Loss= 0.60
Total Validation Accuracy= 0.6228
Epoch  7, CIFAR-10 Batch 5:  Training Accuracy= 0.9000, Batch Loss= 0.48
Total Validation Accuracy= 0.6322
Epoch  8, CIFAR-10 Batch 1:  Training Accuracy= 0.8250, Batch Loss= 0.54
Total Validation Accuracy= 0.6324
Epoch  8, CIFAR-10 Batch 2:  Training Accuracy= 0.8500, Batch Loss= 0.42
Total Validation Accuracy= 0.6360
Epoch  8, CIFAR-10 Batch 3:  Training Accuracy= 0.9750, Batch Loss= 0.40
Total Validation Accuracy= 0.6308
Epoch  8, CIFAR-10 Batch 4:  Training Accuracy= 0.9000, Batch Loss= 0.48
Total Validation Accuracy= 0.6380
Epoch  8, CIFAR-10 Batch 5:  Training Accuracy= 0.9500, Batch Loss= 0.30
Total Validation Accuracy= 0.6500
Epoch  9, CIFAR-10 Batch 1:  Training Accuracy= 0.8500, Batch Loss= 0.47
Total Validation Accuracy= 0.6410
Epoch  9, CIFAR-10 Batch 2:  Training Accuracy= 0.8750, Batch Loss= 0.36
Total Validation Accuracy= 0.6428
Epoch  9, CIFAR-10 Batch 3:  Training Accuracy= 0.9500, Batch Loss= 0.24
Total Validation Accuracy= 0.6618
Epoch  9, CIFAR-10 Batch 4:  Training Accuracy= 0.8750, Batch Loss= 0.38
Total Validation Accuracy= 0.6556
Epoch  9, CIFAR-10 Batch 5:  Training Accuracy= 0.9500, Batch Loss= 0.26
Total Validation Accuracy= 0.6608
Epoch 10, CIFAR-10 Batch 1:  Training Accuracy= 0.9500, Batch Loss= 0.36
Total Validation Accuracy= 0.6624
Epoch 10, CIFAR-10 Batch 2:  Training Accuracy= 0.8750, Batch Loss= 0.27
Total Validation Accuracy= 0.6608
Epoch 10, CIFAR-10 Batch 3:  Training Accuracy= 0.9500, Batch Loss= 0.20
Total Validation Accuracy= 0.6672
Epoch 10, CIFAR-10 Batch 4:  Training Accuracy= 0.9500, Batch Loss= 0.31
Total Validation Accuracy= 0.6600
Epoch 10, CIFAR-10 Batch 5:  Training Accuracy= 0.9500, Batch Loss= 0.24
Total Validation Accuracy= 0.6658
Epoch 11, CIFAR-10 Batch 1:  Training Accuracy= 0.9250, Batch Loss= 0.32
Total Validation Accuracy= 0.6422
Epoch 11, CIFAR-10 Batch 2:  Training Accuracy= 0.9250, Batch Loss= 0.21
Total Validation Accuracy= 0.6644
Epoch 11, CIFAR-10 Batch 3:  Training Accuracy= 0.9750, Batch Loss= 0.16
Total Validation Accuracy= 0.6602
Epoch 11, CIFAR-10 Batch 4:  Training Accuracy= 0.9500, Batch Loss= 0.23
Total Validation Accuracy= 0.6756
Epoch 11, CIFAR-10 Batch 5:  Training Accuracy= 0.9750, Batch Loss= 0.15
Total Validation Accuracy= 0.6824
Epoch 12, CIFAR-10 Batch 1:  Training Accuracy= 1.0000, Batch Loss= 0.19
Total Validation Accuracy= 0.6572
Epoch 12, CIFAR-10 Batch 2:  Training Accuracy= 0.9250, Batch Loss= 0.17
Total Validation Accuracy= 0.6764
Epoch 12, CIFAR-10 Batch 3:  Training Accuracy= 0.9750, Batch Loss= 0.13
Total Validation Accuracy= 0.6644
Epoch 12, CIFAR-10 Batch 4:  Training Accuracy= 1.0000, Batch Loss= 0.17
Total Validation Accuracy= 0.6748
Epoch 12, CIFAR-10 Batch 5:  Training Accuracy= 1.0000, Batch Loss= 0.11
Total Validation Accuracy= 0.6840
Epoch 13, CIFAR-10 Batch 1:  Training Accuracy= 1.0000, Batch Loss= 0.19
Total Validation Accuracy= 0.6564
Epoch 13, CIFAR-10 Batch 2:  Training Accuracy= 0.9500, Batch Loss= 0.19
Total Validation Accuracy= 0.6748
Epoch 13, CIFAR-10 Batch 3:  Training Accuracy= 1.0000, Batch Loss= 0.08
Total Validation Accuracy= 0.6764
Epoch 13, CIFAR-10 Batch 4:  Training Accuracy= 0.9750, Batch Loss= 0.17
Total Validation Accuracy= 0.6768
Epoch 13, CIFAR-10 Batch 5:  Training Accuracy= 1.0000, Batch Loss= 0.07
Total Validation Accuracy= 0.6940
Epoch 14, CIFAR-10 Batch 1:  Training Accuracy= 1.0000, Batch Loss= 0.14
Total Validation Accuracy= 0.6782
Epoch 14, CIFAR-10 Batch 2:  Training Accuracy= 1.0000, Batch Loss= 0.11
Total Validation Accuracy= 0.6860
Epoch 14, CIFAR-10 Batch 3:  Training Accuracy= 1.0000, Batch Loss= 0.07
Total Validation Accuracy= 0.6572
Epoch 14, CIFAR-10 Batch 4:  Training Accuracy= 0.9500, Batch Loss= 0.14
Total Validation Accuracy= 0.6730
Epoch 14, CIFAR-10 Batch 5:  Training Accuracy= 1.0000, Batch Loss= 0.05
Total Validation Accuracy= 0.6968
Epoch 15, CIFAR-10 Batch 1:  Training Accuracy= 0.9750, Batch Loss= 0.12
Total Validation Accuracy= 0.6822
Epoch 15, CIFAR-10 Batch 2:  Training Accuracy= 0.9750, Batch Loss= 0.10
Total Validation Accuracy= 0.6870
Epoch 15, CIFAR-10 Batch 3:  Training Accuracy= 1.0000, Batch Loss= 0.06
Total Validation Accuracy= 0.6834
Epoch 15, CIFAR-10 Batch 4:  Training Accuracy= 0.9500, Batch Loss= 0.13
Total Validation Accuracy= 0.6682
Epoch 15, CIFAR-10 Batch 5:  Training Accuracy= 1.0000, Batch Loss= 0.04
Total Validation Accuracy= 0.6924
Epoch 16, CIFAR-10 Batch 1:  Training Accuracy= 0.9750, Batch Loss= 0.12
Total Validation Accuracy= 0.6854
Epoch 16, CIFAR-10 Batch 2:  Training Accuracy= 1.0000, Batch Loss= 0.05
Total Validation Accuracy= 0.6968
Epoch 16, CIFAR-10 Batch 3:  Training Accuracy= 1.0000, Batch Loss= 0.04
Total Validation Accuracy= 0.6728
Epoch 16, CIFAR-10 Batch 4:  Training Accuracy= 1.0000, Batch Loss= 0.06
Total Validation Accuracy= 0.6888
Epoch 16, CIFAR-10 Batch 5:  Training Accuracy= 1.0000, Batch Loss= 0.03
Total Validation Accuracy= 0.6846
Epoch 17, CIFAR-10 Batch 1:  Training Accuracy= 1.0000, Batch Loss= 0.08
Total Validation Accuracy= 0.6764
Epoch 17, CIFAR-10 Batch 2:  Training Accuracy= 1.0000, Batch Loss= 0.05
Total Validation Accuracy= 0.6954
Epoch 17, CIFAR-10 Batch 3:  Training Accuracy= 1.0000, Batch Loss= 0.04
Total Validation Accuracy= 0.6802
Epoch 17, CIFAR-10 Batch 4:  Training Accuracy= 1.0000, Batch Loss= 0.07
Total Validation Accuracy= 0.6956
Epoch 17, CIFAR-10 Batch 5:  Training Accuracy= 1.0000, Batch Loss= 0.03
Total Validation Accuracy= 0.6898
Epoch 18, CIFAR-10 Batch 1:  Training Accuracy= 1.0000, Batch Loss= 0.07
Total Validation Accuracy= 0.6878
Epoch 18, CIFAR-10 Batch 2:  Training Accuracy= 1.0000, Batch Loss= 0.04
Total Validation Accuracy= 0.6890
Epoch 18, CIFAR-10 Batch 3:  Training Accuracy= 1.0000, Batch Loss= 0.04
Total Validation Accuracy= 0.6872
Epoch 18, CIFAR-10 Batch 4:  Training Accuracy= 0.9750, Batch Loss= 0.07
Total Validation Accuracy= 0.6902
Epoch 18, CIFAR-10 Batch 5:  Training Accuracy= 0.9750, Batch Loss= 0.06
Total Validation Accuracy= 0.6886
Epoch 19, CIFAR-10 Batch 1:  Training Accuracy= 1.0000, Batch Loss= 0.06
Total Validation Accuracy= 0.6962
Epoch 19, CIFAR-10 Batch 2:  Training Accuracy= 1.0000, Batch Loss= 0.02
Total Validation Accuracy= 0.6892
Epoch 19, CIFAR-10 Batch 3:  Training Accuracy= 1.0000, Batch Loss= 0.03
Total Validation Accuracy= 0.6850
Epoch 19, CIFAR-10 Batch 4:  Training Accuracy= 1.0000, Batch Loss= 0.03
Total Validation Accuracy= 0.6890
Epoch 19, CIFAR-10 Batch 5:  Training Accuracy= 1.0000, Batch Loss= 0.02
Total Validation Accuracy= 0.6876
Epoch 20, CIFAR-10 Batch 1:  Training Accuracy= 1.0000, Batch Loss= 0.03
Total Validation Accuracy= 0.6992
Epoch 20, CIFAR-10 Batch 2:  Training Accuracy= 1.0000, Batch Loss= 0.01
Total Validation Accuracy= 0.6838
Epoch 20, CIFAR-10 Batch 3:  Training Accuracy= 1.0000, Batch Loss= 0.01
Total Validation Accuracy= 0.6930
Epoch 20, CIFAR-10 Batch 4:  Training Accuracy= 1.0000, Batch Loss= 0.02
Total Validation Accuracy= 0.6928
Epoch 20, CIFAR-10 Batch 5:  Training Accuracy= 1.0000, Batch Loss= 0.01
Total Validation Accuracy= 0.6918
Epoch 21, CIFAR-10 Batch 1:  Training Accuracy= 1.0000, Batch Loss= 0.03
Total Validation Accuracy= 0.7052
Epoch 21, CIFAR-10 Batch 2:  Training Accuracy= 1.0000, Batch Loss= 0.01
Total Validation Accuracy= 0.6750
Epoch 21, CIFAR-10 Batch 3:  Training Accuracy= 1.0000, Batch Loss= 0.01
Total Validation Accuracy= 0.6924
Epoch 21, CIFAR-10 Batch 4:  Training Accuracy= 1.0000, Batch Loss= 0.02
Total Validation Accuracy= 0.6990
Epoch 21, CIFAR-10 Batch 5:  Training Accuracy= 1.0000, Batch Loss= 0.01
Total Validation Accuracy= 0.6738
Epoch 22, CIFAR-10 Batch 1:  Training Accuracy= 1.0000, Batch Loss= 0.01
Total Validation Accuracy= 0.6942
Epoch 22, CIFAR-10 Batch 2:  Training Accuracy= 1.0000, Batch Loss= 0.01
Total Validation Accuracy= 0.6858
Epoch 22, CIFAR-10 Batch 3:  Training Accuracy= 1.0000, Batch Loss= 0.01
Total Validation Accuracy= 0.6756
Epoch 22, CIFAR-10 Batch 4:  Training Accuracy= 1.0000, Batch Loss= 0.01
Total Validation Accuracy= 0.6982
Epoch 22, CIFAR-10 Batch 5:  Training Accuracy= 1.0000, Batch Loss= 0.01
Total Validation Accuracy= 0.6906
Epoch 23, CIFAR-10 Batch 1:  Training Accuracy= 1.0000, Batch Loss= 0.02
Total Validation Accuracy= 0.6886
Epoch 23, CIFAR-10 Batch 2:  Training Accuracy= 1.0000, Batch Loss= 0.01
Total Validation Accuracy= 0.6956
Epoch 23, CIFAR-10 Batch 3:  Training Accuracy= 1.0000, Batch Loss= 0.01
Total Validation Accuracy= 0.6974
Epoch 23, CIFAR-10 Batch 4:  Training Accuracy= 1.0000, Batch Loss= 0.01
Total Validation Accuracy= 0.6898
Epoch 23, CIFAR-10 Batch 5:  Training Accuracy= 1.0000, Batch Loss= 0.00
Total Validation Accuracy= 0.6974
Epoch 24, CIFAR-10 Batch 1:  Training Accuracy= 1.0000, Batch Loss= 0.02
Total Validation Accuracy= 0.6930
Epoch 24, CIFAR-10 Batch 2:  Training Accuracy= 1.0000, Batch Loss= 0.01
Total Validation Accuracy= 0.6908
Epoch 24, CIFAR-10 Batch 3:  Training Accuracy= 1.0000, Batch Loss= 0.00
Total Validation Accuracy= 0.6868
Epoch 24, CIFAR-10 Batch 4:  Training Accuracy= 1.0000, Batch Loss= 0.00
Total Validation Accuracy= 0.6812
Epoch 24, CIFAR-10 Batch 5:  Training Accuracy= 1.0000, Batch Loss= 0.00
Total Validation Accuracy= 0.6950
Epoch 25, CIFAR-10 Batch 1:  Training Accuracy= 1.0000, Batch Loss= 0.02
Total Validation Accuracy= 0.6848
Epoch 25, CIFAR-10 Batch 2:  Training Accuracy= 1.0000, Batch Loss= 0.00
Total Validation Accuracy= 0.6962
Epoch 25, CIFAR-10 Batch 3:  Training Accuracy= 1.0000, Batch Loss= 0.01
Total Validation Accuracy= 0.6786
Epoch 25, CIFAR-10 Batch 4:  Training Accuracy= 1.0000, Batch Loss= 0.01
Total Validation Accuracy= 0.6650
Epoch 25, CIFAR-10 Batch 5:  Training Accuracy= 1.0000, Batch Loss= 0.01
Total Validation Accuracy= 0.6914
Epoch 26, CIFAR-10 Batch 1:  Training Accuracy= 1.0000, Batch Loss= 0.01
Total Validation Accuracy= 0.6942
Epoch 26, CIFAR-10 Batch 2:  Training Accuracy= 1.0000, Batch Loss= 0.01
Total Validation Accuracy= 0.6936
Epoch 26, CIFAR-10 Batch 3:  Training Accuracy= 1.0000, Batch Loss= 0.00
Total Validation Accuracy= 0.6870
Epoch 26, CIFAR-10 Batch 4:  Training Accuracy= 1.0000, Batch Loss= 0.00
Total Validation Accuracy= 0.6862
Epoch 26, CIFAR-10 Batch 5:  Training Accuracy= 1.0000, Batch Loss= 0.00
Total Validation Accuracy= 0.7022
Epoch 27, CIFAR-10 Batch 1:  Training Accuracy= 1.0000, Batch Loss= 0.01
Total Validation Accuracy= 0.6972
Epoch 27, CIFAR-10 Batch 2:  Training Accuracy= 1.0000, Batch Loss= 0.00
Total Validation Accuracy= 0.7062
Epoch 27, CIFAR-10 Batch 3:  Training Accuracy= 1.0000, Batch Loss= 0.00
Total Validation Accuracy= 0.6934
Epoch 27, CIFAR-10 Batch 4:  Training Accuracy= 1.0000, Batch Loss= 0.01
Total Validation Accuracy= 0.6798
Epoch 27, CIFAR-10 Batch 5:  Training Accuracy= 1.0000, Batch Loss= 0.00
Total Validation Accuracy= 0.6968
Epoch 28, CIFAR-10 Batch 1:  Training Accuracy= 1.0000, Batch Loss= 0.00
Total Validation Accuracy= 0.6862
Epoch 28, CIFAR-10 Batch 2:  Training Accuracy= 1.0000, Batch Loss= 0.00
Total Validation Accuracy= 0.6994
Epoch 28, CIFAR-10 Batch 3:  Training Accuracy= 1.0000, Batch Loss= 0.01
Total Validation Accuracy= 0.6994
Epoch 28, CIFAR-10 Batch 4:  Training Accuracy= 1.0000, Batch Loss= 0.00
Total Validation Accuracy= 0.6812
Epoch 28, CIFAR-10 Batch 5:  Training Accuracy= 1.0000, Batch Loss= 0.00
Total Validation Accuracy= 0.6972
Epoch 29, CIFAR-10 Batch 1:  Training Accuracy= 1.0000, Batch Loss= 0.00
Total Validation Accuracy= 0.6804
Epoch 29, CIFAR-10 Batch 2:  Training Accuracy= 1.0000, Batch Loss= 0.01
Total Validation Accuracy= 0.6868
Epoch 29, CIFAR-10 Batch 3:  Training Accuracy= 1.0000, Batch Loss= 0.00
Total Validation Accuracy= 0.6986
Epoch 29, CIFAR-10 Batch 4:  Training Accuracy= 1.0000, Batch Loss= 0.00
Total Validation Accuracy= 0.6814
Epoch 29, CIFAR-10 Batch 5:  Training Accuracy= 1.0000, Batch Loss= 0.00
Total Validation Accuracy= 0.7004
Epoch 30, CIFAR-10 Batch 1:  Training Accuracy= 1.0000, Batch Loss= 0.00
Total Validation Accuracy= 0.6922
Epoch 30, CIFAR-10 Batch 2:  Training Accuracy= 1.0000, Batch Loss= 0.03
Total Validation Accuracy= 0.6940
Epoch 30, CIFAR-10 Batch 3:  Training Accuracy= 1.0000, Batch Loss= 0.00
Total Validation Accuracy= 0.7082
Epoch 30, CIFAR-10 Batch 4:  Training Accuracy= 1.0000, Batch Loss= 0.00
Total Validation Accuracy= 0.6760
Epoch 30, CIFAR-10 Batch 5:  Training Accuracy= 1.0000, Batch Loss= 0.00
Total Validation Accuracy= 0.6916

Checkpoint

The model has been saved to disk.

Test Model

Test your model against the test dataset. This will be your final accuracy. You should have an accuracy greater than 50%. If you don't, keep tweaking the model architecture and parameters.


In [15]:
"""
DON'T MODIFY ANYTHING IN THIS CELL
"""
%matplotlib inline
%config InlineBackend.figure_format = 'retina'

import tensorflow as tf
import pickle
import helper
import random

# Set batch size if not already set
try:
    if batch_size:
        pass
except NameError:
    batch_size = 64

save_model_path = './image_classification'
n_samples = 4
top_n_predictions = 3

def test_model():
    """
    Test the saved model against the test dataset
    """

    test_features, test_labels = pickle.load(open('preprocess_test.p', mode='rb'))
    loaded_graph = tf.Graph()

    with tf.Session(graph=loaded_graph) as sess:
        # Load model
        loader = tf.train.import_meta_graph(save_model_path + '.meta')
        loader.restore(sess, save_model_path)

        # Get Tensors from loaded model
        loaded_x = loaded_graph.get_tensor_by_name('x:0')
        loaded_y = loaded_graph.get_tensor_by_name('y:0')
        loaded_keep_prob = loaded_graph.get_tensor_by_name('keep_prob:0')
        loaded_logits = loaded_graph.get_tensor_by_name('logits:0')
        loaded_acc = loaded_graph.get_tensor_by_name('accuracy:0')
        
        # Get accuracy in batches for memory limitations
        test_batch_acc_total = 0
        test_batch_count = 0
        
        for test_feature_batch, test_label_batch in helper.batch_features_labels(test_features, test_labels, batch_size):
            test_batch_acc_total += sess.run(
                loaded_acc,
                feed_dict={loaded_x: test_feature_batch, loaded_y: test_label_batch, loaded_keep_prob: 1.0})
            test_batch_count += 1

        print('Testing Accuracy: {}\n'.format(test_batch_acc_total/test_batch_count))

        # Print Random Samples
        random_test_features, random_test_labels = tuple(zip(*random.sample(list(zip(test_features, test_labels)), n_samples)))
        random_test_predictions = sess.run(
            tf.nn.top_k(tf.nn.softmax(loaded_logits), top_n_predictions),
            feed_dict={loaded_x: random_test_features, loaded_y: random_test_labels, loaded_keep_prob: 1.0})
        helper.display_image_predictions(random_test_features, random_test_labels, random_test_predictions)


test_model()


INFO:tensorflow:Restoring parameters from ./image_classification
Testing Accuracy: 0.6889928343949044

Why 50-80% Accuracy?

You might be wondering why you can't get an accuracy any higher. First things first, 50% isn't bad for a simple CNN. Pure guessing would get you 10% accuracy. However, you might notice people are getting scores well above 80%. That's because we haven't taught you all there is to know about neural networks. We still need to cover a few more techniques.

Submitting This Project

When submitting this project, make sure to run all the cells before saving the notebook. Save the notebook file as "dlnd_image_classification.ipynb" and save it as a HTML file under "File" -> "Download as". Include the "helper.py" and "problem_unittests.py" files in your submission.


In [ ]: