notebook.community

Edit and run

Language Translation

In this project, you’re going to take a peek into the realm of neural network machine translation. You’ll be training a sequence to sequence model on a dataset of English and French sentences that can translate new sentences from English to French.

Get the Data

Since translating the whole language of English to French will take lots of time to train, we have provided you with a small portion of the English corpus.





In [1]:



    


"""
DON'T MODIFY ANYTHING IN THIS CELL
"""
import helper
import problem_unittests as tests

source_path = 'data/small_vocab_en'
target_path = 'data/small_vocab_fr'
source_text = helper.load_data(source_path)
target_text = helper.load_data(target_path)

Explore the Data

Play around with view_sentence_range to view different parts of the data.





In [2]:



    


view_sentence_range = (0, 10)

"""
DON'T MODIFY ANYTHING IN THIS CELL
"""
import numpy as np

print('Dataset Stats')
print('Roughly the number of unique words: {}'.format(len({word: None for word in source_text.split()})))

sentences = source_text.split('\n')
word_counts = [len(sentence.split()) for sentence in sentences]
print('Number of sentences: {}'.format(len(sentences)))
print('Average number of words in a sentence: {}'.format(np.average(word_counts)))

print()
print('English sentences {} to {}:'.format(*view_sentence_range))
print('\n'.join(source_text.split('\n')[view_sentence_range[0]:view_sentence_range[1]]))
print()
print('French sentences {} to {}:'.format(*view_sentence_range))
print('\n'.join(target_text.split('\n')[view_sentence_range[0]:view_sentence_range[1]]))



    


















    






Dataset Stats
Roughly the number of unique words: 227
Number of sentences: 137861
Average number of words in a sentence: 13.225277634719028

English sentences 0 to 10:
new jersey is sometimes quiet during autumn , and it is snowy in april .
the united states is usually chilly during july , and it is usually freezing in november .
california is usually quiet during march , and it is usually hot in june .
the united states is sometimes mild during june , and it is cold in september .
your least liked fruit is the grape , but my least liked is the apple .
his favorite fruit is the orange , but my favorite is the grape .
paris is relaxing during december , but it is usually chilly in july .
new jersey is busy during spring , and it is never hot in march .
our least liked fruit is the lemon , but my least liked is the grape .
the united states is sometimes busy during january , and it is sometimes warm in november .

French sentences 0 to 10:
new jersey est parfois calme pendant l' automne , et il est neigeux en avril .
les états-unis est généralement froid en juillet , et il gèle habituellement en novembre .
california est généralement calme en mars , et il est généralement chaud en juin .
les états-unis est parfois légère en juin , et il fait froid en septembre .
votre moins aimé fruit est le raisin , mais mon moins aimé est la pomme .
son fruit préféré est l'orange , mais mon préféré est le raisin .
paris est relaxant en décembre , mais il est généralement froid en juillet .
new jersey est occupé au printemps , et il est jamais chaude en mars .
notre fruit est moins aimé le citron , mais mon moins aimé est le raisin .
les états-unis est parfois occupé en janvier , et il est parfois chaud en novembre .

















In [47]:



    


unique_eng_words = set([word for word in source_text.split()])
print(unique_eng_words)



    


















    






{'cold', 'to', 'peach.', 'orange.', 'rainy', 'banana.', 'liked', 'strawberry', 'why', 'did', 'have', 'bears', 'store', 'oranges', 'may', 'cats', 'that', 'lemons', 'blue', '?', 'difficult', 'next', 'mangoes.', 'most', 'visit', 'lemon.', 'new', 'were', 'pears', 'wants', 'fall', 'china', 'car', 'lake', 'feared', 'paris', 'in', 'lion', 'dog', 'school', 'rabbit', "aren't", 'hot', 'eiffel', 'might', 'when', 'it', 'lions', 'tower', 'shiny', 'rabbits', 'rusty', '.', 'apple.', 'translate', "didn't", 'favorite', 'pear.', 'apples.', 'french', 'july', 'lime', 'limes', 'wanted', 'bananas.', 'march', 'translating', 'least', 'bird', 'mice', 'pleasant', 'bear', 'snowy', 'pears.', 'grape', 'during', 'freezing', 'has', 'lime.', 'elephant', 'snake', 'never', 'shark', 'she', 'favorite.', 'drove', 'peaches.', 'are', 'truck', 'does', 'november', 'april', 'spring', 'animal', 'june', 'states', 'beautiful', 'and', 'my', 'he', 'busy', 'plans', 'usually', 'dry', 'we', 'snakes', 'dogs', 'where', 'grapes', 'loved.', 'dislikes', 'likes', 'august', 'went', 'would', 'relaxing', 'winter', 'like', 'september', 'red', 'thinks', 'peaches', 'dislike', 'jersey', 'january', 'monkey', 'little', 'liked.', ',', 'going', 'their', 'a', 'yellow', 'green', 'field', 'cat', 'white', 'limes.', 'strawberries.', 'the', 'is', 'february', 'summer', "it's", 'wet', 'his', 'chilly', 'france', 'think', 'between', 'california', 'horses', 'monkeys', 'grapefruit', 'weather', 'grape.', 'drives', 'horse', 'animals', 'apple', 'orange', 'banana', 'wonderful', 'fun', 'birds', 'warm', 'lemons.', 'grapes.', 'her', 'autumn', 'elephants', 'strawberries', 'mango.', 'want', "isn't", 'go', 'how', 'sharks', 'fruit.', 'i', 'they', 'our', 'lemon', 'india', 'peach', 'chinese', 'oranges.', 'this', 'automobile', 'sometimes', 'united', 'spanish', 'grapefruit.', 'driving', 'been', 'strawberry.', 'portuguese', 'easy', 'big', 'december', 'disliked', 'loved', 'apples', 'am', 'mild', 'english', 'was', 'old', 'last', 'bananas', 'fruit', 'mango', 'you', 'do', 'mouse', 'saw', 'pear', 'plan', 'football', 'grocery', 'october', 'black', 'quiet', 'nice', 'mangoes', 'but', 'your'}

Implement Preprocessing Function

Text to Word Ids

As you did with other RNNs, you must turn the text into a number so the computer can understand it. In the function text_to_ids(), you'll turn source_text and target_text from words to ids. However, you need to add the <EOS> word id at the end of target_text. This will help the neural network predict when the sentence should end.

You can get the <EOS> word id by doing:

target_vocab_to_int['<EOS>']

You can get other word ids using source_vocab_to_int and target_vocab_to_int.





In [5]:



    


def text_to_ids(source_text, target_text, source_vocab_to_int, target_vocab_to_int):
    """
    Convert source and target text to proper word ids
    :param source_text: String that contains all the source text.
    :param target_text: String that contains all the target text.
    :param source_vocab_to_int: Dictionary to go from the source words to an id
    :param target_vocab_to_int: Dictionary to go from the target words to an id
    :return: A tuple of lists (source_id_text, target_id_text)
    """    
    
    source_id_text = [[source_vocab_to_int[word] for word in sentence.split()] 
                      for sentence in source_text.split('\n')]
    
    target_id_text = [[target_vocab_to_int[word] for word in sentence.split()] 
                      + [target_vocab_to_int['<EOS>']]
                      for sentence in target_text.split('\n')]
    
    return source_id_text, target_id_text

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



    


















    






Tests Passed

Preprocess all the data and save it

Running the code cell below will preprocess all the data and save it to file.





In [8]:



    


"""
DON'T MODIFY ANYTHING IN THIS CELL
"""
helper.preprocess_and_save_data(source_path, target_path, text_to_ids)

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 [9]:



    


"""
DON'T MODIFY ANYTHING IN THIS CELL
"""
import numpy as np
import helper
import problem_unittests as tests

(source_int_text, target_int_text), (source_vocab_to_int, target_vocab_to_int), _ = helper.load_preprocess()

Check the Version of TensorFlow and Access to GPU

This will check to make sure you have the correct version of TensorFlow and access to a GPU





In [10]:



    


"""
DON'T MODIFY ANYTHING IN THIS CELL
"""
from distutils.version import LooseVersion
import warnings
import tensorflow as tf
from tensorflow.python.layers.core import Dense

# Check TensorFlow Version
assert LooseVersion(tf.__version__) >= LooseVersion('1.1'), 'Please use TensorFlow version 1.1 or newer'
print('TensorFlow Version: {}'.format(tf.__version__))

# Check for a GPU
if not tf.test.gpu_device_name():
    warnings.warn('No GPU found. Please use a GPU to train your neural network.')
else:
    print('Default GPU Device: {}'.format(tf.test.gpu_device_name()))



    


















    






TensorFlow Version: 1.0.1
Default GPU Device: /gpu:0

Build the Neural Network

You'll build the components necessary to build a Sequence-to-Sequence model by implementing the following functions below:

  • model_inputs
  • process_decoder_input
  • encoding_layer
  • decoding_layer_train
  • decoding_layer_infer
  • decoding_layer
  • seq2seq_model

Input

Implement the model_inputs() function to create TF Placeholders for the Neural Network. It should create the following placeholders:

  • Input text placeholder named "input" using the TF Placeholder name parameter with rank 2.
  • Targets placeholder with rank 2.
  • Learning rate placeholder with rank 0.
  • Keep probability placeholder named "keep_prob" using the TF Placeholder name parameter with rank 0.
  • Target sequence length placeholder named "target_sequence_length" with rank 1
  • Max target sequence length tensor named "max_target_len" getting its value from applying tf.reduce_max on the target_sequence_length placeholder. Rank 0.
  • Source sequence length placeholder named "source_sequence_length" with rank 1

Return the placeholders in the following the tuple (input, targets, learning rate, keep probability, target sequence length, max target sequence length, source sequence length)





In [11]:



    


def model_inputs():
    """
    Create TF Placeholders for input, targets, learning rate, and lengths of source and target sequences.
    :return: Tuple (input, targets, learning rate, keep probability, target sequence length,
    max target sequence length, source sequence length)
    """
    input_ = tf.placeholder(tf.int32, [None, None], name="input")
    targets = tf.placeholder(tf.int32, [None, None], name="targets")
    lr = tf.placeholder(tf.float32, name="learning_rate")
    keep_prob = tf.placeholder(tf.float32, name="keep_prob")
    
    return input_, targets, lr, keep_prob

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



    


















    






Tests Passed

Process Decoder Input

Implement process_decoder_input by removing the last word id from each batch in target_data and concat the GO ID to the begining of each batch.





In [12]:



    


def process_decoder_input(target_data, target_vocab_to_int, batch_size):
    """
    Preprocess target data for encoding
    :param target_data: Target Placehoder
    :param target_vocab_to_int: Dictionary to go from the target words to an id
    :param batch_size: Batch Size
    :return: Preprocessed target data
    """
    ending = tf.strided_slice(target_data, [0, 0], [batch_size, -1], [1, 1])
    dec_input = tf.concat([tf.fill([batch_size, 1], target_vocab_to_int['<GO>']), ending], 1)
    
    return dec_input

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



    


















    






Tests Passed

Encoding

Implement encoding_layer() to create a Encoder RNN layer:





In [13]:



    


from imp import reload
reload(tests)

def encoding_layer(rnn_inputs, rnn_size, num_layers, keep_prob, 
                   source_sequence_length, source_vocab_size, 
                   encoding_embedding_size):
    """
    Create encoding layer
    :param rnn_inputs: Inputs for the RNN
    :param rnn_size: RNN Size
    :param num_layers: Number of layers
    :param keep_prob: Dropout keep probability
    :param source_sequence_length: a list of the lengths of each sequence in the batch
    :param source_vocab_size: vocabulary size of source data
    :param encoding_embedding_size: embedding size of source data
    :return: tuple (RNN output, RNN state)
    """
    # Encoder
    cell = tf.contrib.rnn.BasicLSTMCell(rnn_size)
    drop = tf.contrib.rnn.DropoutWrapper(cell, output_keep_prob=keep_prob)
    enc_cell = tf.contrib.rnn.MultiRNNCell([drop] * num_layers)
    _, enc_state = tf.nn.dynamic_rnn(enc_cell, rnn_inputs, dtype=tf.float32)
    
    return enc_state

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



    


















    






Tests Passed

Decoding - Training

Create a training decoding layer:





In [15]:



    


def decoding_layer_train(encoder_state, dec_cell, dec_embed_input, 
                         target_sequence_length, max_summary_length, 
                         output_layer, keep_prob):
    """
    Create a decoding layer for training
    :param encoder_state: Encoder State
    :param dec_cell: Decoder RNN Cell
    :param dec_embed_input: Decoder embedded input
    :param target_sequence_length: The lengths of each sequence in the target batch
    :param max_summary_length: The length of the longest sequence in the batch
    :param output_layer: Function to apply the output layer
    :param keep_prob: Dropout keep probability
    :return: BasicDecoderOutput containing training logits and sample_id
    """
    
    # Training Decoder
    train_decoder_fn = tf.contrib.seq2seq.simple_decoder_fn_train(encoder_state)
    train_pred, _, _ = tf.contrib.seq2seq.dynamic_rnn_decoder(
        dec_cell, train_decoder_fn, dec_embed_input, sequence_length, scope=decoding_scope)

    # Apply output function
    train_logits =  output_fn(train_pred)
    
    return train_logits



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



    


















    






Tests Passed

Decoding - Inference

Create inference decoder:





In [17]:



    


def decoding_layer_infer(encoder_state, dec_cell, dec_embeddings, start_of_sequence_id,
                         end_of_sequence_id, max_target_sequence_length,
                         vocab_size, output_layer, batch_size, keep_prob):
    """
    Create a decoding layer for inference
    :param encoder_state: Encoder state
    :param dec_cell: Decoder RNN Cell
    :param dec_embeddings: Decoder embeddings
    :param start_of_sequence_id: GO ID
    :param end_of_sequence_id: EOS Id
    :param max_target_sequence_length: Maximum length of target sequences
    :param vocab_size: Size of decoder/target vocabulary
    :param decoding_scope: TenorFlow Variable Scope for decoding
    :param output_layer: Function to apply the output layer
    :param batch_size: Batch size
    :param keep_prob: Dropout keep probability
    :return: BasicDecoderOutput containing inference logits and sample_id
    """
    infer_decoder_fn = tf.contrib.seq2seq.simple_decoder_fn_inference(
        output_fn, encoder_state, dec_embeddings,start_of_sequence_id, end_of_sequence_id, 
        maximum_length, vocab_size)
    inference_logits, _, _ = tf.contrib.seq2seq.dynamic_rnn_decoder(dec_cell, infer_decoder_fn, scope=decoding_scope)
    
    return inference_logits



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



    


















    






Tests Passed

Build the Decoding Layer

Implement decoding_layer() to create a Decoder RNN layer.

  • Embed the target sequences
  • Construct the decoder LSTM cell (just like you constructed the encoder cell above)
  • Create an output layer to map the outputs of the decoder to the elements of our vocabulary
  • Use the your decoding_layer_train(encoder_state, dec_cell, dec_embed_input, target_sequence_length, max_target_sequence_length, output_layer, keep_prob) function to get the training logits.
  • Use your decoding_layer_infer(encoder_state, dec_cell, dec_embeddings, start_of_sequence_id, end_of_sequence_id, max_target_sequence_length, vocab_size, output_layer, batch_size, keep_prob) function to get the inference logits.

Note: You'll need to use tf.variable_scope to share variables between training and inference.





In [23]:



    


def decoding_layer(dec_input, encoder_state,
                   target_sequence_length, max_target_sequence_length,
                   rnn_size,
                   num_layers, target_vocab_to_int, target_vocab_size,
                   batch_size, keep_prob, decoding_embedding_size):
    """
    Create decoding layer
    :param dec_input: Decoder input
    :param encoder_state: Encoder state
    :param target_sequence_length: The lengths of each sequence in the target batch
    :param max_target_sequence_length: Maximum length of target sequences
    :param rnn_size: RNN Size
    :param num_layers: Number of layers
    :param target_vocab_to_int: Dictionary to go from the target words to an id
    :param target_vocab_size: Size of target vocabulary
    :param batch_size: The size of the batch
    :param keep_prob: Dropout keep probability
    :param decoding_embedding_size: Decoding embedding size
    :return: Tuple of (Training BasicDecoderOutput, Inference BasicDecoderOutput)
    """
    # TODO: Implement Function
    
    # Decoder RNNs
    
    with tf.variable_scope('decoding') as decoding_scope:
        cell = tf.contrib.rnn.BasicLSTMCell(rnn_size)
        drop = tf.contrib.rnn.DropoutWrapper(cell, output_keep_prob=keep_prob)
        dec_cell = tf.contrib.rnn.MultiRNNCell([drop] * num_layers)
        
        output_fn = lambda x: tf.contrib.layers.fully_connected(x, vocab_size, None, scope=decoding_scope)
        
        train_logits = decoding_layer_train(encoder_state, dec_cell, dec_embed_input, sequence_length, decoding_scope,
                         output_fn, keep_prob)
        
        
        start_of_sequence_id = target_vocab_to_int['<GO>']
        end_of_sequence_id = target_vocab_to_int['<EOS>']
        decoding_scope.reuse_variables()
        inference_logits = decoding_layer_infer(encoder_state, dec_cell, dec_embeddings, start_of_sequence_id, end_of_sequence_id,
                         sequence_length, vocab_size, decoding_scope, output_fn, keep_prob)
        
    
    
    return train_logits, inference_logits



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



    


















    






Tests Passed

Build the Neural Network

Apply the functions you implemented above to:

  • Encode the input using your encoding_layer(rnn_inputs, rnn_size, num_layers, keep_prob, source_sequence_length, source_vocab_size, encoding_embedding_size).
  • Process target data using your process_decoder_input(target_data, target_vocab_to_int, batch_size) function.
  • Decode the encoded input using your decoding_layer(dec_input, enc_state, target_sequence_length, max_target_sentence_length, rnn_size, num_layers, target_vocab_to_int, target_vocab_size, batch_size, keep_prob, dec_embedding_size) function.




In [25]:



    


def seq2seq_model(input_data, target_data, keep_prob, batch_size,
                  source_sequence_length, target_sequence_length,
                  max_target_sentence_length,
                  source_vocab_size, target_vocab_size,
                  enc_embedding_size, dec_embedding_size,
                  rnn_size, num_layers, target_vocab_to_int):
    """
    Build the Sequence-to-Sequence part of the neural network
    :param input_data: Input placeholder
    :param target_data: Target placeholder
    :param keep_prob: Dropout keep probability placeholder
    :param batch_size: Batch Size
    :param source_sequence_length: Sequence Lengths of source sequences in the batch
    :param target_sequence_length: Sequence Lengths of target sequences in the batch
    :param source_vocab_size: Source vocabulary size
    :param target_vocab_size: Target vocabulary size
    :param enc_embedding_size: Decoder embedding size
    :param dec_embedding_size: Encoder embedding size
    :param rnn_size: RNN Size
    :param num_layers: Number of layers
    :param target_vocab_to_int: Dictionary to go from the target words to an id
    :return: Tuple of (Training BasicDecoderOutput, Inference BasicDecoderOutput)
    """
    
    # Apply embedding to the input data for the encoder
    enc_embed_input = tf.contrib.layers.embed_sequence(input_data, source_vocab_size, enc_embedding_size)
    
    # Encode the input using your encoding_layer
    encoder_state = encoding_layer(enc_embed_input, rnn_size, num_layers, keep_prob)
    
    # Process target data using your process_encoding_input
    dec_input = process_decoding_input(target_data, target_vocab_to_int, batch_size)
    
    # Embed dec_input
    dec_embeddings = tf.Variable(tf.random_uniform([target_vocab_size, dec_embedding_size]))
    dec_embed_input = tf.nn.embedding_lookup(dec_embeddings, dec_input)

    # Decode the encoded input 
    train_logits, inference_logits = decoding_layer(dec_embed_input, dec_embeddings, encoder_state, 
                                                    target_vocab_size, sequence_length, rnn_size,
                                                    num_layers, target_vocab_to_int, keep_prob)
    
    
    
    return (train_logits, inference_logits)


"""
DON'T MODIFY ANYTHING IN THIS CELL THAT IS BELOW THIS LINE
"""
tests.test_seq2seq_model(seq2seq_model)



    


















    






Tests Passed

Neural Network Training

Hyperparameters

Tune the following parameters:

  • Set epochs to the number of epochs.
  • Set batch_size to the batch size.
  • Set rnn_size to the size of the RNNs.
  • Set num_layers to the number of layers.
  • Set encoding_embedding_size to the size of the embedding for the encoder.
  • Set decoding_embedding_size to the size of the embedding for the decoder.
  • Set learning_rate to the learning rate.
  • Set keep_probability to the Dropout keep probability
  • Set display_step to state how many steps between each debug output statement




In [33]:



    


# Number of Epochs
epochs = 10
# Batch Size
batch_size = 1024
# RNN Size
rnn_size = 512
# Number of Layers
num_layers = 2
# Embedding Size
encoding_embedding_size = 300
decoding_embedding_size = 300
# Learning Rate
learning_rate = 0.01
# Dropout Keep Probability
keep_probability = 0.5

Build the Graph

Build the graph using the neural network you implemented.





In [34]:



    


"""
DON'T MODIFY ANYTHING IN THIS CELL
"""
save_path = 'checkpoints/dev'
(source_int_text, target_int_text), (source_vocab_to_int, target_vocab_to_int), _ = helper.load_preprocess()
max_target_sentence_length = max([len(sentence) for sentence in source_int_text])

train_graph = tf.Graph()
with train_graph.as_default():
    input_data, targets, lr, keep_prob, target_sequence_length, max_target_sequence_length, source_sequence_length = model_inputs()

    #sequence_length = tf.placeholder_with_default(max_target_sentence_length, None, name='sequence_length')
    input_shape = tf.shape(input_data)

    train_logits, inference_logits = seq2seq_model(tf.reverse(input_data, [-1]),
                                                   targets,
                                                   keep_prob,
                                                   batch_size,
                                                   source_sequence_length,
                                                   target_sequence_length,
                                                   max_target_sequence_length,
                                                   len(source_vocab_to_int),
                                                   len(target_vocab_to_int),
                                                   encoding_embedding_size,
                                                   decoding_embedding_size,
                                                   rnn_size,
                                                   num_layers,
                                                   target_vocab_to_int)


    training_logits = tf.identity(train_logits.rnn_output, name='logits')
    inference_logits = tf.identity(inference_logits.sample_id, name='predictions')

    masks = tf.sequence_mask(target_sequence_length, max_target_sequence_length, dtype=tf.float32, name='masks')

    with tf.name_scope("optimization"):
        # Loss function
        cost = tf.contrib.seq2seq.sequence_loss(
            training_logits,
            targets,
            masks)

        # Optimizer
        optimizer = tf.train.AdamOptimizer(lr)

        # Gradient Clipping
        gradients = optimizer.compute_gradients(cost)
        capped_gradients = [(tf.clip_by_value(grad, -1., 1.), var) for grad, var in gradients if grad is not None]
        train_op = optimizer.apply_gradients(capped_gradients)

Batch and pad the source and target sequences





In [ ]:



    


"""
DON'T MODIFY ANYTHING IN THIS CELL
"""
def pad_sentence_batch(sentence_batch, pad_int):
    """Pad sentences with <PAD> so that each sentence of a batch has the same length"""
    max_sentence = max([len(sentence) for sentence in sentence_batch])
    return [sentence + [pad_int] * (max_sentence - len(sentence)) for sentence in sentence_batch]


def get_batches(sources, targets, batch_size, source_pad_int, target_pad_int):
    """Batch targets, sources, and the lengths of their sentences together"""
    for batch_i in range(0, len(sources)//batch_size):
        start_i = batch_i * batch_size

        # Slice the right amount for the batch
        sources_batch = sources[start_i:start_i + batch_size]
        targets_batch = targets[start_i:start_i + batch_size]

        # Pad
        pad_sources_batch = np.array(pad_sentence_batch(sources_batch, source_pad_int))
        pad_targets_batch = np.array(pad_sentence_batch(targets_batch, target_pad_int))

        # Need the lengths for the _lengths parameters
        pad_targets_lengths = []
        for target in pad_targets_batch:
            pad_targets_lengths.append(len(target))

        pad_source_lengths = []
        for source in pad_sources_batch:
            pad_source_lengths.append(len(source))

        yield pad_sources_batch, pad_targets_batch, pad_source_lengths, pad_targets_lengths

Train

Train the neural network on the preprocessed data. If you have a hard time getting a good loss, check the forms to see if anyone is having the same problem.





In [35]:



    


"""
DON'T MODIFY ANYTHING IN THIS CELL
"""
def get_accuracy(target, logits):
    """
    Calculate accuracy
    """
    max_seq = max(target.shape[1], logits.shape[1])
    if max_seq - target.shape[1]:
        target = np.pad(
            target,
            [(0,0),(0,max_seq - target.shape[1])],
            'constant')
    if max_seq - logits.shape[1]:
        logits = np.pad(
            logits,
            [(0,0),(0,max_seq - logits.shape[1])],
            'constant')

    return np.mean(np.equal(target, logits))

# Split data to training and validation sets
train_source = source_int_text[batch_size:]
train_target = target_int_text[batch_size:]
valid_source = source_int_text[:batch_size]
valid_target = target_int_text[:batch_size]
(valid_sources_batch, valid_targets_batch, valid_sources_lengths, valid_targets_lengths ) = next(get_batches(valid_source,
                                                                                                             valid_target,
                                                                                                             batch_size,
                                                                                                             source_vocab_to_int['<PAD>'],
                                                                                                             target_vocab_to_int['<PAD>']))                                                                                                  
with tf.Session(graph=train_graph) as sess:
    sess.run(tf.global_variables_initializer())

    for epoch_i in range(epochs):
        for batch_i, (source_batch, target_batch, sources_lengths, targets_lengths) in enumerate(
                get_batches(train_source, train_target, batch_size,
                            source_vocab_to_int['<PAD>'],
                            target_vocab_to_int['<PAD>'])):

            _, loss = sess.run(
                [train_op, cost],
                {input_data: source_batch,
                 targets: target_batch,
                 lr: learning_rate,
                 target_sequence_length: targets_lengths,
                 source_sequence_length: sources_lengths,
                 keep_prob: keep_probability})


            if batch_i % display_step == 0 and batch_i > 0:


                batch_train_logits = sess.run(
                    inference_logits,
                    {input_data: source_batch,
                     source_sequence_length: sources_lengths,
                     target_sequence_length: targets_lengths,
                     keep_prob: 1.0})


                batch_valid_logits = sess.run(
                    inference_logits,
                    {input_data: valid_sources_batch,
                     source_sequence_length: valid_sources_lengths,
                     target_sequence_length: valid_targets_lengths,
                     keep_prob: 1.0})

                train_acc = get_accuracy(target_batch, batch_train_logits)

                valid_acc = get_accuracy(valid_targets_batch, batch_valid_logits)

                print('Epoch {:>3} Batch {:>4}/{} - Train Accuracy: {:>6.4f}, Validation Accuracy: {:>6.4f}, Loss: {:>6.4f}'
                      .format(epoch_i, batch_i, len(source_int_text) // batch_size, train_acc, valid_acc, loss))

    # Save Model
    saver = tf.train.Saver()
    saver.save(sess, save_path)
    print('Model Trained and Saved')



    


















    






Epoch   0 Batch    0/134 - Train Accuracy:  0.267, Validation Accuracy:  0.310, Loss:  5.875
Epoch   0 Batch    1/134 - Train Accuracy:  0.211, Validation Accuracy:  0.283, Loss:  7.318
Epoch   0 Batch    2/134 - Train Accuracy:  0.300, Validation Accuracy:  0.334, Loss:  4.296
Epoch   0 Batch    3/134 - Train Accuracy:  0.207, Validation Accuracy:  0.242, Loss:  3.731
Epoch   0 Batch    4/134 - Train Accuracy:  0.272, Validation Accuracy:  0.313, Loss:  4.387
Epoch   0 Batch    5/134 - Train Accuracy:  0.295, Validation Accuracy:  0.328, Loss:  3.951
Epoch   0 Batch    6/134 - Train Accuracy:  0.326, Validation Accuracy:  0.329, Loss:  3.656
Epoch   0 Batch    7/134 - Train Accuracy:  0.282, Validation Accuracy:  0.326, Loss:  3.656
Epoch   0 Batch    8/134 - Train Accuracy:  0.286, Validation Accuracy:  0.328, Loss:  3.594
Epoch   0 Batch    9/134 - Train Accuracy:  0.325, Validation Accuracy:  0.328, Loss:  3.420
Epoch   0 Batch   10/134 - Train Accuracy:  0.316, Validation Accuracy:  0.346, Loss:  3.510
Epoch   0 Batch   11/134 - Train Accuracy:  0.319, Validation Accuracy:  0.346, Loss:  3.438
Epoch   0 Batch   12/134 - Train Accuracy:  0.346, Validation Accuracy:  0.346, Loss:  3.269
Epoch   0 Batch   13/134 - Train Accuracy:  0.306, Validation Accuracy:  0.346, Loss:  3.351
Epoch   0 Batch   14/134 - Train Accuracy:  0.312, Validation Accuracy:  0.349, Loss:  3.272
Epoch   0 Batch   15/134 - Train Accuracy:  0.316, Validation Accuracy:  0.349, Loss:  3.196
Epoch   0 Batch   16/134 - Train Accuracy:  0.319, Validation Accuracy:  0.349, Loss:  3.101
Epoch   0 Batch   17/134 - Train Accuracy:  0.338, Validation Accuracy:  0.367, Loss:  3.089
Epoch   0 Batch   18/134 - Train Accuracy:  0.354, Validation Accuracy:  0.383, Loss:  3.041
Epoch   0 Batch   19/134 - Train Accuracy:  0.352, Validation Accuracy:  0.384, Loss:  3.005
Epoch   0 Batch   20/134 - Train Accuracy:  0.385, Validation Accuracy:  0.388, Loss:  2.853
Epoch   0 Batch   21/134 - Train Accuracy:  0.363, Validation Accuracy:  0.388, Loss:  2.909
Epoch   0 Batch   22/134 - Train Accuracy:  0.321, Validation Accuracy:  0.386, Loss:  3.021
Epoch   0 Batch   23/134 - Train Accuracy:  0.385, Validation Accuracy:  0.385, Loss:  2.723
Epoch   0 Batch   24/134 - Train Accuracy:  0.330, Validation Accuracy:  0.386, Loss:  2.944
Epoch   0 Batch   25/134 - Train Accuracy:  0.354, Validation Accuracy:  0.385, Loss:  2.813
Epoch   0 Batch   26/134 - Train Accuracy:  0.329, Validation Accuracy:  0.388, Loss:  2.921
Epoch   0 Batch   27/134 - Train Accuracy:  0.368, Validation Accuracy:  0.397, Loss:  2.754
Epoch   0 Batch   28/134 - Train Accuracy:  0.383, Validation Accuracy:  0.410, Loss:  2.742
Epoch   0 Batch   29/134 - Train Accuracy:  0.409, Validation Accuracy:  0.416, Loss:  2.606
Epoch   0 Batch   30/134 - Train Accuracy:  0.407, Validation Accuracy:  0.411, Loss:  2.572
Epoch   0 Batch   31/134 - Train Accuracy:  0.379, Validation Accuracy:  0.413, Loss:  2.650
Epoch   0 Batch   32/134 - Train Accuracy:  0.413, Validation Accuracy:  0.420, Loss:  2.523
Epoch   0 Batch   33/134 - Train Accuracy:  0.404, Validation Accuracy:  0.437, Loss:  2.606
Epoch   0 Batch   34/134 - Train Accuracy:  0.426, Validation Accuracy:  0.450, Loss:  2.575
Epoch   0 Batch   35/134 - Train Accuracy:  0.446, Validation Accuracy:  0.450, Loss:  2.437
Epoch   0 Batch   36/134 - Train Accuracy:  0.430, Validation Accuracy:  0.456, Loss:  2.499
Epoch   0 Batch   37/134 - Train Accuracy:  0.430, Validation Accuracy:  0.466, Loss:  2.504
Epoch   0 Batch   38/134 - Train Accuracy:  0.435, Validation Accuracy:  0.466, Loss:  2.456
Epoch   0 Batch   39/134 - Train Accuracy:  0.444, Validation Accuracy:  0.467, Loss:  2.405
Epoch   0 Batch   40/134 - Train Accuracy:  0.442, Validation Accuracy:  0.465, Loss:  2.391
Epoch   0 Batch   41/134 - Train Accuracy:  0.457, Validation Accuracy:  0.478, Loss:  2.356
Epoch   0 Batch   42/134 - Train Accuracy:  0.456, Validation Accuracy:  0.488, Loss:  2.379
Epoch   0 Batch   43/134 - Train Accuracy:  0.466, Validation Accuracy:  0.499, Loss:  2.361
Epoch   0 Batch   44/134 - Train Accuracy:  0.484, Validation Accuracy:  0.508, Loss:  2.315
Epoch   0 Batch   45/134 - Train Accuracy:  0.481, Validation Accuracy:  0.507, Loss:  2.291
Epoch   0 Batch   46/134 - Train Accuracy:  0.510, Validation Accuracy:  0.512, Loss:  2.189
Epoch   0 Batch   47/134 - Train Accuracy:  0.492, Validation Accuracy:  0.518, Loss:  2.242
Epoch   0 Batch   48/134 - Train Accuracy:  0.503, Validation Accuracy:  0.519, Loss:  2.207
Epoch   0 Batch   49/134 - Train Accuracy:  0.505, Validation Accuracy:  0.523, Loss:  2.185
Epoch   0 Batch   50/134 - Train Accuracy:  0.502, Validation Accuracy:  0.528, Loss:  2.187
Epoch   0 Batch   51/134 - Train Accuracy:  0.499, Validation Accuracy:  0.525, Loss:  2.157
Epoch   0 Batch   52/134 - Train Accuracy:  0.475, Validation Accuracy:  0.508, Loss:  2.168
Epoch   0 Batch   53/134 - Train Accuracy:  0.493, Validation Accuracy:  0.526, Loss:  2.164
Epoch   0 Batch   54/134 - Train Accuracy:  0.510, Validation Accuracy:  0.540, Loss:  2.121
Epoch   0 Batch   55/134 - Train Accuracy:  0.493, Validation Accuracy:  0.523, Loss:  2.100
Epoch   0 Batch   56/134 - Train Accuracy:  0.535, Validation Accuracy:  0.541, Loss:  1.987
Epoch   0 Batch   57/134 - Train Accuracy:  0.521, Validation Accuracy:  0.545, Loss:  2.043
Epoch   0 Batch   58/134 - Train Accuracy:  0.531, Validation Accuracy:  0.554, Loss:  2.015
Epoch   0 Batch   59/134 - Train Accuracy:  0.508, Validation Accuracy:  0.552, Loss:  2.110
Epoch   0 Batch   60/134 - Train Accuracy:  0.528, Validation Accuracy:  0.549, Loss:  1.978
Epoch   0 Batch   61/134 - Train Accuracy:  0.505, Validation Accuracy:  0.534, Loss:  1.986
Epoch   0 Batch   62/134 - Train Accuracy:  0.543, Validation Accuracy:  0.561, Loss:  1.938
Epoch   0 Batch   63/134 - Train Accuracy:  0.542, Validation Accuracy:  0.558, Loss:  1.919
Epoch   0 Batch   64/134 - Train Accuracy:  0.517, Validation Accuracy:  0.551, Loss:  1.939
Epoch   0 Batch   65/134 - Train Accuracy:  0.529, Validation Accuracy:  0.553, Loss:  1.900
Epoch   0 Batch   66/134 - Train Accuracy:  0.543, Validation Accuracy:  0.564, Loss:  1.856
Epoch   0 Batch   67/134 - Train Accuracy:  0.488, Validation Accuracy:  0.548, Loss:  1.978
Epoch   0 Batch   68/134 - Train Accuracy:  0.527, Validation Accuracy:  0.566, Loss:  1.921
Epoch   0 Batch   69/134 - Train Accuracy:  0.545, Validation Accuracy:  0.564, Loss:  1.821
Epoch   0 Batch   70/134 - Train Accuracy:  0.538, Validation Accuracy:  0.552, Loss:  1.755
Epoch   0 Batch   71/134 - Train Accuracy:  0.539, Validation Accuracy:  0.556, Loss:  1.764
Epoch   0 Batch   72/134 - Train Accuracy:  0.524, Validation Accuracy:  0.553, Loss:  1.764
Epoch   0 Batch   73/134 - Train Accuracy:  0.569, Validation Accuracy:  0.576, Loss:  1.690
Epoch   0 Batch   74/134 - Train Accuracy:  0.551, Validation Accuracy:  0.575, Loss:  1.717
Epoch   0 Batch   75/134 - Train Accuracy:  0.563, Validation Accuracy:  0.568, Loss:  1.629
Epoch   0 Batch   76/134 - Train Accuracy:  0.559, Validation Accuracy:  0.580, Loss:  1.678
Epoch   0 Batch   77/134 - Train Accuracy:  0.580, Validation Accuracy:  0.575, Loss:  1.575
Epoch   0 Batch   78/134 - Train Accuracy:  0.543, Validation Accuracy:  0.569, Loss:  1.645
Epoch   0 Batch   79/134 - Train Accuracy:  0.535, Validation Accuracy:  0.562, Loss:  1.634
Epoch   0 Batch   80/134 - Train Accuracy:  0.555, Validation Accuracy:  0.581, Loss:  1.610
Epoch   0 Batch   81/134 - Train Accuracy:  0.562, Validation Accuracy:  0.581, Loss:  1.585
Epoch   0 Batch   82/134 - Train Accuracy:  0.577, Validation Accuracy:  0.578, Loss:  1.476
Epoch   0 Batch   83/134 - Train Accuracy:  0.542, Validation Accuracy:  0.566, Loss:  1.555
Epoch   0 Batch   84/134 - Train Accuracy:  0.551, Validation Accuracy:  0.573, Loss:  1.536
Epoch   0 Batch   85/134 - Train Accuracy:  0.508, Validation Accuracy:  0.533, Loss:  1.543
Epoch   0 Batch   86/134 - Train Accuracy:  0.558, Validation Accuracy:  0.582, Loss:  1.572
Epoch   0 Batch   87/134 - Train Accuracy:  0.566, Validation Accuracy:  0.589, Loss:  1.498
Epoch   0 Batch   88/134 - Train Accuracy:  0.557, Validation Accuracy:  0.566, Loss:  1.402
Epoch   0 Batch   89/134 - Train Accuracy:  0.544, Validation Accuracy:  0.565, Loss:  1.443
Epoch   0 Batch   90/134 - Train Accuracy:  0.558, Validation Accuracy:  0.581, Loss:  1.430
Epoch   0 Batch   91/134 - Train Accuracy:  0.634, Validation Accuracy:  0.594, Loss:  1.233
Epoch   0 Batch   92/134 - Train Accuracy:  0.569, Validation Accuracy:  0.582, Loss:  1.376
Epoch   0 Batch   93/134 - Train Accuracy:  0.569, Validation Accuracy:  0.578, Loss:  1.336
Epoch   0 Batch   94/134 - Train Accuracy:  0.571, Validation Accuracy:  0.585, Loss:  1.341
Epoch   0 Batch   95/134 - Train Accuracy:  0.570, Validation Accuracy:  0.586, Loss:  1.329
Epoch   0 Batch   96/134 - Train Accuracy:  0.566, Validation Accuracy:  0.593, Loss:  1.329
Epoch   0 Batch   97/134 - Train Accuracy:  0.562, Validation Accuracy:  0.592, Loss:  1.326
Epoch   0 Batch   98/134 - Train Accuracy:  0.528, Validation Accuracy:  0.568, Loss:  1.363
Epoch   0 Batch   99/134 - Train Accuracy:  0.537, Validation Accuracy:  0.556, Loss:  1.274
Epoch   0 Batch  100/134 - Train Accuracy:  0.520, Validation Accuracy:  0.542, Loss:  1.266
Epoch   0 Batch  101/134 - Train Accuracy:  0.466, Validation Accuracy:  0.521, Loss:  1.300
Epoch   0 Batch  102/134 - Train Accuracy:  0.499, Validation Accuracy:  0.518, Loss:  1.239
Epoch   0 Batch  103/134 - Train Accuracy:  0.511, Validation Accuracy:  0.510, Loss:  1.177
Epoch   0 Batch  104/134 - Train Accuracy:  0.499, Validation Accuracy:  0.512, Loss:  1.190
Epoch   0 Batch  105/134 - Train Accuracy:  0.511, Validation Accuracy:  0.534, Loss:  1.186
Epoch   0 Batch  106/134 - Train Accuracy:  0.523, Validation Accuracy:  0.540, Loss:  1.176
Epoch   0 Batch  107/134 - Train Accuracy:  0.526, Validation Accuracy:  0.522, Loss:  1.112
Epoch   0 Batch  108/134 - Train Accuracy:  0.467, Validation Accuracy:  0.514, Loss:  1.219
Epoch   0 Batch  109/134 - Train Accuracy:  0.510, Validation Accuracy:  0.511, Loss:  1.095
Epoch   0 Batch  110/134 - Train Accuracy:  0.489, Validation Accuracy:  0.505, Loss:  1.120
Epoch   0 Batch  111/134 - Train Accuracy:  0.576, Validation Accuracy:  0.585, Loss:  1.127
Epoch   0 Batch  112/134 - Train Accuracy:  0.588, Validation Accuracy:  0.595, Loss:  1.101
Epoch   0 Batch  113/134 - Train Accuracy:  0.630, Validation Accuracy:  0.596, Loss:  0.976
Epoch   0 Batch  114/134 - Train Accuracy:  0.582, Validation Accuracy:  0.600, Loss:  1.100
Epoch   0 Batch  115/134 - Train Accuracy:  0.511, Validation Accuracy:  0.558, Loss:  1.151
Epoch   0 Batch  116/134 - Train Accuracy:  0.544, Validation Accuracy:  0.562, Loss:  1.086
Epoch   0 Batch  117/134 - Train Accuracy:  0.590, Validation Accuracy:  0.613, Loss:  1.058
Epoch   0 Batch  118/134 - Train Accuracy:  0.591, Validation Accuracy:  0.606, Loss:  1.057
Epoch   0 Batch  119/134 - Train Accuracy:  0.617, Validation Accuracy:  0.597, Loss:  0.963
Epoch   0 Batch  120/134 - Train Accuracy:  0.577, Validation Accuracy:  0.599, Loss:  1.038
Epoch   0 Batch  121/134 - Train Accuracy:  0.595, Validation Accuracy:  0.602, Loss:  1.018
Epoch   0 Batch  122/134 - Train Accuracy:  0.571, Validation Accuracy:  0.594, Loss:  1.026
Epoch   0 Batch  123/134 - Train Accuracy:  0.585, Validation Accuracy:  0.600, Loss:  1.012
Epoch   0 Batch  124/134 - Train Accuracy:  0.611, Validation Accuracy:  0.602, Loss:  0.949
Epoch   0 Batch  125/134 - Train Accuracy:  0.588, Validation Accuracy:  0.601, Loss:  0.997
Epoch   0 Batch  126/134 - Train Accuracy:  0.571, Validation Accuracy:  0.593, Loss:  0.992
Epoch   0 Batch  127/134 - Train Accuracy:  0.582, Validation Accuracy:  0.592, Loss:  0.952
Epoch   0 Batch  128/134 - Train Accuracy:  0.572, Validation Accuracy:  0.605, Loss:  0.982
Epoch   0 Batch  129/134 - Train Accuracy:  0.604, Validation Accuracy:  0.615, Loss:  0.964
Epoch   0 Batch  130/134 - Train Accuracy:  0.603, Validation Accuracy:  0.626, Loss:  0.973
Epoch   0 Batch  131/134 - Train Accuracy:  0.595, Validation Accuracy:  0.627, Loss:  0.993
Epoch   0 Batch  132/134 - Train Accuracy:  0.604, Validation Accuracy:  0.621, Loss:  0.935
Epoch   1 Batch    0/134 - Train Accuracy:  0.596, Validation Accuracy:  0.616, Loss:  0.941
Epoch   1 Batch    1/134 - Train Accuracy:  0.588, Validation Accuracy:  0.622, Loss:  0.968
Epoch   1 Batch    2/134 - Train Accuracy:  0.601, Validation Accuracy:  0.625, Loss:  0.912
Epoch   1 Batch    3/134 - Train Accuracy:  0.610, Validation Accuracy:  0.624, Loss:  0.919
Epoch   1 Batch    4/134 - Train Accuracy:  0.598, Validation Accuracy:  0.618, Loss:  0.914
Epoch   1 Batch    5/134 - Train Accuracy:  0.595, Validation Accuracy:  0.619, Loss:  0.916
Epoch   1 Batch    6/134 - Train Accuracy:  0.607, Validation Accuracy:  0.614, Loss:  0.852
Epoch   1 Batch    7/134 - Train Accuracy:  0.605, Validation Accuracy:  0.621, Loss:  0.894
Epoch   1 Batch    8/134 - Train Accuracy:  0.610, Validation Accuracy:  0.627, Loss:  0.869
Epoch   1 Batch    9/134 - Train Accuracy:  0.650, Validation Accuracy:  0.638, Loss:  0.830
Epoch   1 Batch   10/134 - Train Accuracy:  0.624, Validation Accuracy:  0.637, Loss:  0.871
Epoch   1 Batch   11/134 - Train Accuracy:  0.616, Validation Accuracy:  0.635, Loss:  0.867
Epoch   1 Batch   12/134 - Train Accuracy:  0.629, Validation Accuracy:  0.629, Loss:  0.820
Epoch   1 Batch   13/134 - Train Accuracy:  0.593, Validation Accuracy:  0.626, Loss:  0.863
Epoch   1 Batch   14/134 - Train Accuracy:  0.622, Validation Accuracy:  0.626, Loss:  0.838
Epoch   1 Batch   15/134 - Train Accuracy:  0.615, Validation Accuracy:  0.632, Loss:  0.823
Epoch   1 Batch   16/134 - Train Accuracy:  0.626, Validation Accuracy:  0.641, Loss:  0.826
Epoch   1 Batch   17/134 - Train Accuracy:  0.630, Validation Accuracy:  0.646, Loss:  0.835
Epoch   1 Batch   18/134 - Train Accuracy:  0.628, Validation Accuracy:  0.647, Loss:  0.819
Epoch   1 Batch   19/134 - Train Accuracy:  0.631, Validation Accuracy:  0.650, Loss:  0.818
Epoch   1 Batch   20/134 - Train Accuracy:  0.647, Validation Accuracy:  0.650, Loss:  0.785
Epoch   1 Batch   21/134 - Train Accuracy:  0.628, Validation Accuracy:  0.642, Loss:  0.802
Epoch   1 Batch   22/134 - Train Accuracy:  0.603, Validation Accuracy:  0.638, Loss:  0.832
Epoch   1 Batch   23/134 - Train Accuracy:  0.650, Validation Accuracy:  0.644, Loss:  0.745
Epoch   1 Batch   24/134 - Train Accuracy:  0.611, Validation Accuracy:  0.650, Loss:  0.820
Epoch   1 Batch   25/134 - Train Accuracy:  0.618, Validation Accuracy:  0.647, Loss:  0.780
Epoch   1 Batch   26/134 - Train Accuracy:  0.624, Validation Accuracy:  0.648, Loss:  0.807
Epoch   1 Batch   27/134 - Train Accuracy:  0.649, Validation Accuracy:  0.657, Loss:  0.774
Epoch   1 Batch   28/134 - Train Accuracy:  0.642, Validation Accuracy:  0.660, Loss:  0.769
Epoch   1 Batch   29/134 - Train Accuracy:  0.652, Validation Accuracy:  0.657, Loss:  0.723
Epoch   1 Batch   30/134 - Train Accuracy:  0.652, Validation Accuracy:  0.660, Loss:  0.726
Epoch   1 Batch   31/134 - Train Accuracy:  0.635, Validation Accuracy:  0.662, Loss:  0.763
Epoch   1 Batch   32/134 - Train Accuracy:  0.646, Validation Accuracy:  0.661, Loss:  0.729
Epoch   1 Batch   33/134 - Train Accuracy:  0.628, Validation Accuracy:  0.661, Loss:  0.751
Epoch   1 Batch   34/134 - Train Accuracy:  0.639, Validation Accuracy:  0.655, Loss:  0.767
Epoch   1 Batch   35/134 - Train Accuracy:  0.652, Validation Accuracy:  0.657, Loss:  0.723
Epoch   1 Batch   36/134 - Train Accuracy:  0.648, Validation Accuracy:  0.659, Loss:  0.737
Epoch   1 Batch   37/134 - Train Accuracy:  0.640, Validation Accuracy:  0.668, Loss:  0.746
Epoch   1 Batch   38/134 - Train Accuracy:  0.659, Validation Accuracy:  0.670, Loss:  0.725
Epoch   1 Batch   39/134 - Train Accuracy:  0.656, Validation Accuracy:  0.672, Loss:  0.714
Epoch   1 Batch   40/134 - Train Accuracy:  0.660, Validation Accuracy:  0.669, Loss:  0.717
Epoch   1 Batch   41/134 - Train Accuracy:  0.652, Validation Accuracy:  0.667, Loss:  0.716
Epoch   1 Batch   42/134 - Train Accuracy:  0.634, Validation Accuracy:  0.664, Loss:  0.709
Epoch   1 Batch   43/134 - Train Accuracy:  0.645, Validation Accuracy:  0.668, Loss:  0.719
Epoch   1 Batch   44/134 - Train Accuracy:  0.651, Validation Accuracy:  0.671, Loss:  0.708
Epoch   1 Batch   45/134 - Train Accuracy:  0.660, Validation Accuracy:  0.678, Loss:  0.699
Epoch   1 Batch   46/134 - Train Accuracy:  0.669, Validation Accuracy:  0.677, Loss:  0.676
Epoch   1 Batch   47/134 - Train Accuracy:  0.651, Validation Accuracy:  0.675, Loss:  0.693
Epoch   1 Batch   48/134 - Train Accuracy:  0.660, Validation Accuracy:  0.680, Loss:  0.696
Epoch   1 Batch   49/134 - Train Accuracy:  0.675, Validation Accuracy:  0.679, Loss:  0.678
Epoch   1 Batch   50/134 - Train Accuracy:  0.654, Validation Accuracy:  0.676, Loss:  0.691
Epoch   1 Batch   51/134 - Train Accuracy:  0.657, Validation Accuracy:  0.676, Loss:  0.676
Epoch   1 Batch   52/134 - Train Accuracy:  0.647, Validation Accuracy:  0.677, Loss:  0.687
Epoch   1 Batch   53/134 - Train Accuracy:  0.649, Validation Accuracy:  0.678, Loss:  0.681
Epoch   1 Batch   54/134 - Train Accuracy:  0.644, Validation Accuracy:  0.680, Loss:  0.673
Epoch   1 Batch   55/134 - Train Accuracy:  0.658, Validation Accuracy:  0.678, Loss:  0.672
Epoch   1 Batch   56/134 - Train Accuracy:  0.671, Validation Accuracy:  0.680, Loss:  0.638
Epoch   1 Batch   57/134 - Train Accuracy:  0.662, Validation Accuracy:  0.681, Loss:  0.660
Epoch   1 Batch   58/134 - Train Accuracy:  0.670, Validation Accuracy:  0.683, Loss:  0.649
Epoch   1 Batch   59/134 - Train Accuracy:  0.661, Validation Accuracy:  0.685, Loss:  0.684
Epoch   1 Batch   60/134 - Train Accuracy:  0.669, Validation Accuracy:  0.685, Loss:  0.647
Epoch   1 Batch   61/134 - Train Accuracy:  0.664, Validation Accuracy:  0.678, Loss:  0.643
Epoch   1 Batch   62/134 - Train Accuracy:  0.659, Validation Accuracy:  0.681, Loss:  0.640
Epoch   1 Batch   63/134 - Train Accuracy:  0.668, Validation Accuracy:  0.686, Loss:  0.638
Epoch   1 Batch   64/134 - Train Accuracy:  0.664, Validation Accuracy:  0.685, Loss:  0.641
Epoch   1 Batch   65/134 - Train Accuracy:  0.662, Validation Accuracy:  0.681, Loss:  0.640
Epoch   1 Batch   66/134 - Train Accuracy:  0.664, Validation Accuracy:  0.684, Loss:  0.623
Epoch   1 Batch   67/134 - Train Accuracy:  0.637, Validation Accuracy:  0.680, Loss:  0.682
Epoch   1 Batch   68/134 - Train Accuracy:  0.654, Validation Accuracy:  0.683, Loss:  0.659
Epoch   1 Batch   69/134 - Train Accuracy:  0.679, Validation Accuracy:  0.688, Loss:  0.627
Epoch   1 Batch   70/134 - Train Accuracy:  0.686, Validation Accuracy:  0.692, Loss:  0.609
Epoch   1 Batch   71/134 - Train Accuracy:  0.675, Validation Accuracy:  0.694, Loss:  0.609
Epoch   1 Batch   72/134 - Train Accuracy:  0.675, Validation Accuracy:  0.695, Loss:  0.611
Epoch   1 Batch   73/134 - Train Accuracy:  0.679, Validation Accuracy:  0.695, Loss:  0.592
Epoch   1 Batch   74/134 - Train Accuracy:  0.682, Validation Accuracy:  0.691, Loss:  0.615
Epoch   1 Batch   75/134 - Train Accuracy:  0.692, Validation Accuracy:  0.695, Loss:  0.585
Epoch   1 Batch   76/134 - Train Accuracy:  0.678, Validation Accuracy:  0.693, Loss:  0.602
Epoch   1 Batch   77/134 - Train Accuracy:  0.703, Validation Accuracy:  0.696, Loss:  0.581
Epoch   1 Batch   78/134 - Train Accuracy:  0.675, Validation Accuracy:  0.697, Loss:  0.608
Epoch   1 Batch   79/134 - Train Accuracy:  0.682, Validation Accuracy:  0.702, Loss:  0.598
Epoch   1 Batch   80/134 - Train Accuracy:  0.674, Validation Accuracy:  0.695, Loss:  0.603
Epoch   1 Batch   81/134 - Train Accuracy:  0.683, Validation Accuracy:  0.697, Loss:  0.589
Epoch   1 Batch   82/134 - Train Accuracy:  0.705, Validation Accuracy:  0.702, Loss:  0.555
Epoch   1 Batch   83/134 - Train Accuracy:  0.692, Validation Accuracy:  0.706, Loss:  0.589
Epoch   1 Batch   84/134 - Train Accuracy:  0.686, Validation Accuracy:  0.703, Loss:  0.586
Epoch   1 Batch   85/134 - Train Accuracy:  0.688, Validation Accuracy:  0.704, Loss:  0.590
Epoch   1 Batch   86/134 - Train Accuracy:  0.694, Validation Accuracy:  0.709, Loss:  0.573
Epoch   1 Batch   87/134 - Train Accuracy:  0.685, Validation Accuracy:  0.705, Loss:  0.598
Epoch   1 Batch   88/134 - Train Accuracy:  0.708, Validation Accuracy:  0.709, Loss:  0.549
Epoch   1 Batch   89/134 - Train Accuracy:  0.689, Validation Accuracy:  0.702, Loss:  0.573
Epoch   1 Batch   90/134 - Train Accuracy:  0.685, Validation Accuracy:  0.710, Loss:  0.583
Epoch   1 Batch   91/134 - Train Accuracy:  0.735, Validation Accuracy:  0.713, Loss:  0.498
Epoch   1 Batch   92/134 - Train Accuracy:  0.701, Validation Accuracy:  0.711, Loss:  0.560
Epoch   1 Batch   93/134 - Train Accuracy:  0.706, Validation Accuracy:  0.715, Loss:  0.558
Epoch   1 Batch   94/134 - Train Accuracy:  0.701, Validation Accuracy:  0.721, Loss:  0.551
Epoch   1 Batch   95/134 - Train Accuracy:  0.705, Validation Accuracy:  0.720, Loss:  0.568
Epoch   1 Batch   96/134 - Train Accuracy:  0.709, Validation Accuracy:  0.712, Loss:  0.554
Epoch   1 Batch   97/134 - Train Accuracy:  0.680, Validation Accuracy:  0.714, Loss:  0.556
Epoch   1 Batch   98/134 - Train Accuracy:  0.682, Validation Accuracy:  0.716, Loss:  0.588
Epoch   1 Batch   99/134 - Train Accuracy:  0.722, Validation Accuracy:  0.730, Loss:  0.550
Epoch   1 Batch  100/134 - Train Accuracy:  0.712, Validation Accuracy:  0.731, Loss:  0.549
Epoch   1 Batch  101/134 - Train Accuracy:  0.698, Validation Accuracy:  0.721, Loss:  0.578
Epoch   1 Batch  102/134 - Train Accuracy:  0.705, Validation Accuracy:  0.717, Loss:  0.536
Epoch   1 Batch  103/134 - Train Accuracy:  0.725, Validation Accuracy:  0.720, Loss:  0.517
Epoch   1 Batch  104/134 - Train Accuracy:  0.719, Validation Accuracy:  0.722, Loss:  0.524
Epoch   1 Batch  105/134 - Train Accuracy:  0.705, Validation Accuracy:  0.721, Loss:  0.544
Epoch   1 Batch  106/134 - Train Accuracy:  0.706, Validation Accuracy:  0.721, Loss:  0.536
Epoch   1 Batch  107/134 - Train Accuracy:  0.728, Validation Accuracy:  0.719, Loss:  0.512
Epoch   1 Batch  108/134 - Train Accuracy:  0.702, Validation Accuracy:  0.727, Loss:  0.556
Epoch   1 Batch  109/134 - Train Accuracy:  0.729, Validation Accuracy:  0.724, Loss:  0.499
Epoch   1 Batch  110/134 - Train Accuracy:  0.735, Validation Accuracy:  0.718, Loss:  0.518
Epoch   1 Batch  111/134 - Train Accuracy:  0.721, Validation Accuracy:  0.724, Loss:  0.523
Epoch   1 Batch  112/134 - Train Accuracy:  0.723, Validation Accuracy:  0.724, Loss:  0.509
Epoch   1 Batch  113/134 - Train Accuracy:  0.739, Validation Accuracy:  0.724, Loss:  0.461
Epoch   1 Batch  114/134 - Train Accuracy:  0.703, Validation Accuracy:  0.722, Loss:  0.515
Epoch   1 Batch  115/134 - Train Accuracy:  0.692, Validation Accuracy:  0.725, Loss:  0.553
Epoch   1 Batch  116/134 - Train Accuracy:  0.735, Validation Accuracy:  0.739, Loss:  0.510
Epoch   1 Batch  117/134 - Train Accuracy:  0.733, Validation Accuracy:  0.741, Loss:  0.504
Epoch   1 Batch  118/134 - Train Accuracy:  0.726, Validation Accuracy:  0.741, Loss:  0.506
Epoch   1 Batch  119/134 - Train Accuracy:  0.746, Validation Accuracy:  0.732, Loss:  0.460
Epoch   1 Batch  120/134 - Train Accuracy:  0.722, Validation Accuracy:  0.732, Loss:  0.516
Epoch   1 Batch  121/134 - Train Accuracy:  0.725, Validation Accuracy:  0.730, Loss:  0.495
Epoch   1 Batch  122/134 - Train Accuracy:  0.719, Validation Accuracy:  0.743, Loss:  0.510
Epoch   1 Batch  123/134 - Train Accuracy:  0.737, Validation Accuracy:  0.745, Loss:  0.488
Epoch   1 Batch  124/134 - Train Accuracy:  0.747, Validation Accuracy:  0.741, Loss:  0.463
Epoch   1 Batch  125/134 - Train Accuracy:  0.729, Validation Accuracy:  0.729, Loss:  0.498
Epoch   1 Batch  126/134 - Train Accuracy:  0.701, Validation Accuracy:  0.726, Loss:  0.501
Epoch   1 Batch  127/134 - Train Accuracy:  0.731, Validation Accuracy:  0.728, Loss:  0.483
Epoch   1 Batch  128/134 - Train Accuracy:  0.714, Validation Accuracy:  0.723, Loss:  0.504
Epoch   1 Batch  129/134 - Train Accuracy:  0.720, Validation Accuracy:  0.734, Loss:  0.493
Epoch   1 Batch  130/134 - Train Accuracy:  0.740, Validation Accuracy:  0.749, Loss:  0.494
Epoch   1 Batch  131/134 - Train Accuracy:  0.737, Validation Accuracy:  0.747, Loss:  0.507
Epoch   1 Batch  132/134 - Train Accuracy:  0.738, Validation Accuracy:  0.740, Loss:  0.472
Epoch   2 Batch    0/134 - Train Accuracy:  0.732, Validation Accuracy:  0.744, Loss:  0.485
Epoch   2 Batch    1/134 - Train Accuracy:  0.715, Validation Accuracy:  0.737, Loss:  0.499
Epoch   2 Batch    2/134 - Train Accuracy:  0.738, Validation Accuracy:  0.744, Loss:  0.475
Epoch   2 Batch    3/134 - Train Accuracy:  0.734, Validation Accuracy:  0.749, Loss:  0.476
Epoch   2 Batch    4/134 - Train Accuracy:  0.742, Validation Accuracy:  0.752, Loss:  0.489
Epoch   2 Batch    5/134 - Train Accuracy:  0.734, Validation Accuracy:  0.750, Loss:  0.487
Epoch   2 Batch    6/134 - Train Accuracy:  0.751, Validation Accuracy:  0.762, Loss:  0.449
Epoch   2 Batch    7/134 - Train Accuracy:  0.733, Validation Accuracy:  0.755, Loss:  0.471
Epoch   2 Batch    8/134 - Train Accuracy:  0.745, Validation Accuracy:  0.749, Loss:  0.460
Epoch   2 Batch    9/134 - Train Accuracy:  0.762, Validation Accuracy:  0.753, Loss:  0.445
Epoch   2 Batch   10/134 - Train Accuracy:  0.727, Validation Accuracy:  0.750, Loss:  0.459
Epoch   2 Batch   11/134 - Train Accuracy:  0.749, Validation Accuracy:  0.749, Loss:  0.464
Epoch   2 Batch   12/134 - Train Accuracy:  0.741, Validation Accuracy:  0.756, Loss:  0.440
Epoch   2 Batch   13/134 - Train Accuracy:  0.724, Validation Accuracy:  0.755, Loss:  0.470
Epoch   2 Batch   14/134 - Train Accuracy:  0.739, Validation Accuracy:  0.746, Loss:  0.459
Epoch   2 Batch   15/134 - Train Accuracy:  0.747, Validation Accuracy:  0.745, Loss:  0.446
Epoch   2 Batch   16/134 - Train Accuracy:  0.745, Validation Accuracy:  0.744, Loss:  0.445
Epoch   2 Batch   17/134 - Train Accuracy:  0.735, Validation Accuracy:  0.750, Loss:  0.461
Epoch   2 Batch   18/134 - Train Accuracy:  0.749, Validation Accuracy:  0.761, Loss:  0.453
Epoch   2 Batch   19/134 - Train Accuracy:  0.756, Validation Accuracy:  0.767, Loss:  0.456
Epoch   2 Batch   20/134 - Train Accuracy:  0.764, Validation Accuracy:  0.766, Loss:  0.442
Epoch   2 Batch   21/134 - Train Accuracy:  0.755, Validation Accuracy:  0.757, Loss:  0.444
Epoch   2 Batch   22/134 - Train Accuracy:  0.740, Validation Accuracy:  0.753, Loss:  0.469
Epoch   2 Batch   23/134 - Train Accuracy:  0.765, Validation Accuracy:  0.748, Loss:  0.416
Epoch   2 Batch   24/134 - Train Accuracy:  0.723, Validation Accuracy:  0.741, Loss:  0.451
Epoch   2 Batch   25/134 - Train Accuracy:  0.744, Validation Accuracy:  0.749, Loss:  0.428
Epoch   2 Batch   26/134 - Train Accuracy:  0.765, Validation Accuracy:  0.763, Loss:  0.454
Epoch   2 Batch   27/134 - Train Accuracy:  0.755, Validation Accuracy:  0.760, Loss:  0.437
Epoch   2 Batch   28/134 - Train Accuracy:  0.756, Validation Accuracy:  0.769, Loss:  0.442
Epoch   2 Batch   29/134 - Train Accuracy:  0.787, Validation Accuracy:  0.774, Loss:  0.406
Epoch   2 Batch   30/134 - Train Accuracy:  0.769, Validation Accuracy:  0.770, Loss:  0.412
Epoch   2 Batch   31/134 - Train Accuracy:  0.748, Validation Accuracy:  0.761, Loss:  0.424
Epoch   2 Batch   32/134 - Train Accuracy:  0.750, Validation Accuracy:  0.755, Loss:  0.410
Epoch   2 Batch   33/134 - Train Accuracy:  0.762, Validation Accuracy:  0.776, Loss:  0.427
Epoch   2 Batch   34/134 - Train Accuracy:  0.771, Validation Accuracy:  0.772, Loss:  0.433
Epoch   2 Batch   35/134 - Train Accuracy:  0.784, Validation Accuracy:  0.768, Loss:  0.410
Epoch   2 Batch   36/134 - Train Accuracy:  0.769, Validation Accuracy:  0.770, Loss:  0.422
Epoch   2 Batch   37/134 - Train Accuracy:  0.767, Validation Accuracy:  0.779, Loss:  0.419
Epoch   2 Batch   38/134 - Train Accuracy:  0.772, Validation Accuracy:  0.776, Loss:  0.416
Epoch   2 Batch   39/134 - Train Accuracy:  0.766, Validation Accuracy:  0.777, Loss:  0.410
Epoch   2 Batch   40/134 - Train Accuracy:  0.782, Validation Accuracy:  0.780, Loss:  0.407
Epoch   2 Batch   41/134 - Train Accuracy:  0.776, Validation Accuracy:  0.778, Loss:  0.412
Epoch   2 Batch   42/134 - Train Accuracy:  0.773, Validation Accuracy:  0.770, Loss:  0.406
Epoch   2 Batch   43/134 - Train Accuracy:  0.748, Validation Accuracy:  0.761, Loss:  0.411
Epoch   2 Batch   44/134 - Train Accuracy:  0.762, Validation Accuracy:  0.766, Loss:  0.404
Epoch   2 Batch   45/134 - Train Accuracy:  0.787, Validation Accuracy:  0.781, Loss:  0.397
Epoch   2 Batch   46/134 - Train Accuracy:  0.790, Validation Accuracy:  0.789, Loss:  0.390
Epoch   2 Batch   47/134 - Train Accuracy:  0.776, Validation Accuracy:  0.786, Loss:  0.403
Epoch   2 Batch   48/134 - Train Accuracy:  0.783, Validation Accuracy:  0.785, Loss:  0.396
Epoch   2 Batch   49/134 - Train Accuracy:  0.786, Validation Accuracy:  0.782, Loss:  0.385
Epoch   2 Batch   50/134 - Train Accuracy:  0.775, Validation Accuracy:  0.787, Loss:  0.398
Epoch   2 Batch   51/134 - Train Accuracy:  0.780, Validation Accuracy:  0.784, Loss:  0.391
Epoch   2 Batch   52/134 - Train Accuracy:  0.771, Validation Accuracy:  0.775, Loss:  0.385
Epoch   2 Batch   53/134 - Train Accuracy:  0.775, Validation Accuracy:  0.782, Loss:  0.389
Epoch   2 Batch   54/134 - Train Accuracy:  0.777, Validation Accuracy:  0.793, Loss:  0.382
Epoch   2 Batch   55/134 - Train Accuracy:  0.784, Validation Accuracy:  0.795, Loss:  0.389
Epoch   2 Batch   56/134 - Train Accuracy:  0.790, Validation Accuracy:  0.793, Loss:  0.370
Epoch   2 Batch   57/134 - Train Accuracy:  0.799, Validation Accuracy:  0.796, Loss:  0.384
Epoch   2 Batch   58/134 - Train Accuracy:  0.808, Validation Accuracy:  0.796, Loss:  0.374
Epoch   2 Batch   59/134 - Train Accuracy:  0.785, Validation Accuracy:  0.794, Loss:  0.396
Epoch   2 Batch   60/134 - Train Accuracy:  0.785, Validation Accuracy:  0.787, Loss:  0.368
Epoch   2 Batch   61/134 - Train Accuracy:  0.787, Validation Accuracy:  0.782, Loss:  0.372
Epoch   2 Batch   62/134 - Train Accuracy:  0.789, Validation Accuracy:  0.798, Loss:  0.372
Epoch   2 Batch   63/134 - Train Accuracy:  0.800, Validation Accuracy:  0.802, Loss:  0.367
Epoch   2 Batch   64/134 - Train Accuracy:  0.788, Validation Accuracy:  0.800, Loss:  0.372
Epoch   2 Batch   65/134 - Train Accuracy:  0.786, Validation Accuracy:  0.801, Loss:  0.372
Epoch   2 Batch   66/134 - Train Accuracy:  0.807, Validation Accuracy:  0.804, Loss:  0.362
Epoch   2 Batch   67/134 - Train Accuracy:  0.777, Validation Accuracy:  0.806, Loss:  0.389
Epoch   2 Batch   68/134 - Train Accuracy:  0.797, Validation Accuracy:  0.806, Loss:  0.388
Epoch   2 Batch   69/134 - Train Accuracy:  0.815, Validation Accuracy:  0.807, Loss:  0.364
Epoch   2 Batch   70/134 - Train Accuracy:  0.811, Validation Accuracy:  0.804, Loss:  0.350
Epoch   2 Batch   71/134 - Train Accuracy:  0.797, Validation Accuracy:  0.804, Loss:  0.361
Epoch   2 Batch   72/134 - Train Accuracy:  0.798, Validation Accuracy:  0.806, Loss:  0.353
Epoch   2 Batch   73/134 - Train Accuracy:  0.807, Validation Accuracy:  0.815, Loss:  0.346
Epoch   2 Batch   74/134 - Train Accuracy:  0.804, Validation Accuracy:  0.815, Loss:  0.361
Epoch   2 Batch   75/134 - Train Accuracy:  0.810, Validation Accuracy:  0.804, Loss:  0.345
Epoch   2 Batch   76/134 - Train Accuracy:  0.812, Validation Accuracy:  0.811, Loss:  0.354
Epoch   2 Batch   77/134 - Train Accuracy:  0.812, Validation Accuracy:  0.813, Loss:  0.342
Epoch   2 Batch   78/134 - Train Accuracy:  0.801, Validation Accuracy:  0.808, Loss:  0.355
Epoch   2 Batch   79/134 - Train Accuracy:  0.806, Validation Accuracy:  0.814, Loss:  0.353
Epoch   2 Batch   80/134 - Train Accuracy:  0.803, Validation Accuracy:  0.814, Loss:  0.352
Epoch   2 Batch   81/134 - Train Accuracy:  0.819, Validation Accuracy:  0.814, Loss:  0.350
Epoch   2 Batch   82/134 - Train Accuracy:  0.824, Validation Accuracy:  0.814, Loss:  0.324
Epoch   2 Batch   83/134 - Train Accuracy:  0.813, Validation Accuracy:  0.824, Loss:  0.345
Epoch   2 Batch   84/134 - Train Accuracy:  0.812, Validation Accuracy:  0.824, Loss:  0.341
Epoch   2 Batch   85/134 - Train Accuracy:  0.816, Validation Accuracy:  0.813, Loss:  0.347
Epoch   2 Batch   86/134 - Train Accuracy:  0.823, Validation Accuracy:  0.824, Loss:  0.343
Epoch   2 Batch   87/134 - Train Accuracy:  0.813, Validation Accuracy:  0.818, Loss:  0.353
Epoch   2 Batch   88/134 - Train Accuracy:  0.826, Validation Accuracy:  0.825, Loss:  0.332
Epoch   2 Batch   89/134 - Train Accuracy:  0.822, Validation Accuracy:  0.822, Loss:  0.335
Epoch   2 Batch   90/134 - Train Accuracy:  0.824, Validation Accuracy:  0.824, Loss:  0.343
Epoch   2 Batch   91/134 - Train Accuracy:  0.853, Validation Accuracy:  0.829, Loss:  0.290
Epoch   2 Batch   92/134 - Train Accuracy:  0.829, Validation Accuracy:  0.825, Loss:  0.324
Epoch   2 Batch   93/134 - Train Accuracy:  0.825, Validation Accuracy:  0.827, Loss:  0.319
Epoch   2 Batch   94/134 - Train Accuracy:  0.828, Validation Accuracy:  0.830, Loss:  0.323
Epoch   2 Batch   95/134 - Train Accuracy:  0.824, Validation Accuracy:  0.832, Loss:  0.335
Epoch   2 Batch   96/134 - Train Accuracy:  0.825, Validation Accuracy:  0.822, Loss:  0.329
Epoch   2 Batch   97/134 - Train Accuracy:  0.822, Validation Accuracy:  0.820, Loss:  0.328
Epoch   2 Batch   98/134 - Train Accuracy:  0.815, Validation Accuracy:  0.822, Loss:  0.342
Epoch   2 Batch   99/134 - Train Accuracy:  0.823, Validation Accuracy:  0.821, Loss:  0.326
Epoch   2 Batch  100/134 - Train Accuracy:  0.829, Validation Accuracy:  0.827, Loss:  0.324
Epoch   2 Batch  101/134 - Train Accuracy:  0.822, Validation Accuracy:  0.828, Loss:  0.339
Epoch   2 Batch  102/134 - Train Accuracy:  0.824, Validation Accuracy:  0.826, Loss:  0.314
Epoch   2 Batch  103/134 - Train Accuracy:  0.834, Validation Accuracy:  0.829, Loss:  0.307
Epoch   2 Batch  104/134 - Train Accuracy:  0.831, Validation Accuracy:  0.831, Loss:  0.311
Epoch   2 Batch  105/134 - Train Accuracy:  0.828, Validation Accuracy:  0.834, Loss:  0.322
Epoch   2 Batch  106/134 - Train Accuracy:  0.826, Validation Accuracy:  0.834, Loss:  0.319
Epoch   2 Batch  107/134 - Train Accuracy:  0.843, Validation Accuracy:  0.837, Loss:  0.302
Epoch   2 Batch  108/134 - Train Accuracy:  0.831, Validation Accuracy:  0.833, Loss:  0.329
Epoch   2 Batch  109/134 - Train Accuracy:  0.845, Validation Accuracy:  0.837, Loss:  0.295
Epoch   2 Batch  110/134 - Train Accuracy:  0.845, Validation Accuracy:  0.841, Loss:  0.310
Epoch   2 Batch  111/134 - Train Accuracy:  0.835, Validation Accuracy:  0.839, Loss:  0.312
Epoch   2 Batch  112/134 - Train Accuracy:  0.838, Validation Accuracy:  0.843, Loss:  0.303
Epoch   2 Batch  113/134 - Train Accuracy:  0.851, Validation Accuracy:  0.839, Loss:  0.278
Epoch   2 Batch  114/134 - Train Accuracy:  0.832, Validation Accuracy:  0.836, Loss:  0.301
Epoch   2 Batch  115/134 - Train Accuracy:  0.829, Validation Accuracy:  0.846, Loss:  0.321
Epoch   2 Batch  116/134 - Train Accuracy:  0.843, Validation Accuracy:  0.839, Loss:  0.307
Epoch   2 Batch  117/134 - Train Accuracy:  0.847, Validation Accuracy:  0.836, Loss:  0.295
Epoch   2 Batch  118/134 - Train Accuracy:  0.841, Validation Accuracy:  0.842, Loss:  0.299
Epoch   2 Batch  119/134 - Train Accuracy:  0.858, Validation Accuracy:  0.846, Loss:  0.270
Epoch   2 Batch  120/134 - Train Accuracy:  0.828, Validation Accuracy:  0.850, Loss:  0.305
Epoch   2 Batch  121/134 - Train Accuracy:  0.844, Validation Accuracy:  0.848, Loss:  0.293
Epoch   2 Batch  122/134 - Train Accuracy:  0.833, Validation Accuracy:  0.842, Loss:  0.307
Epoch   2 Batch  123/134 - Train Accuracy:  0.842, Validation Accuracy:  0.842, Loss:  0.285
Epoch   2 Batch  124/134 - Train Accuracy:  0.854, Validation Accuracy:  0.844, Loss:  0.274
Epoch   2 Batch  125/134 - Train Accuracy:  0.856, Validation Accuracy:  0.844, Loss:  0.290
Epoch   2 Batch  126/134 - Train Accuracy:  0.833, Validation Accuracy:  0.849, Loss:  0.294
Epoch   2 Batch  127/134 - Train Accuracy:  0.841, Validation Accuracy:  0.859, Loss:  0.281
Epoch   2 Batch  128/134 - Train Accuracy:  0.842, Validation Accuracy:  0.853, Loss:  0.296
Epoch   2 Batch  129/134 - Train Accuracy:  0.844, Validation Accuracy:  0.852, Loss:  0.291
Epoch   2 Batch  130/134 - Train Accuracy:  0.853, Validation Accuracy:  0.854, Loss:  0.291
Epoch   2 Batch  131/134 - Train Accuracy:  0.836, Validation Accuracy:  0.850, Loss:  0.302
Epoch   2 Batch  132/134 - Train Accuracy:  0.856, Validation Accuracy:  0.855, Loss:  0.284
Epoch   3 Batch    0/134 - Train Accuracy:  0.847, Validation Accuracy:  0.848, Loss:  0.279
Epoch   3 Batch    1/134 - Train Accuracy:  0.842, Validation Accuracy:  0.850, Loss:  0.296
Epoch   3 Batch    2/134 - Train Accuracy:  0.859, Validation Accuracy:  0.858, Loss:  0.274
Epoch   3 Batch    3/134 - Train Accuracy:  0.855, Validation Accuracy:  0.854, Loss:  0.277
Epoch   3 Batch    4/134 - Train Accuracy:  0.853, Validation Accuracy:  0.857, Loss:  0.286
Epoch   3 Batch    5/134 - Train Accuracy:  0.845, Validation Accuracy:  0.857, Loss:  0.281
Epoch   3 Batch    6/134 - Train Accuracy:  0.854, Validation Accuracy:  0.855, Loss:  0.261
Epoch   3 Batch    7/134 - Train Accuracy:  0.850, Validation Accuracy:  0.862, Loss:  0.274
Epoch   3 Batch    8/134 - Train Accuracy:  0.856, Validation Accuracy:  0.862, Loss:  0.270
Epoch   3 Batch    9/134 - Train Accuracy:  0.864, Validation Accuracy:  0.865, Loss:  0.261
Epoch   3 Batch   10/134 - Train Accuracy:  0.856, Validation Accuracy:  0.866, Loss:  0.269
Epoch   3 Batch   11/134 - Train Accuracy:  0.861, Validation Accuracy:  0.870, Loss:  0.268
Epoch   3 Batch   12/134 - Train Accuracy:  0.857, Validation Accuracy:  0.868, Loss:  0.259
Epoch   3 Batch   13/134 - Train Accuracy:  0.857, Validation Accuracy:  0.867, Loss:  0.270
Epoch   3 Batch   14/134 - Train Accuracy:  0.871, Validation Accuracy:  0.865, Loss:  0.262
Epoch   3 Batch   15/134 - Train Accuracy:  0.869, Validation Accuracy:  0.862, Loss:  0.259
Epoch   3 Batch   16/134 - Train Accuracy:  0.871, Validation Accuracy:  0.867, Loss:  0.257
Epoch   3 Batch   17/134 - Train Accuracy:  0.849, Validation Accuracy:  0.862, Loss:  0.275
Epoch   3 Batch   18/134 - Train Accuracy:  0.857, Validation Accuracy:  0.860, Loss:  0.267
Epoch   3 Batch   19/134 - Train Accuracy:  0.859, Validation Accuracy:  0.864, Loss:  0.261
Epoch   3 Batch   20/134 - Train Accuracy:  0.871, Validation Accuracy:  0.866, Loss:  0.251
Epoch   3 Batch   21/134 - Train Accuracy:  0.860, Validation Accuracy:  0.862, Loss:  0.259
Epoch   3 Batch   22/134 - Train Accuracy:  0.860, Validation Accuracy:  0.869, Loss:  0.268
Epoch   3 Batch   23/134 - Train Accuracy:  0.876, Validation Accuracy:  0.871, Loss:  0.238
Epoch   3 Batch   24/134 - Train Accuracy:  0.856, Validation Accuracy:  0.863, Loss:  0.261
Epoch   3 Batch   25/134 - Train Accuracy:  0.861, Validation Accuracy:  0.864, Loss:  0.250
Epoch   3 Batch   26/134 - Train Accuracy:  0.868, Validation Accuracy:  0.867, Loss:  0.265
Epoch   3 Batch   27/134 - Train Accuracy:  0.867, Validation Accuracy:  0.869, Loss:  0.248
Epoch   3 Batch   28/134 - Train Accuracy:  0.864, Validation Accuracy:  0.868, Loss:  0.257
Epoch   3 Batch   29/134 - Train Accuracy:  0.880, Validation Accuracy:  0.869, Loss:  0.232
Epoch   3 Batch   30/134 - Train Accuracy:  0.867, Validation Accuracy:  0.868, Loss:  0.242
Epoch   3 Batch   31/134 - Train Accuracy:  0.867, Validation Accuracy:  0.869, Loss:  0.247
Epoch   3 Batch   32/134 - Train Accuracy:  0.864, Validation Accuracy:  0.870, Loss:  0.235
Epoch   3 Batch   33/134 - Train Accuracy:  0.863, Validation Accuracy:  0.870, Loss:  0.252
Epoch   3 Batch   34/134 - Train Accuracy:  0.855, Validation Accuracy:  0.865, Loss:  0.261
Epoch   3 Batch   35/134 - Train Accuracy:  0.872, Validation Accuracy:  0.870, Loss:  0.243
Epoch   3 Batch   36/134 - Train Accuracy:  0.869, Validation Accuracy:  0.876, Loss:  0.241
Epoch   3 Batch   37/134 - Train Accuracy:  0.869, Validation Accuracy:  0.875, Loss:  0.244
Epoch   3 Batch   38/134 - Train Accuracy:  0.867, Validation Accuracy:  0.875, Loss:  0.242
Epoch   3 Batch   39/134 - Train Accuracy:  0.870, Validation Accuracy:  0.875, Loss:  0.236
Epoch   3 Batch   40/134 - Train Accuracy:  0.877, Validation Accuracy:  0.880, Loss:  0.234
Epoch   3 Batch   41/134 - Train Accuracy:  0.868, Validation Accuracy:  0.878, Loss:  0.240
Epoch   3 Batch   42/134 - Train Accuracy:  0.874, Validation Accuracy:  0.877, Loss:  0.236
Epoch   3 Batch   43/134 - Train Accuracy:  0.863, Validation Accuracy:  0.881, Loss:  0.248
Epoch   3 Batch   44/134 - Train Accuracy:  0.878, Validation Accuracy:  0.883, Loss:  0.234
Epoch   3 Batch   45/134 - Train Accuracy:  0.881, Validation Accuracy:  0.881, Loss:  0.230
Epoch   3 Batch   46/134 - Train Accuracy:  0.879, Validation Accuracy:  0.886, Loss:  0.230
Epoch   3 Batch   47/134 - Train Accuracy:  0.878, Validation Accuracy:  0.886, Loss:  0.227
Epoch   3 Batch   48/134 - Train Accuracy:  0.878, Validation Accuracy:  0.885, Loss:  0.232
Epoch   3 Batch   49/134 - Train Accuracy:  0.887, Validation Accuracy:  0.884, Loss:  0.223
Epoch   3 Batch   50/134 - Train Accuracy:  0.881, Validation Accuracy:  0.880, Loss:  0.227
Epoch   3 Batch   51/134 - Train Accuracy:  0.876, Validation Accuracy:  0.884, Loss:  0.224
Epoch   3 Batch   52/134 - Train Accuracy:  0.877, Validation Accuracy:  0.887, Loss:  0.231
Epoch   3 Batch   53/134 - Train Accuracy:  0.895, Validation Accuracy:  0.890, Loss:  0.222
Epoch   3 Batch   54/134 - Train Accuracy:  0.865, Validation Accuracy:  0.883, Loss:  0.222
Epoch   3 Batch   55/134 - Train Accuracy:  0.876, Validation Accuracy:  0.883, Loss:  0.224
Epoch   3 Batch   56/134 - Train Accuracy:  0.887, Validation Accuracy:  0.882, Loss:  0.213
Epoch   3 Batch   57/134 - Train Accuracy:  0.886, Validation Accuracy:  0.881, Loss:  0.220
Epoch   3 Batch   58/134 - Train Accuracy:  0.893, Validation Accuracy:  0.882, Loss:  0.214
Epoch   3 Batch   59/134 - Train Accuracy:  0.877, Validation Accuracy:  0.880, Loss:  0.226
Epoch   3 Batch   60/134 - Train Accuracy:  0.885, Validation Accuracy:  0.887, Loss:  0.216
Epoch   3 Batch   61/134 - Train Accuracy:  0.886, Validation Accuracy:  0.891, Loss:  0.212
Epoch   3 Batch   62/134 - Train Accuracy:  0.874, Validation Accuracy:  0.889, Loss:  0.212
Epoch   3 Batch   63/134 - Train Accuracy:  0.885, Validation Accuracy:  0.887, Loss:  0.213
Epoch   3 Batch   64/134 - Train Accuracy:  0.879, Validation Accuracy:  0.888, Loss:  0.216
Epoch   3 Batch   65/134 - Train Accuracy:  0.877, Validation Accuracy:  0.888, Loss:  0.216
Epoch   3 Batch   66/134 - Train Accuracy:  0.890, Validation Accuracy:  0.888, Loss:  0.210
Epoch   3 Batch   67/134 - Train Accuracy:  0.876, Validation Accuracy:  0.892, Loss:  0.227
Epoch   3 Batch   68/134 - Train Accuracy:  0.888, Validation Accuracy:  0.895, Loss:  0.221
Epoch   3 Batch   69/134 - Train Accuracy:  0.893, Validation Accuracy:  0.899, Loss:  0.210
Epoch   3 Batch   70/134 - Train Accuracy:  0.892, Validation Accuracy:  0.902, Loss:  0.206
Epoch   3 Batch   71/134 - Train Accuracy:  0.895, Validation Accuracy:  0.895, Loss:  0.205
Epoch   3 Batch   72/134 - Train Accuracy:  0.889, Validation Accuracy:  0.893, Loss:  0.208
Epoch   3 Batch   73/134 - Train Accuracy:  0.886, Validation Accuracy:  0.900, Loss:  0.205
Epoch   3 Batch   74/134 - Train Accuracy:  0.873, Validation Accuracy:  0.897, Loss:  0.213
Epoch   3 Batch   75/134 - Train Accuracy:  0.895, Validation Accuracy:  0.902, Loss:  0.194
Epoch   3 Batch   76/134 - Train Accuracy:  0.902, Validation Accuracy:  0.905, Loss:  0.200
Epoch   3 Batch   77/134 - Train Accuracy:  0.895, Validation Accuracy:  0.903, Loss:  0.195
Epoch   3 Batch   78/134 - Train Accuracy:  0.890, Validation Accuracy:  0.902, Loss:  0.200
Epoch   3 Batch   79/134 - Train Accuracy:  0.898, Validation Accuracy:  0.904, Loss:  0.198
Epoch   3 Batch   80/134 - Train Accuracy:  0.892, Validation Accuracy:  0.899, Loss:  0.200
Epoch   3 Batch   81/134 - Train Accuracy:  0.903, Validation Accuracy:  0.902, Loss:  0.203
Epoch   3 Batch   82/134 - Train Accuracy:  0.896, Validation Accuracy:  0.901, Loss:  0.188
Epoch   3 Batch   83/134 - Train Accuracy:  0.884, Validation Accuracy:  0.896, Loss:  0.200
Epoch   3 Batch   84/134 - Train Accuracy:  0.884, Validation Accuracy:  0.899, Loss:  0.194
Epoch   3 Batch   85/134 - Train Accuracy:  0.899, Validation Accuracy:  0.901, Loss:  0.200
Epoch   3 Batch   86/134 - Train Accuracy:  0.904, Validation Accuracy:  0.897, Loss:  0.194
Epoch   3 Batch   87/134 - Train Accuracy:  0.878, Validation Accuracy:  0.897, Loss:  0.208
Epoch   3 Batch   88/134 - Train Accuracy:  0.904, Validation Accuracy:  0.899, Loss:  0.186
Epoch   3 Batch   89/134 - Train Accuracy:  0.896, Validation Accuracy:  0.903, Loss:  0.187
Epoch   3 Batch   90/134 - Train Accuracy:  0.901, Validation Accuracy:  0.906, Loss:  0.201
Epoch   3 Batch   91/134 - Train Accuracy:  0.909, Validation Accuracy:  0.897, Loss:  0.163
Epoch   3 Batch   92/134 - Train Accuracy:  0.904, Validation Accuracy:  0.905, Loss:  0.187
Epoch   3 Batch   93/134 - Train Accuracy:  0.896, Validation Accuracy:  0.901, Loss:  0.181
Epoch   3 Batch   94/134 - Train Accuracy:  0.896, Validation Accuracy:  0.900, Loss:  0.184
Epoch   3 Batch   95/134 - Train Accuracy:  0.893, Validation Accuracy:  0.903, Loss:  0.187
Epoch   3 Batch   96/134 - Train Accuracy:  0.898, Validation Accuracy:  0.904, Loss:  0.187
Epoch   3 Batch   97/134 - Train Accuracy:  0.894, Validation Accuracy:  0.903, Loss:  0.185
Epoch   3 Batch   98/134 - Train Accuracy:  0.889, Validation Accuracy:  0.898, Loss:  0.196
Epoch   3 Batch   99/134 - Train Accuracy:  0.888, Validation Accuracy:  0.897, Loss:  0.190
Epoch   3 Batch  100/134 - Train Accuracy:  0.904, Validation Accuracy:  0.905, Loss:  0.184
Epoch   3 Batch  101/134 - Train Accuracy:  0.904, Validation Accuracy:  0.899, Loss:  0.192
Epoch   3 Batch  102/134 - Train Accuracy:  0.887, Validation Accuracy:  0.892, Loss:  0.181
Epoch   3 Batch  103/134 - Train Accuracy:  0.902, Validation Accuracy:  0.900, Loss:  0.176
Epoch   3 Batch  104/134 - Train Accuracy:  0.908, Validation Accuracy:  0.903, Loss:  0.173
Epoch   3 Batch  105/134 - Train Accuracy:  0.905, Validation Accuracy:  0.906, Loss:  0.188
Epoch   3 Batch  106/134 - Train Accuracy:  0.895, Validation Accuracy:  0.906, Loss:  0.184
Epoch   3 Batch  107/134 - Train Accuracy:  0.908, Validation Accuracy:  0.907, Loss:  0.173
Epoch   3 Batch  108/134 - Train Accuracy:  0.898, Validation Accuracy:  0.908, Loss:  0.185
Epoch   3 Batch  109/134 - Train Accuracy:  0.910, Validation Accuracy:  0.906, Loss:  0.166
Epoch   3 Batch  110/134 - Train Accuracy:  0.903, Validation Accuracy:  0.905, Loss:  0.174
Epoch   3 Batch  111/134 - Train Accuracy:  0.896, Validation Accuracy:  0.907, Loss:  0.175
Epoch   3 Batch  112/134 - Train Accuracy:  0.895, Validation Accuracy:  0.914, Loss:  0.172
Epoch   3 Batch  113/134 - Train Accuracy:  0.910, Validation Accuracy:  0.910, Loss:  0.163
Epoch   3 Batch  114/134 - Train Accuracy:  0.905, Validation Accuracy:  0.908, Loss:  0.169
Epoch   3 Batch  115/134 - Train Accuracy:  0.892, Validation Accuracy:  0.911, Loss:  0.175
Epoch   3 Batch  116/134 - Train Accuracy:  0.907, Validation Accuracy:  0.913, Loss:  0.174
Epoch   3 Batch  117/134 - Train Accuracy:  0.914, Validation Accuracy:  0.912, Loss:  0.162
Epoch   3 Batch  118/134 - Train Accuracy:  0.906, Validation Accuracy:  0.910, Loss:  0.167
Epoch   3 Batch  119/134 - Train Accuracy:  0.917, Validation Accuracy:  0.912, Loss:  0.152
Epoch   3 Batch  120/134 - Train Accuracy:  0.904, Validation Accuracy:  0.912, Loss:  0.171
Epoch   3 Batch  121/134 - Train Accuracy:  0.912, Validation Accuracy:  0.912, Loss:  0.157
Epoch   3 Batch  122/134 - Train Accuracy:  0.907, Validation Accuracy:  0.914, Loss:  0.165
Epoch   3 Batch  123/134 - Train Accuracy:  0.914, Validation Accuracy:  0.916, Loss:  0.157
Epoch   3 Batch  124/134 - Train Accuracy:  0.919, Validation Accuracy:  0.913, Loss:  0.152
Epoch   3 Batch  125/134 - Train Accuracy:  0.918, Validation Accuracy:  0.915, Loss:  0.156
Epoch   3 Batch  126/134 - Train Accuracy:  0.900, Validation Accuracy:  0.918, Loss:  0.166
Epoch   3 Batch  127/134 - Train Accuracy:  0.907, Validation Accuracy:  0.914, Loss:  0.154
Epoch   3 Batch  128/134 - Train Accuracy:  0.907, Validation Accuracy:  0.912, Loss:  0.175
Epoch   3 Batch  129/134 - Train Accuracy:  0.904, Validation Accuracy:  0.916, Loss:  0.166
Epoch   3 Batch  130/134 - Train Accuracy:  0.913, Validation Accuracy:  0.920, Loss:  0.165
Epoch   3 Batch  131/134 - Train Accuracy:  0.894, Validation Accuracy:  0.921, Loss:  0.166
Epoch   3 Batch  132/134 - Train Accuracy:  0.911, Validation Accuracy:  0.913, Loss:  0.157
Epoch   4 Batch    0/134 - Train Accuracy:  0.914, Validation Accuracy:  0.917, Loss:  0.152
Epoch   4 Batch    1/134 - Train Accuracy:  0.900, Validation Accuracy:  0.904, Loss:  0.160
Epoch   4 Batch    2/134 - Train Accuracy:  0.920, Validation Accuracy:  0.917, Loss:  0.151
Epoch   4 Batch    3/134 - Train Accuracy:  0.916, Validation Accuracy:  0.915, Loss:  0.154
Epoch   4 Batch    4/134 - Train Accuracy:  0.915, Validation Accuracy:  0.921, Loss:  0.156
Epoch   4 Batch    5/134 - Train Accuracy:  0.918, Validation Accuracy:  0.927, Loss:  0.159
Epoch   4 Batch    6/134 - Train Accuracy:  0.918, Validation Accuracy:  0.922, Loss:  0.141
Epoch   4 Batch    7/134 - Train Accuracy:  0.913, Validation Accuracy:  0.924, Loss:  0.153
Epoch   4 Batch    8/134 - Train Accuracy:  0.920, Validation Accuracy:  0.923, Loss:  0.143
Epoch   4 Batch    9/134 - Train Accuracy:  0.924, Validation Accuracy:  0.922, Loss:  0.139
Epoch   4 Batch   10/134 - Train Accuracy:  0.915, Validation Accuracy:  0.923, Loss:  0.153
Epoch   4 Batch   11/134 - Train Accuracy:  0.918, Validation Accuracy:  0.923, Loss:  0.151
Epoch   4 Batch   12/134 - Train Accuracy:  0.912, Validation Accuracy:  0.927, Loss:  0.142
Epoch   4 Batch   13/134 - Train Accuracy:  0.905, Validation Accuracy:  0.932, Loss:  0.151
Epoch   4 Batch   14/134 - Train Accuracy:  0.925, Validation Accuracy:  0.935, Loss:  0.143
Epoch   4 Batch   15/134 - Train Accuracy:  0.927, Validation Accuracy:  0.927, Loss:  0.139
Epoch   4 Batch   16/134 - Train Accuracy:  0.923, Validation Accuracy:  0.927, Loss:  0.137
Epoch   4 Batch   17/134 - Train Accuracy:  0.910, Validation Accuracy:  0.927, Loss:  0.149
Epoch   4 Batch   18/134 - Train Accuracy:  0.923, Validation Accuracy:  0.925, Loss:  0.139
Epoch   4 Batch   19/134 - Train Accuracy:  0.922, Validation Accuracy:  0.926, Loss:  0.136
Epoch   4 Batch   20/134 - Train Accuracy:  0.929, Validation Accuracy:  0.929, Loss:  0.131
Epoch   4 Batch   21/134 - Train Accuracy:  0.928, Validation Accuracy:  0.926, Loss:  0.143
Epoch   4 Batch   22/134 - Train Accuracy:  0.918, Validation Accuracy:  0.925, Loss:  0.143
Epoch   4 Batch   23/134 - Train Accuracy:  0.930, Validation Accuracy:  0.923, Loss:  0.129
Epoch   4 Batch   24/134 - Train Accuracy:  0.914, Validation Accuracy:  0.922, Loss:  0.144
Epoch   4 Batch   25/134 - Train Accuracy:  0.917, Validation Accuracy:  0.920, Loss:  0.134
Epoch   4 Batch   26/134 - Train Accuracy:  0.926, Validation Accuracy:  0.922, Loss:  0.145
Epoch   4 Batch   27/134 - Train Accuracy:  0.927, Validation Accuracy:  0.925, Loss:  0.135
Epoch   4 Batch   28/134 - Train Accuracy:  0.927, Validation Accuracy:  0.927, Loss:  0.138
Epoch   4 Batch   29/134 - Train Accuracy:  0.931, Validation Accuracy:  0.928, Loss:  0.122
Epoch   4 Batch   30/134 - Train Accuracy:  0.922, Validation Accuracy:  0.926, Loss:  0.132
Epoch   4 Batch   31/134 - Train Accuracy:  0.926, Validation Accuracy:  0.934, Loss:  0.134
Epoch   4 Batch   32/134 - Train Accuracy:  0.923, Validation Accuracy:  0.933, Loss:  0.130
Epoch   4 Batch   33/134 - Train Accuracy:  0.917, Validation Accuracy:  0.936, Loss:  0.142
Epoch   4 Batch   34/134 - Train Accuracy:  0.917, Validation Accuracy:  0.935, Loss:  0.141
Epoch   4 Batch   35/134 - Train Accuracy:  0.925, Validation Accuracy:  0.931, Loss:  0.136
Epoch   4 Batch   36/134 - Train Accuracy:  0.926, Validation Accuracy:  0.929, Loss:  0.130
Epoch   4 Batch   37/134 - Train Accuracy:  0.922, Validation Accuracy:  0.935, Loss:  0.131
Epoch   4 Batch   38/134 - Train Accuracy:  0.931, Validation Accuracy:  0.931, Loss:  0.137
Epoch   4 Batch   39/134 - Train Accuracy:  0.922, Validation Accuracy:  0.932, Loss:  0.126
Epoch   4 Batch   40/134 - Train Accuracy:  0.927, Validation Accuracy:  0.928, Loss:  0.126
Epoch   4 Batch   41/134 - Train Accuracy:  0.919, Validation Accuracy:  0.928, Loss:  0.135
Epoch   4 Batch   42/134 - Train Accuracy:  0.929, Validation Accuracy:  0.930, Loss:  0.123
Epoch   4 Batch   43/134 - Train Accuracy:  0.918, Validation Accuracy:  0.929, Loss:  0.132
Epoch   4 Batch   44/134 - Train Accuracy:  0.929, Validation Accuracy:  0.931, Loss:  0.125
Epoch   4 Batch   45/134 - Train Accuracy:  0.928, Validation Accuracy:  0.930, Loss:  0.119
Epoch   4 Batch   46/134 - Train Accuracy:  0.929, Validation Accuracy:  0.932, Loss:  0.126
Epoch   4 Batch   47/134 - Train Accuracy:  0.925, Validation Accuracy:  0.935, Loss:  0.122
Epoch   4 Batch   48/134 - Train Accuracy:  0.934, Validation Accuracy:  0.934, Loss:  0.126
Epoch   4 Batch   49/134 - Train Accuracy:  0.934, Validation Accuracy:  0.938, Loss:  0.118
Epoch   4 Batch   50/134 - Train Accuracy:  0.931, Validation Accuracy:  0.937, Loss:  0.123
Epoch   4 Batch   51/134 - Train Accuracy:  0.932, Validation Accuracy:  0.936, Loss:  0.126
Epoch   4 Batch   52/134 - Train Accuracy:  0.933, Validation Accuracy:  0.940, Loss:  0.121
Epoch   4 Batch   53/134 - Train Accuracy:  0.947, Validation Accuracy:  0.938, Loss:  0.113
Epoch   4 Batch   54/134 - Train Accuracy:  0.926, Validation Accuracy:  0.933, Loss:  0.124
Epoch   4 Batch   55/134 - Train Accuracy:  0.932, Validation Accuracy:  0.936, Loss:  0.121
Epoch   4 Batch   56/134 - Train Accuracy:  0.933, Validation Accuracy:  0.935, Loss:  0.115
Epoch   4 Batch   57/134 - Train Accuracy:  0.929, Validation Accuracy:  0.925, Loss:  0.118
Epoch   4 Batch   58/134 - Train Accuracy:  0.932, Validation Accuracy:  0.928, Loss:  0.112
Epoch   4 Batch   59/134 - Train Accuracy:  0.923, Validation Accuracy:  0.935, Loss:  0.125
Epoch   4 Batch   60/134 - Train Accuracy:  0.933, Validation Accuracy:  0.936, Loss:  0.112
Epoch   4 Batch   61/134 - Train Accuracy:  0.940, Validation Accuracy:  0.934, Loss:  0.105
Epoch   4 Batch   62/134 - Train Accuracy:  0.931, Validation Accuracy:  0.938, Loss:  0.113
Epoch   4 Batch   63/134 - Train Accuracy:  0.931, Validation Accuracy:  0.935, Loss:  0.115
Epoch   4 Batch   64/134 - Train Accuracy:  0.928, Validation Accuracy:  0.931, Loss:  0.114
Epoch   4 Batch   65/134 - Train Accuracy:  0.922, Validation Accuracy:  0.935, Loss:  0.117
Epoch   4 Batch   66/134 - Train Accuracy:  0.928, Validation Accuracy:  0.935, Loss:  0.112
Epoch   4 Batch   67/134 - Train Accuracy:  0.925, Validation Accuracy:  0.941, Loss:  0.121
Epoch   4 Batch   68/134 - Train Accuracy:  0.926, Validation Accuracy:  0.943, Loss:  0.117
Epoch   4 Batch   69/134 - Train Accuracy:  0.937, Validation Accuracy:  0.943, Loss:  0.112
Epoch   4 Batch   70/134 - Train Accuracy:  0.937, Validation Accuracy:  0.944, Loss:  0.109
Epoch   4 Batch   71/134 - Train Accuracy:  0.941, Validation Accuracy:  0.940, Loss:  0.104
Epoch   4 Batch   72/134 - Train Accuracy:  0.939, Validation Accuracy:  0.939, Loss:  0.106
Epoch   4 Batch   73/134 - Train Accuracy:  0.937, Validation Accuracy:  0.937, Loss:  0.111
Epoch   4 Batch   74/134 - Train Accuracy:  0.931, Validation Accuracy:  0.937, Loss:  0.121
Epoch   4 Batch   75/134 - Train Accuracy:  0.941, Validation Accuracy:  0.943, Loss:  0.108
Epoch   4 Batch   76/134 - Train Accuracy:  0.946, Validation Accuracy:  0.947, Loss:  0.104
Epoch   4 Batch   77/134 - Train Accuracy:  0.937, Validation Accuracy:  0.942, Loss:  0.106
Epoch   4 Batch   78/134 - Train Accuracy:  0.929, Validation Accuracy:  0.938, Loss:  0.108
Epoch   4 Batch   79/134 - Train Accuracy:  0.941, Validation Accuracy:  0.941, Loss:  0.109
Epoch   4 Batch   80/134 - Train Accuracy:  0.945, Validation Accuracy:  0.941, Loss:  0.106
Epoch   4 Batch   81/134 - Train Accuracy:  0.945, Validation Accuracy:  0.939, Loss:  0.113
Epoch   4 Batch   82/134 - Train Accuracy:  0.942, Validation Accuracy:  0.940, Loss:  0.099
Epoch   4 Batch   83/134 - Train Accuracy:  0.936, Validation Accuracy:  0.941, Loss:  0.107
Epoch   4 Batch   84/134 - Train Accuracy:  0.932, Validation Accuracy:  0.938, Loss:  0.105
Epoch   4 Batch   85/134 - Train Accuracy:  0.941, Validation Accuracy:  0.940, Loss:  0.109
Epoch   4 Batch   86/134 - Train Accuracy:  0.940, Validation Accuracy:  0.941, Loss:  0.103
Epoch   4 Batch   87/134 - Train Accuracy:  0.923, Validation Accuracy:  0.941, Loss:  0.120
Epoch   4 Batch   88/134 - Train Accuracy:  0.944, Validation Accuracy:  0.939, Loss:  0.097
Epoch   4 Batch   89/134 - Train Accuracy:  0.938, Validation Accuracy:  0.942, Loss:  0.102
Epoch   4 Batch   90/134 - Train Accuracy:  0.935, Validation Accuracy:  0.944, Loss:  0.110
Epoch   4 Batch   91/134 - Train Accuracy:  0.949, Validation Accuracy:  0.946, Loss:  0.088
Epoch   4 Batch   92/134 - Train Accuracy:  0.939, Validation Accuracy:  0.941, Loss:  0.101
Epoch   4 Batch   93/134 - Train Accuracy:  0.932, Validation Accuracy:  0.942, Loss:  0.098
Epoch   4 Batch   94/134 - Train Accuracy:  0.935, Validation Accuracy:  0.944, Loss:  0.102
Epoch   4 Batch   95/134 - Train Accuracy:  0.931, Validation Accuracy:  0.948, Loss:  0.108
Epoch   4 Batch   96/134 - Train Accuracy:  0.930, Validation Accuracy:  0.946, Loss:  0.102
Epoch   4 Batch   97/134 - Train Accuracy:  0.934, Validation Accuracy:  0.946, Loss:  0.099
Epoch   4 Batch   98/134 - Train Accuracy:  0.925, Validation Accuracy:  0.937, Loss:  0.102
Epoch   4 Batch   99/134 - Train Accuracy:  0.936, Validation Accuracy:  0.937, Loss:  0.097
Epoch   4 Batch  100/134 - Train Accuracy:  0.937, Validation Accuracy:  0.945, Loss:  0.100
Epoch   4 Batch  101/134 - Train Accuracy:  0.936, Validation Accuracy:  0.944, Loss:  0.105
Epoch   4 Batch  102/134 - Train Accuracy:  0.931, Validation Accuracy:  0.939, Loss:  0.100
Epoch   4 Batch  103/134 - Train Accuracy:  0.945, Validation Accuracy:  0.940, Loss:  0.097
Epoch   4 Batch  104/134 - Train Accuracy:  0.941, Validation Accuracy:  0.942, Loss:  0.093
Epoch   4 Batch  105/134 - Train Accuracy:  0.935, Validation Accuracy:  0.943, Loss:  0.106
Epoch   4 Batch  106/134 - Train Accuracy:  0.935, Validation Accuracy:  0.940, Loss:  0.097
Epoch   4 Batch  107/134 - Train Accuracy:  0.941, Validation Accuracy:  0.942, Loss:  0.098
Epoch   4 Batch  108/134 - Train Accuracy:  0.938, Validation Accuracy:  0.943, Loss:  0.099
Epoch   4 Batch  109/134 - Train Accuracy:  0.935, Validation Accuracy:  0.942, Loss:  0.091
Epoch   4 Batch  110/134 - Train Accuracy:  0.940, Validation Accuracy:  0.942, Loss:  0.097
Epoch   4 Batch  111/134 - Train Accuracy:  0.938, Validation Accuracy:  0.946, Loss:  0.099
Epoch   4 Batch  112/134 - Train Accuracy:  0.939, Validation Accuracy:  0.947, Loss:  0.092
Epoch   4 Batch  113/134 - Train Accuracy:  0.944, Validation Accuracy:  0.946, Loss:  0.090
Epoch   4 Batch  114/134 - Train Accuracy:  0.942, Validation Accuracy:  0.947, Loss:  0.098
Epoch   4 Batch  115/134 - Train Accuracy:  0.931, Validation Accuracy:  0.944, Loss:  0.099
Epoch   4 Batch  116/134 - Train Accuracy:  0.939, Validation Accuracy:  0.941, Loss:  0.097
Epoch   4 Batch  117/134 - Train Accuracy:  0.945, Validation Accuracy:  0.940, Loss:  0.088
Epoch   4 Batch  118/134 - Train Accuracy:  0.936, Validation Accuracy:  0.941, Loss:  0.093
Epoch   4 Batch  119/134 - Train Accuracy:  0.947, Validation Accuracy:  0.946, Loss:  0.092
Epoch   4 Batch  120/134 - Train Accuracy:  0.942, Validation Accuracy:  0.944, Loss:  0.102
Epoch   4 Batch  121/134 - Train Accuracy:  0.938, Validation Accuracy:  0.947, Loss:  0.095
Epoch   4 Batch  122/134 - Train Accuracy:  0.932, Validation Accuracy:  0.940, Loss:  0.099
Epoch   4 Batch  123/134 - Train Accuracy:  0.942, Validation Accuracy:  0.939, Loss:  0.092
Epoch   4 Batch  124/134 - Train Accuracy:  0.938, Validation Accuracy:  0.939, Loss:  0.092
Epoch   4 Batch  125/134 - Train Accuracy:  0.948, Validation Accuracy:  0.946, Loss:  0.088
Epoch   4 Batch  126/134 - Train Accuracy:  0.934, Validation Accuracy:  0.949, Loss:  0.098
Epoch   4 Batch  127/134 - Train Accuracy:  0.935, Validation Accuracy:  0.945, Loss:  0.087
Epoch   4 Batch  128/134 - Train Accuracy:  0.937, Validation Accuracy:  0.946, Loss:  0.103
Epoch   4 Batch  129/134 - Train Accuracy:  0.941, Validation Accuracy:  0.949, Loss:  0.093
Epoch   4 Batch  130/134 - Train Accuracy:  0.938, Validation Accuracy:  0.945, Loss:  0.092
Epoch   4 Batch  131/134 - Train Accuracy:  0.940, Validation Accuracy:  0.952, Loss:  0.098
Epoch   4 Batch  132/134 - Train Accuracy:  0.933, Validation Accuracy:  0.942, Loss:  0.083
Epoch   5 Batch    0/134 - Train Accuracy:  0.935, Validation Accuracy:  0.947, Loss:  0.088
Epoch   5 Batch    1/134 - Train Accuracy:  0.933, Validation Accuracy:  0.946, Loss:  0.094
Epoch   5 Batch    2/134 - Train Accuracy:  0.952, Validation Accuracy:  0.956, Loss:  0.087
Epoch   5 Batch    3/134 - Train Accuracy:  0.949, Validation Accuracy:  0.952, Loss:  0.091
Epoch   5 Batch    4/134 - Train Accuracy:  0.943, Validation Accuracy:  0.950, Loss:  0.092
Epoch   5 Batch    5/134 - Train Accuracy:  0.941, Validation Accuracy:  0.952, Loss:  0.095
Epoch   5 Batch    6/134 - Train Accuracy:  0.944, Validation Accuracy:  0.952, Loss:  0.078
Epoch   5 Batch    7/134 - Train Accuracy:  0.938, Validation Accuracy:  0.953, Loss:  0.088
Epoch   5 Batch    8/134 - Train Accuracy:  0.942, Validation Accuracy:  0.949, Loss:  0.086
Epoch   5 Batch    9/134 - Train Accuracy:  0.952, Validation Accuracy:  0.951, Loss:  0.080
Epoch   5 Batch   10/134 - Train Accuracy:  0.939, Validation Accuracy:  0.948, Loss:  0.086
Epoch   5 Batch   11/134 - Train Accuracy:  0.941, Validation Accuracy:  0.947, Loss:  0.089
Epoch   5 Batch   12/134 - Train Accuracy:  0.944, Validation Accuracy:  0.950, Loss:  0.083
Epoch   5 Batch   13/134 - Train Accuracy:  0.937, Validation Accuracy:  0.946, Loss:  0.085
Epoch   5 Batch   14/134 - Train Accuracy:  0.946, Validation Accuracy:  0.949, Loss:  0.083
Epoch   5 Batch   15/134 - Train Accuracy:  0.948, Validation Accuracy:  0.950, Loss:  0.083
Epoch   5 Batch   16/134 - Train Accuracy:  0.940, Validation Accuracy:  0.948, Loss:  0.081
Epoch   5 Batch   17/134 - Train Accuracy:  0.941, Validation Accuracy:  0.949, Loss:  0.091
Epoch   5 Batch   18/134 - Train Accuracy:  0.944, Validation Accuracy:  0.945, Loss:  0.085
Epoch   5 Batch   19/134 - Train Accuracy:  0.943, Validation Accuracy:  0.945, Loss:  0.084
Epoch   5 Batch   20/134 - Train Accuracy:  0.948, Validation Accuracy:  0.947, Loss:  0.081
Epoch   5 Batch   21/134 - Train Accuracy:  0.946, Validation Accuracy:  0.947, Loss:  0.085
Epoch   5 Batch   22/134 - Train Accuracy:  0.943, Validation Accuracy:  0.947, Loss:  0.085
Epoch   5 Batch   23/134 - Train Accuracy:  0.946, Validation Accuracy:  0.946, Loss:  0.074
Epoch   5 Batch   24/134 - Train Accuracy:  0.940, Validation Accuracy:  0.949, Loss:  0.090
Epoch   5 Batch   25/134 - Train Accuracy:  0.940, Validation Accuracy:  0.950, Loss:  0.075
Epoch   5 Batch   26/134 - Train Accuracy:  0.948, Validation Accuracy:  0.951, Loss:  0.084
Epoch   5 Batch   27/134 - Train Accuracy:  0.937, Validation Accuracy:  0.948, Loss:  0.082
Epoch   5 Batch   28/134 - Train Accuracy:  0.948, Validation Accuracy:  0.950, Loss:  0.086
Epoch   5 Batch   29/134 - Train Accuracy:  0.953, Validation Accuracy:  0.954, Loss:  0.071
Epoch   5 Batch   30/134 - Train Accuracy:  0.942, Validation Accuracy:  0.956, Loss:  0.082
Epoch   5 Batch   31/134 - Train Accuracy:  0.950, Validation Accuracy:  0.951, Loss:  0.082
Epoch   5 Batch   32/134 - Train Accuracy:  0.947, Validation Accuracy:  0.948, Loss:  0.079
Epoch   5 Batch   33/134 - Train Accuracy:  0.940, Validation Accuracy:  0.946, Loss:  0.089
Epoch   5 Batch   34/134 - Train Accuracy:  0.943, Validation Accuracy:  0.954, Loss:  0.088
Epoch   5 Batch   35/134 - Train Accuracy:  0.946, Validation Accuracy:  0.953, Loss:  0.087
Epoch   5 Batch   36/134 - Train Accuracy:  0.943, Validation Accuracy:  0.955, Loss:  0.080
Epoch   5 Batch   37/134 - Train Accuracy:  0.941, Validation Accuracy:  0.952, Loss:  0.084
Epoch   5 Batch   38/134 - Train Accuracy:  0.949, Validation Accuracy:  0.952, Loss:  0.079
Epoch   5 Batch   39/134 - Train Accuracy:  0.944, Validation Accuracy:  0.954, Loss:  0.083
Epoch   5 Batch   40/134 - Train Accuracy:  0.947, Validation Accuracy:  0.949, Loss:  0.081
Epoch   5 Batch   41/134 - Train Accuracy:  0.941, Validation Accuracy:  0.951, Loss:  0.082
Epoch   5 Batch   42/134 - Train Accuracy:  0.944, Validation Accuracy:  0.950, Loss:  0.078
Epoch   5 Batch   43/134 - Train Accuracy:  0.941, Validation Accuracy:  0.953, Loss:  0.086
Epoch   5 Batch   44/134 - Train Accuracy:  0.949, Validation Accuracy:  0.957, Loss:  0.079
Epoch   5 Batch   45/134 - Train Accuracy:  0.952, Validation Accuracy:  0.955, Loss:  0.074
Epoch   5 Batch   46/134 - Train Accuracy:  0.950, Validation Accuracy:  0.955, Loss:  0.082
Epoch   5 Batch   47/134 - Train Accuracy:  0.948, Validation Accuracy:  0.951, Loss:  0.078
Epoch   5 Batch   48/134 - Train Accuracy:  0.946, Validation Accuracy:  0.952, Loss:  0.076
Epoch   5 Batch   49/134 - Train Accuracy:  0.951, Validation Accuracy:  0.950, Loss:  0.077
Epoch   5 Batch   50/134 - Train Accuracy:  0.950, Validation Accuracy:  0.950, Loss:  0.081
Epoch   5 Batch   51/134 - Train Accuracy:  0.949, Validation Accuracy:  0.952, Loss:  0.081
Epoch   5 Batch   52/134 - Train Accuracy:  0.949, Validation Accuracy:  0.952, Loss:  0.081
Epoch   5 Batch   53/134 - Train Accuracy:  0.961, Validation Accuracy:  0.951, Loss:  0.074
Epoch   5 Batch   54/134 - Train Accuracy:  0.944, Validation Accuracy:  0.945, Loss:  0.081
Epoch   5 Batch   55/134 - Train Accuracy:  0.949, Validation Accuracy:  0.944, Loss:  0.080
Epoch   5 Batch   56/134 - Train Accuracy:  0.946, Validation Accuracy:  0.950, Loss:  0.073
Epoch   5 Batch   57/134 - Train Accuracy:  0.950, Validation Accuracy:  0.947, Loss:  0.077
Epoch   5 Batch   58/134 - Train Accuracy:  0.948, Validation Accuracy:  0.947, Loss:  0.070
Epoch   5 Batch   59/134 - Train Accuracy:  0.940, Validation Accuracy:  0.951, Loss:  0.080
Epoch   5 Batch   60/134 - Train Accuracy:  0.944, Validation Accuracy:  0.949, Loss:  0.074
Epoch   5 Batch   61/134 - Train Accuracy:  0.948, Validation Accuracy:  0.948, Loss:  0.070
Epoch   5 Batch   62/134 - Train Accuracy:  0.950, Validation Accuracy:  0.948, Loss:  0.070
Epoch   5 Batch   63/134 - Train Accuracy:  0.945, Validation Accuracy:  0.951, Loss:  0.073
Epoch   5 Batch   64/134 - Train Accuracy:  0.947, Validation Accuracy:  0.950, Loss:  0.074
Epoch   5 Batch   65/134 - Train Accuracy:  0.941, Validation Accuracy:  0.951, Loss:  0.075
Epoch   5 Batch   66/134 - Train Accuracy:  0.949, Validation Accuracy:  0.951, Loss:  0.072
Epoch   5 Batch   67/134 - Train Accuracy:  0.943, Validation Accuracy:  0.952, Loss:  0.081
Epoch   5 Batch   68/134 - Train Accuracy:  0.939, Validation Accuracy:  0.951, Loss:  0.080
Epoch   5 Batch   69/134 - Train Accuracy:  0.946, Validation Accuracy:  0.954, Loss:  0.074
Epoch   5 Batch   70/134 - Train Accuracy:  0.948, Validation Accuracy:  0.951, Loss:  0.073
Epoch   5 Batch   71/134 - Train Accuracy:  0.959, Validation Accuracy:  0.954, Loss:  0.066
Epoch   5 Batch   72/134 - Train Accuracy:  0.949, Validation Accuracy:  0.952, Loss:  0.069
Epoch   5 Batch   73/134 - Train Accuracy:  0.954, Validation Accuracy:  0.956, Loss:  0.076
Epoch   5 Batch   74/134 - Train Accuracy:  0.945, Validation Accuracy:  0.959, Loss:  0.082
Epoch   5 Batch   75/134 - Train Accuracy:  0.953, Validation Accuracy:  0.958, Loss:  0.073
Epoch   5 Batch   76/134 - Train Accuracy:  0.958, Validation Accuracy:  0.958, Loss:  0.070
Epoch   5 Batch   77/134 - Train Accuracy:  0.954, Validation Accuracy:  0.956, Loss:  0.071
Epoch   5 Batch   78/134 - Train Accuracy:  0.948, Validation Accuracy:  0.958, Loss:  0.069
Epoch   5 Batch   79/134 - Train Accuracy:  0.949, Validation Accuracy:  0.954, Loss:  0.073
Epoch   5 Batch   80/134 - Train Accuracy:  0.954, Validation Accuracy:  0.955, Loss:  0.069
Epoch   5 Batch   81/134 - Train Accuracy:  0.958, Validation Accuracy:  0.955, Loss:  0.074
Epoch   5 Batch   82/134 - Train Accuracy:  0.952, Validation Accuracy:  0.957, Loss:  0.067
Epoch   5 Batch   83/134 - Train Accuracy:  0.949, Validation Accuracy:  0.956, Loss:  0.072
Epoch   5 Batch   84/134 - Train Accuracy:  0.945, Validation Accuracy:  0.953, Loss:  0.070
Epoch   5 Batch   85/134 - Train Accuracy:  0.950, Validation Accuracy:  0.949, Loss:  0.075
Epoch   5 Batch   86/134 - Train Accuracy:  0.956, Validation Accuracy:  0.951, Loss:  0.069
Epoch   5 Batch   87/134 - Train Accuracy:  0.934, Validation Accuracy:  0.956, Loss:  0.083
Epoch   5 Batch   88/134 - Train Accuracy:  0.952, Validation Accuracy:  0.952, Loss:  0.069
Epoch   5 Batch   89/134 - Train Accuracy:  0.947, Validation Accuracy:  0.956, Loss:  0.069
Epoch   5 Batch   90/134 - Train Accuracy:  0.949, Validation Accuracy:  0.956, Loss:  0.075
Epoch   5 Batch   91/134 - Train Accuracy:  0.958, Validation Accuracy:  0.957, Loss:  0.056
Epoch   5 Batch   92/134 - Train Accuracy:  0.959, Validation Accuracy:  0.959, Loss:  0.066
Epoch   5 Batch   93/134 - Train Accuracy:  0.951, Validation Accuracy:  0.958, Loss:  0.068
Epoch   5 Batch   94/134 - Train Accuracy:  0.954, Validation Accuracy:  0.958, Loss:  0.067
Epoch   5 Batch   95/134 - Train Accuracy:  0.949, Validation Accuracy:  0.958, Loss:  0.073
Epoch   5 Batch   96/134 - Train Accuracy:  0.940, Validation Accuracy:  0.954, Loss:  0.070
Epoch   5 Batch   97/134 - Train Accuracy:  0.945, Validation Accuracy:  0.954, Loss:  0.066
Epoch   5 Batch   98/134 - Train Accuracy:  0.949, Validation Accuracy:  0.959, Loss:  0.073
Epoch   5 Batch   99/134 - Train Accuracy:  0.958, Validation Accuracy:  0.960, Loss:  0.069
Epoch   5 Batch  100/134 - Train Accuracy:  0.953, Validation Accuracy:  0.957, Loss:  0.071
Epoch   5 Batch  101/134 - Train Accuracy:  0.950, Validation Accuracy:  0.957, Loss:  0.074
Epoch   5 Batch  102/134 - Train Accuracy:  0.946, Validation Accuracy:  0.959, Loss:  0.069
Epoch   5 Batch  103/134 - Train Accuracy:  0.948, Validation Accuracy:  0.956, Loss:  0.071
Epoch   5 Batch  104/134 - Train Accuracy:  0.951, Validation Accuracy:  0.957, Loss:  0.066
Epoch   5 Batch  105/134 - Train Accuracy:  0.955, Validation Accuracy:  0.958, Loss:  0.078
Epoch   5 Batch  106/134 - Train Accuracy:  0.945, Validation Accuracy:  0.959, Loss:  0.068
Epoch   5 Batch  107/134 - Train Accuracy:  0.953, Validation Accuracy:  0.959, Loss:  0.072
Epoch   5 Batch  108/134 - Train Accuracy:  0.945, Validation Accuracy:  0.955, Loss:  0.071
Epoch   5 Batch  109/134 - Train Accuracy:  0.947, Validation Accuracy:  0.956, Loss:  0.064
Epoch   5 Batch  110/134 - Train Accuracy:  0.956, Validation Accuracy:  0.957, Loss:  0.066
Epoch   5 Batch  111/134 - Train Accuracy:  0.949, Validation Accuracy:  0.954, Loss:  0.076
Epoch   5 Batch  112/134 - Train Accuracy:  0.949, Validation Accuracy:  0.953, Loss:  0.067
Epoch   5 Batch  113/134 - Train Accuracy:  0.957, Validation Accuracy:  0.957, Loss:  0.070
Epoch   5 Batch  114/134 - Train Accuracy:  0.951, Validation Accuracy:  0.955, Loss:  0.069
Epoch   5 Batch  115/134 - Train Accuracy:  0.945, Validation Accuracy:  0.953, Loss:  0.067
Epoch   5 Batch  116/134 - Train Accuracy:  0.956, Validation Accuracy:  0.954, Loss:  0.070
Epoch   5 Batch  117/134 - Train Accuracy:  0.964, Validation Accuracy:  0.961, Loss:  0.059
Epoch   5 Batch  118/134 - Train Accuracy:  0.955, Validation Accuracy:  0.957, Loss:  0.065
Epoch   5 Batch  119/134 - Train Accuracy:  0.959, Validation Accuracy:  0.954, Loss:  0.061
Epoch   5 Batch  120/134 - Train Accuracy:  0.942, Validation Accuracy:  0.951, Loss:  0.072
Epoch   5 Batch  121/134 - Train Accuracy:  0.952, Validation Accuracy:  0.955, Loss:  0.065
Epoch   5 Batch  122/134 - Train Accuracy:  0.951, Validation Accuracy:  0.955, Loss:  0.069
Epoch   5 Batch  123/134 - Train Accuracy:  0.962, Validation Accuracy:  0.956, Loss:  0.063
Epoch   5 Batch  124/134 - Train Accuracy:  0.951, Validation Accuracy:  0.956, Loss:  0.062
Epoch   5 Batch  125/134 - Train Accuracy:  0.961, Validation Accuracy:  0.957, Loss:  0.060
Epoch   5 Batch  126/134 - Train Accuracy:  0.952, Validation Accuracy:  0.958, Loss:  0.071
Epoch   5 Batch  127/134 - Train Accuracy:  0.953, Validation Accuracy:  0.957, Loss:  0.068
Epoch   5 Batch  128/134 - Train Accuracy:  0.951, Validation Accuracy:  0.958, Loss:  0.069
Epoch   5 Batch  129/134 - Train Accuracy:  0.951, Validation Accuracy:  0.960, Loss:  0.070
Epoch   5 Batch  130/134 - Train Accuracy:  0.955, Validation Accuracy:  0.960, Loss:  0.064
Epoch   5 Batch  131/134 - Train Accuracy:  0.950, Validation Accuracy:  0.961, Loss:  0.071
Epoch   5 Batch  132/134 - Train Accuracy:  0.950, Validation Accuracy:  0.962, Loss:  0.061
Epoch   6 Batch    0/134 - Train Accuracy:  0.954, Validation Accuracy:  0.959, Loss:  0.065
Epoch   6 Batch    1/134 - Train Accuracy:  0.945, Validation Accuracy:  0.955, Loss:  0.068
Epoch   6 Batch    2/134 - Train Accuracy:  0.958, Validation Accuracy:  0.960, Loss:  0.061
Epoch   6 Batch    3/134 - Train Accuracy:  0.957, Validation Accuracy:  0.961, Loss:  0.066
Epoch   6 Batch    4/134 - Train Accuracy:  0.955, Validation Accuracy:  0.962, Loss:  0.064
Epoch   6 Batch    5/134 - Train Accuracy:  0.953, Validation Accuracy:  0.959, Loss:  0.069
Epoch   6 Batch    6/134 - Train Accuracy:  0.958, Validation Accuracy:  0.962, Loss:  0.061
Epoch   6 Batch    7/134 - Train Accuracy:  0.956, Validation Accuracy:  0.964, Loss:  0.063
Epoch   6 Batch    8/134 - Train Accuracy:  0.960, Validation Accuracy:  0.961, Loss:  0.064
Epoch   6 Batch    9/134 - Train Accuracy:  0.956, Validation Accuracy:  0.959, Loss:  0.057
Epoch   6 Batch   10/134 - Train Accuracy:  0.951, Validation Accuracy:  0.963, Loss:  0.063
Epoch   6 Batch   11/134 - Train Accuracy:  0.951, Validation Accuracy:  0.962, Loss:  0.065
Epoch   6 Batch   12/134 - Train Accuracy:  0.952, Validation Accuracy:  0.962, Loss:  0.063
Epoch   6 Batch   13/134 - Train Accuracy:  0.939, Validation Accuracy:  0.956, Loss:  0.065
Epoch   6 Batch   14/134 - Train Accuracy:  0.949, Validation Accuracy:  0.958, Loss:  0.066
Epoch   6 Batch   15/134 - Train Accuracy:  0.951, Validation Accuracy:  0.963, Loss:  0.064
Epoch   6 Batch   16/134 - Train Accuracy:  0.948, Validation Accuracy:  0.956, Loss:  0.061
Epoch   6 Batch   17/134 - Train Accuracy:  0.952, Validation Accuracy:  0.958, Loss:  0.071
Epoch   6 Batch   18/134 - Train Accuracy:  0.950, Validation Accuracy:  0.962, Loss:  0.065
Epoch   6 Batch   19/134 - Train Accuracy:  0.945, Validation Accuracy:  0.957, Loss:  0.063
Epoch   6 Batch   20/134 - Train Accuracy:  0.951, Validation Accuracy:  0.954, Loss:  0.061
Epoch   6 Batch   21/134 - Train Accuracy:  0.946, Validation Accuracy:  0.955, Loss:  0.069
Epoch   6 Batch   22/134 - Train Accuracy:  0.952, Validation Accuracy:  0.958, Loss:  0.064
Epoch   6 Batch   23/134 - Train Accuracy:  0.955, Validation Accuracy:  0.959, Loss:  0.058
Epoch   6 Batch   24/134 - Train Accuracy:  0.946, Validation Accuracy:  0.959, Loss:  0.068
Epoch   6 Batch   25/134 - Train Accuracy:  0.954, Validation Accuracy:  0.960, Loss:  0.058
Epoch   6 Batch   26/134 - Train Accuracy:  0.954, Validation Accuracy:  0.957, Loss:  0.063
Epoch   6 Batch   27/134 - Train Accuracy:  0.945, Validation Accuracy:  0.956, Loss:  0.063
Epoch   6 Batch   28/134 - Train Accuracy:  0.957, Validation Accuracy:  0.954, Loss:  0.069
Epoch   6 Batch   29/134 - Train Accuracy:  0.959, Validation Accuracy:  0.955, Loss:  0.052
Epoch   6 Batch   30/134 - Train Accuracy:  0.949, Validation Accuracy:  0.956, Loss:  0.064
Epoch   6 Batch   31/134 - Train Accuracy:  0.952, Validation Accuracy:  0.957, Loss:  0.064
Epoch   6 Batch   32/134 - Train Accuracy:  0.952, Validation Accuracy:  0.958, Loss:  0.058
Epoch   6 Batch   33/134 - Train Accuracy:  0.950, Validation Accuracy:  0.960, Loss:  0.070
Epoch   6 Batch   34/134 - Train Accuracy:  0.945, Validation Accuracy:  0.960, Loss:  0.069
Epoch   6 Batch   35/134 - Train Accuracy:  0.954, Validation Accuracy:  0.961, Loss:  0.064
Epoch   6 Batch   36/134 - Train Accuracy:  0.955, Validation Accuracy:  0.962, Loss:  0.062
Epoch   6 Batch   37/134 - Train Accuracy:  0.950, Validation Accuracy:  0.959, Loss:  0.066
Epoch   6 Batch   38/134 - Train Accuracy:  0.958, Validation Accuracy:  0.958, Loss:  0.062
Epoch   6 Batch   39/134 - Train Accuracy:  0.952, Validation Accuracy:  0.956, Loss:  0.063
Epoch   6 Batch   40/134 - Train Accuracy:  0.957, Validation Accuracy:  0.958, Loss:  0.063
Epoch   6 Batch   41/134 - Train Accuracy:  0.947, Validation Accuracy:  0.962, Loss:  0.066
Epoch   6 Batch   42/134 - Train Accuracy:  0.951, Validation Accuracy:  0.959, Loss:  0.057
Epoch   6 Batch   43/134 - Train Accuracy:  0.955, Validation Accuracy:  0.957, Loss:  0.064
Epoch   6 Batch   44/134 - Train Accuracy:  0.951, Validation Accuracy:  0.955, Loss:  0.061
Epoch   6 Batch   45/134 - Train Accuracy:  0.959, Validation Accuracy:  0.960, Loss:  0.057
Epoch   6 Batch   46/134 - Train Accuracy:  0.958, Validation Accuracy:  0.961, Loss:  0.064
Epoch   6 Batch   47/134 - Train Accuracy:  0.955, Validation Accuracy:  0.963, Loss:  0.064
Epoch   6 Batch   48/134 - Train Accuracy:  0.956, Validation Accuracy:  0.959, Loss:  0.064
Epoch   6 Batch   49/134 - Train Accuracy:  0.964, Validation Accuracy:  0.962, Loss:  0.057
Epoch   6 Batch   50/134 - Train Accuracy:  0.953, Validation Accuracy:  0.955, Loss:  0.063
Epoch   6 Batch   51/134 - Train Accuracy:  0.956, Validation Accuracy:  0.960, Loss:  0.063
Epoch   6 Batch   52/134 - Train Accuracy:  0.958, Validation Accuracy:  0.962, Loss:  0.059
Epoch   6 Batch   53/134 - Train Accuracy:  0.959, Validation Accuracy:  0.960, Loss:  0.055
Epoch   6 Batch   54/134 - Train Accuracy:  0.948, Validation Accuracy:  0.961, Loss:  0.063
Epoch   6 Batch   55/134 - Train Accuracy:  0.953, Validation Accuracy:  0.962, Loss:  0.061
Epoch   6 Batch   56/134 - Train Accuracy:  0.954, Validation Accuracy:  0.962, Loss:  0.058
Epoch   6 Batch   57/134 - Train Accuracy:  0.960, Validation Accuracy:  0.963, Loss:  0.058
Epoch   6 Batch   58/134 - Train Accuracy:  0.955, Validation Accuracy:  0.961, Loss:  0.055
Epoch   6 Batch   59/134 - Train Accuracy:  0.951, Validation Accuracy:  0.963, Loss:  0.061
Epoch   6 Batch   60/134 - Train Accuracy:  0.955, Validation Accuracy:  0.962, Loss:  0.058
Epoch   6 Batch   61/134 - Train Accuracy:  0.955, Validation Accuracy:  0.962, Loss:  0.055
Epoch   6 Batch   62/134 - Train Accuracy:  0.957, Validation Accuracy:  0.961, Loss:  0.053
Epoch   6 Batch   63/134 - Train Accuracy:  0.957, Validation Accuracy:  0.959, Loss:  0.056
Epoch   6 Batch   64/134 - Train Accuracy:  0.957, Validation Accuracy:  0.956, Loss:  0.059
Epoch   6 Batch   65/134 - Train Accuracy:  0.945, Validation Accuracy:  0.960, Loss:  0.063
Epoch   6 Batch   66/134 - Train Accuracy:  0.954, Validation Accuracy:  0.960, Loss:  0.054
Epoch   6 Batch   67/134 - Train Accuracy:  0.952, Validation Accuracy:  0.963, Loss:  0.061
Epoch   6 Batch   68/134 - Train Accuracy:  0.948, Validation Accuracy:  0.963, Loss:  0.058
Epoch   6 Batch   69/134 - Train Accuracy:  0.955, Validation Accuracy:  0.960, Loss:  0.058
Epoch   6 Batch   70/134 - Train Accuracy:  0.960, Validation Accuracy:  0.961, Loss:  0.058
Epoch   6 Batch   71/134 - Train Accuracy:  0.961, Validation Accuracy:  0.962, Loss:  0.053
Epoch   6 Batch   72/134 - Train Accuracy:  0.956, Validation Accuracy:  0.962, Loss:  0.056
Epoch   6 Batch   73/134 - Train Accuracy:  0.954, Validation Accuracy:  0.958, Loss:  0.062
Epoch   6 Batch   74/134 - Train Accuracy:  0.955, Validation Accuracy:  0.964, Loss:  0.065
Epoch   6 Batch   75/134 - Train Accuracy:  0.956, Validation Accuracy:  0.959, Loss:  0.060
Epoch   6 Batch   76/134 - Train Accuracy:  0.966, Validation Accuracy:  0.964, Loss:  0.056
Epoch   6 Batch   77/134 - Train Accuracy:  0.958, Validation Accuracy:  0.964, Loss:  0.056
Epoch   6 Batch   78/134 - Train Accuracy:  0.954, Validation Accuracy:  0.966, Loss:  0.059
Epoch   6 Batch   79/134 - Train Accuracy:  0.955, Validation Accuracy:  0.964, Loss:  0.058
Epoch   6 Batch   80/134 - Train Accuracy:  0.961, Validation Accuracy:  0.962, Loss:  0.055
Epoch   6 Batch   81/134 - Train Accuracy:  0.966, Validation Accuracy:  0.959, Loss:  0.053
Epoch   6 Batch   82/134 - Train Accuracy:  0.957, Validation Accuracy:  0.962, Loss:  0.053
Epoch   6 Batch   83/134 - Train Accuracy:  0.957, Validation Accuracy:  0.962, Loss:  0.061
Epoch   6 Batch   84/134 - Train Accuracy:  0.956, Validation Accuracy:  0.962, Loss:  0.054
Epoch   6 Batch   85/134 - Train Accuracy:  0.963, Validation Accuracy:  0.961, Loss:  0.060
Epoch   6 Batch   86/134 - Train Accuracy:  0.958, Validation Accuracy:  0.965, Loss:  0.056
Epoch   6 Batch   87/134 - Train Accuracy:  0.950, Validation Accuracy:  0.964, Loss:  0.067
Epoch   6 Batch   88/134 - Train Accuracy:  0.962, Validation Accuracy:  0.962, Loss:  0.055
Epoch   6 Batch   89/134 - Train Accuracy:  0.953, Validation Accuracy:  0.958, Loss:  0.055
Epoch   6 Batch   90/134 - Train Accuracy:  0.955, Validation Accuracy:  0.957, Loss:  0.061
Epoch   6 Batch   91/134 - Train Accuracy:  0.960, Validation Accuracy:  0.961, Loss:  0.045
Epoch   6 Batch   92/134 - Train Accuracy:  0.959, Validation Accuracy:  0.959, Loss:  0.053
Epoch   6 Batch   93/134 - Train Accuracy:  0.953, Validation Accuracy:  0.956, Loss:  0.057
Epoch   6 Batch   94/134 - Train Accuracy:  0.954, Validation Accuracy:  0.956, Loss:  0.056
Epoch   6 Batch   95/134 - Train Accuracy:  0.956, Validation Accuracy:  0.957, Loss:  0.061
Epoch   6 Batch   96/134 - Train Accuracy:  0.950, Validation Accuracy:  0.960, Loss:  0.054
Epoch   6 Batch   97/134 - Train Accuracy:  0.950, Validation Accuracy:  0.962, Loss:  0.058
Epoch   6 Batch   98/134 - Train Accuracy:  0.957, Validation Accuracy:  0.961, Loss:  0.059
Epoch   6 Batch   99/134 - Train Accuracy:  0.965, Validation Accuracy:  0.957, Loss:  0.051
Epoch   6 Batch  100/134 - Train Accuracy:  0.958, Validation Accuracy:  0.959, Loss:  0.061
Epoch   6 Batch  101/134 - Train Accuracy:  0.955, Validation Accuracy:  0.960, Loss:  0.060
Epoch   6 Batch  102/134 - Train Accuracy:  0.952, Validation Accuracy:  0.960, Loss:  0.055
Epoch   6 Batch  103/134 - Train Accuracy:  0.957, Validation Accuracy:  0.960, Loss:  0.056
Epoch   6 Batch  104/134 - Train Accuracy:  0.958, Validation Accuracy:  0.962, Loss:  0.054
Epoch   6 Batch  105/134 - Train Accuracy:  0.954, Validation Accuracy:  0.963, Loss:  0.063
Epoch   6 Batch  106/134 - Train Accuracy:  0.956, Validation Accuracy:  0.966, Loss:  0.056
Epoch   6 Batch  107/134 - Train Accuracy:  0.955, Validation Accuracy:  0.962, Loss:  0.057
Epoch   6 Batch  108/134 - Train Accuracy:  0.950, Validation Accuracy:  0.958, Loss:  0.058
Epoch   6 Batch  109/134 - Train Accuracy:  0.955, Validation Accuracy:  0.960, Loss:  0.055
Epoch   6 Batch  110/134 - Train Accuracy:  0.957, Validation Accuracy:  0.962, Loss:  0.056
Epoch   6 Batch  111/134 - Train Accuracy:  0.955, Validation Accuracy:  0.964, Loss:  0.056
Epoch   6 Batch  112/134 - Train Accuracy:  0.954, Validation Accuracy:  0.963, Loss:  0.053
Epoch   6 Batch  113/134 - Train Accuracy:  0.960, Validation Accuracy:  0.963, Loss:  0.057
Epoch   6 Batch  114/134 - Train Accuracy:  0.959, Validation Accuracy:  0.965, Loss:  0.055
Epoch   6 Batch  115/134 - Train Accuracy:  0.953, Validation Accuracy:  0.968, Loss:  0.052
Epoch   6 Batch  116/134 - Train Accuracy:  0.960, Validation Accuracy:  0.967, Loss:  0.056
Epoch   6 Batch  117/134 - Train Accuracy:  0.969, Validation Accuracy:  0.965, Loss:  0.049
Epoch   6 Batch  118/134 - Train Accuracy:  0.962, Validation Accuracy:  0.963, Loss:  0.054
Epoch   6 Batch  119/134 - Train Accuracy:  0.964, Validation Accuracy:  0.968, Loss:  0.049
Epoch   6 Batch  120/134 - Train Accuracy:  0.951, Validation Accuracy:  0.969, Loss:  0.059
Epoch   6 Batch  121/134 - Train Accuracy:  0.961, Validation Accuracy:  0.967, Loss:  0.051
Epoch   6 Batch  122/134 - Train Accuracy:  0.963, Validation Accuracy:  0.964, Loss:  0.054
Epoch   6 Batch  123/134 - Train Accuracy:  0.965, Validation Accuracy:  0.963, Loss:  0.051
Epoch   6 Batch  124/134 - Train Accuracy:  0.959, Validation Accuracy:  0.964, Loss:  0.053
Epoch   6 Batch  125/134 - Train Accuracy:  0.968, Validation Accuracy:  0.965, Loss:  0.044
Epoch   6 Batch  126/134 - Train Accuracy:  0.951, Validation Accuracy:  0.963, Loss:  0.054
Epoch   6 Batch  127/134 - Train Accuracy:  0.954, Validation Accuracy:  0.963, Loss:  0.055
Epoch   6 Batch  128/134 - Train Accuracy:  0.958, Validation Accuracy:  0.962, Loss:  0.054
Epoch   6 Batch  129/134 - Train Accuracy:  0.958, Validation Accuracy:  0.966, Loss:  0.052
Epoch   6 Batch  130/134 - Train Accuracy:  0.962, Validation Accuracy:  0.965, Loss:  0.054
Epoch   6 Batch  131/134 - Train Accuracy:  0.955, Validation Accuracy:  0.970, Loss:  0.058
Epoch   6 Batch  132/134 - Train Accuracy:  0.957, Validation Accuracy:  0.969, Loss:  0.048
Epoch   7 Batch    0/134 - Train Accuracy:  0.963, Validation Accuracy:  0.968, Loss:  0.050
Epoch   7 Batch    1/134 - Train Accuracy:  0.955, Validation Accuracy:  0.967, Loss:  0.054
Epoch   7 Batch    2/134 - Train Accuracy:  0.963, Validation Accuracy:  0.966, Loss:  0.047
Epoch   7 Batch    3/134 - Train Accuracy:  0.965, Validation Accuracy:  0.967, Loss:  0.053
Epoch   7 Batch    4/134 - Train Accuracy:  0.961, Validation Accuracy:  0.970, Loss:  0.054
Epoch   7 Batch    5/134 - Train Accuracy:  0.954, Validation Accuracy:  0.970, Loss:  0.057
Epoch   7 Batch    6/134 - Train Accuracy:  0.964, Validation Accuracy:  0.969, Loss:  0.046
Epoch   7 Batch    7/134 - Train Accuracy:  0.957, Validation Accuracy:  0.968, Loss:  0.050
Epoch   7 Batch    8/134 - Train Accuracy:  0.956, Validation Accuracy:  0.966, Loss:  0.050
Epoch   7 Batch    9/134 - Train Accuracy:  0.960, Validation Accuracy:  0.967, Loss:  0.044
Epoch   7 Batch   10/134 - Train Accuracy:  0.953, Validation Accuracy:  0.967, Loss:  0.050
Epoch   7 Batch   11/134 - Train Accuracy:  0.954, Validation Accuracy:  0.966, Loss:  0.050
Epoch   7 Batch   12/134 - Train Accuracy:  0.956, Validation Accuracy:  0.966, Loss:  0.049
Epoch   7 Batch   13/134 - Train Accuracy:  0.953, Validation Accuracy:  0.968, Loss:  0.051
Epoch   7 Batch   14/134 - Train Accuracy:  0.956, Validation Accuracy:  0.967, Loss:  0.050
Epoch   7 Batch   15/134 - Train Accuracy:  0.964, Validation Accuracy:  0.965, Loss:  0.047
Epoch   7 Batch   16/134 - Train Accuracy:  0.959, Validation Accuracy:  0.966, Loss:  0.047
Epoch   7 Batch   17/134 - Train Accuracy:  0.954, Validation Accuracy:  0.964, Loss:  0.055
Epoch   7 Batch   18/134 - Train Accuracy:  0.955, Validation Accuracy:  0.966, Loss:  0.051
Epoch   7 Batch   19/134 - Train Accuracy:  0.956, Validation Accuracy:  0.967, Loss:  0.053
Epoch   7 Batch   20/134 - Train Accuracy:  0.962, Validation Accuracy:  0.965, Loss:  0.048
Epoch   7 Batch   21/134 - Train Accuracy:  0.961, Validation Accuracy:  0.967, Loss:  0.053
Epoch   7 Batch   22/134 - Train Accuracy:  0.958, Validation Accuracy:  0.966, Loss:  0.049
Epoch   7 Batch   23/134 - Train Accuracy:  0.960, Validation Accuracy:  0.965, Loss:  0.045
Epoch   7 Batch   24/134 - Train Accuracy:  0.955, Validation Accuracy:  0.963, Loss:  0.053
Epoch   7 Batch   25/134 - Train Accuracy:  0.962, Validation Accuracy:  0.962, Loss:  0.043
Epoch   7 Batch   26/134 - Train Accuracy:  0.960, Validation Accuracy:  0.961, Loss:  0.048
Epoch   7 Batch   27/134 - Train Accuracy:  0.960, Validation Accuracy:  0.962, Loss:  0.048
Epoch   7 Batch   28/134 - Train Accuracy:  0.964, Validation Accuracy:  0.962, Loss:  0.055
Epoch   7 Batch   29/134 - Train Accuracy:  0.968, Validation Accuracy:  0.966, Loss:  0.045
Epoch   7 Batch   30/134 - Train Accuracy:  0.954, Validation Accuracy:  0.966, Loss:  0.051
Epoch   7 Batch   31/134 - Train Accuracy:  0.951, Validation Accuracy:  0.966, Loss:  0.051
Epoch   7 Batch   32/134 - Train Accuracy:  0.961, Validation Accuracy:  0.966, Loss:  0.048
Epoch   7 Batch   33/134 - Train Accuracy:  0.956, Validation Accuracy:  0.965, Loss:  0.054
Epoch   7 Batch   34/134 - Train Accuracy:  0.961, Validation Accuracy:  0.965, Loss:  0.057
Epoch   7 Batch   35/134 - Train Accuracy:  0.955, Validation Accuracy:  0.966, Loss:  0.057
Epoch   7 Batch   36/134 - Train Accuracy:  0.958, Validation Accuracy:  0.963, Loss:  0.051
Epoch   7 Batch   37/134 - Train Accuracy:  0.954, Validation Accuracy:  0.963, Loss:  0.051
Epoch   7 Batch   38/134 - Train Accuracy:  0.969, Validation Accuracy:  0.964, Loss:  0.051
Epoch   7 Batch   39/134 - Train Accuracy:  0.957, Validation Accuracy:  0.965, Loss:  0.051
Epoch   7 Batch   40/134 - Train Accuracy:  0.962, Validation Accuracy:  0.968, Loss:  0.047
Epoch   7 Batch   41/134 - Train Accuracy:  0.959, Validation Accuracy:  0.968, Loss:  0.054
Epoch   7 Batch   42/134 - Train Accuracy:  0.957, Validation Accuracy:  0.964, Loss:  0.045
Epoch   7 Batch   43/134 - Train Accuracy:  0.958, Validation Accuracy:  0.963, Loss:  0.051
Epoch   7 Batch   44/134 - Train Accuracy:  0.958, Validation Accuracy:  0.966, Loss:  0.052
Epoch   7 Batch   45/134 - Train Accuracy:  0.965, Validation Accuracy:  0.966, Loss:  0.044
Epoch   7 Batch   46/134 - Train Accuracy:  0.964, Validation Accuracy:  0.964, Loss:  0.051
Epoch   7 Batch   47/134 - Train Accuracy:  0.958, Validation Accuracy:  0.965, Loss:  0.048
Epoch   7 Batch   48/134 - Train Accuracy:  0.960, Validation Accuracy:  0.967, Loss:  0.050
Epoch   7 Batch   49/134 - Train Accuracy:  0.962, Validation Accuracy:  0.966, Loss:  0.047
Epoch   7 Batch   50/134 - Train Accuracy:  0.958, Validation Accuracy:  0.964, Loss:  0.050
Epoch   7 Batch   51/134 - Train Accuracy:  0.963, Validation Accuracy:  0.967, Loss:  0.053
Epoch   7 Batch   52/134 - Train Accuracy:  0.960, Validation Accuracy:  0.967, Loss:  0.048
Epoch   7 Batch   53/134 - Train Accuracy:  0.968, Validation Accuracy:  0.969, Loss:  0.045
Epoch   7 Batch   54/134 - Train Accuracy:  0.960, Validation Accuracy:  0.969, Loss:  0.048
Epoch   7 Batch   55/134 - Train Accuracy:  0.958, Validation Accuracy:  0.967, Loss:  0.049
Epoch   7 Batch   56/134 - Train Accuracy:  0.961, Validation Accuracy:  0.966, Loss:  0.046
Epoch   7 Batch   57/134 - Train Accuracy:  0.961, Validation Accuracy:  0.961, Loss:  0.050
Epoch   7 Batch   58/134 - Train Accuracy:  0.965, Validation Accuracy:  0.962, Loss:  0.040
Epoch   7 Batch   59/134 - Train Accuracy:  0.956, Validation Accuracy:  0.961, Loss:  0.051
Epoch   7 Batch   60/134 - Train Accuracy:  0.956, Validation Accuracy:  0.960, Loss:  0.047
Epoch   7 Batch   61/134 - Train Accuracy:  0.958, Validation Accuracy:  0.963, Loss:  0.045
Epoch   7 Batch   62/134 - Train Accuracy:  0.963, Validation Accuracy:  0.964, Loss:  0.044
Epoch   7 Batch   63/134 - Train Accuracy:  0.962, Validation Accuracy:  0.962, Loss:  0.047
Epoch   7 Batch   64/134 - Train Accuracy:  0.957, Validation Accuracy:  0.962, Loss:  0.045
Epoch   7 Batch   65/134 - Train Accuracy:  0.953, Validation Accuracy:  0.966, Loss:  0.050
Epoch   7 Batch   66/134 - Train Accuracy:  0.960, Validation Accuracy:  0.962, Loss:  0.048
Epoch   7 Batch   67/134 - Train Accuracy:  0.954, Validation Accuracy:  0.963, Loss:  0.050
Epoch   7 Batch   68/134 - Train Accuracy:  0.955, Validation Accuracy:  0.965, Loss:  0.051
Epoch   7 Batch   69/134 - Train Accuracy:  0.965, Validation Accuracy:  0.965, Loss:  0.049
Epoch   7 Batch   70/134 - Train Accuracy:  0.963, Validation Accuracy:  0.969, Loss:  0.047
Epoch   7 Batch   71/134 - Train Accuracy:  0.966, Validation Accuracy:  0.967, Loss:  0.041
Epoch   7 Batch   72/134 - Train Accuracy:  0.959, Validation Accuracy:  0.969, Loss:  0.043
Epoch   7 Batch   73/134 - Train Accuracy:  0.963, Validation Accuracy:  0.964, Loss:  0.048
Epoch   7 Batch   74/134 - Train Accuracy:  0.955, Validation Accuracy:  0.965, Loss:  0.053
Epoch   7 Batch   75/134 - Train Accuracy:  0.967, Validation Accuracy:  0.967, Loss:  0.047
Epoch   7 Batch   76/134 - Train Accuracy:  0.968, Validation Accuracy:  0.968, Loss:  0.047
Epoch   7 Batch   77/134 - Train Accuracy:  0.965, Validation Accuracy:  0.966, Loss:  0.042
Epoch   7 Batch   78/134 - Train Accuracy:  0.958, Validation Accuracy:  0.968, Loss:  0.043
Epoch   7 Batch   79/134 - Train Accuracy:  0.964, Validation Accuracy:  0.968, Loss:  0.044
Epoch   7 Batch   80/134 - Train Accuracy:  0.968, Validation Accuracy:  0.965, Loss:  0.042
Epoch   7 Batch   81/134 - Train Accuracy:  0.969, Validation Accuracy:  0.964, Loss:  0.044
Epoch   7 Batch   82/134 - Train Accuracy:  0.966, Validation Accuracy:  0.962, Loss:  0.041
Epoch   7 Batch   83/134 - Train Accuracy:  0.963, Validation Accuracy:  0.962, Loss:  0.049
Epoch   7 Batch   84/134 - Train Accuracy:  0.956, Validation Accuracy:  0.964, Loss:  0.044
Epoch   7 Batch   85/134 - Train Accuracy:  0.962, Validation Accuracy:  0.965, Loss:  0.050
Epoch   7 Batch   86/134 - Train Accuracy:  0.962, Validation Accuracy:  0.967, Loss:  0.043
Epoch   7 Batch   87/134 - Train Accuracy:  0.962, Validation Accuracy:  0.965, Loss:  0.052
Epoch   7 Batch   88/134 - Train Accuracy:  0.971, Validation Accuracy:  0.965, Loss:  0.042
Epoch   7 Batch   89/134 - Train Accuracy:  0.957, Validation Accuracy:  0.966, Loss:  0.044
Epoch   7 Batch   90/134 - Train Accuracy:  0.959, Validation Accuracy:  0.970, Loss:  0.047
Epoch   7 Batch   91/134 - Train Accuracy:  0.968, Validation Accuracy:  0.970, Loss:  0.037
Epoch   7 Batch   92/134 - Train Accuracy:  0.969, Validation Accuracy:  0.969, Loss:  0.044
Epoch   7 Batch   93/134 - Train Accuracy:  0.969, Validation Accuracy:  0.967, Loss:  0.042
Epoch   7 Batch   94/134 - Train Accuracy:  0.962, Validation Accuracy:  0.964, Loss:  0.043
Epoch   7 Batch   95/134 - Train Accuracy:  0.963, Validation Accuracy:  0.966, Loss:  0.046
Epoch   7 Batch   96/134 - Train Accuracy:  0.956, Validation Accuracy:  0.963, Loss:  0.044
Epoch   7 Batch   97/134 - Train Accuracy:  0.958, Validation Accuracy:  0.966, Loss:  0.044
Epoch   7 Batch   98/134 - Train Accuracy:  0.961, Validation Accuracy:  0.966, Loss:  0.046
Epoch   7 Batch   99/134 - Train Accuracy:  0.967, Validation Accuracy:  0.968, Loss:  0.042
Epoch   7 Batch  100/134 - Train Accuracy:  0.964, Validation Accuracy:  0.972, Loss:  0.042
Epoch   7 Batch  101/134 - Train Accuracy:  0.958, Validation Accuracy:  0.970, Loss:  0.048
Epoch   7 Batch  102/134 - Train Accuracy:  0.964, Validation Accuracy:  0.970, Loss:  0.046
Epoch   7 Batch  103/134 - Train Accuracy:  0.959, Validation Accuracy:  0.969, Loss:  0.044
Epoch   7 Batch  104/134 - Train Accuracy:  0.965, Validation Accuracy:  0.965, Loss:  0.043
Epoch   7 Batch  105/134 - Train Accuracy:  0.962, Validation Accuracy:  0.966, Loss:  0.050
Epoch   7 Batch  106/134 - Train Accuracy:  0.963, Validation Accuracy:  0.971, Loss:  0.045
Epoch   7 Batch  107/134 - Train Accuracy:  0.961, Validation Accuracy:  0.968, Loss:  0.049
Epoch   7 Batch  108/134 - Train Accuracy:  0.962, Validation Accuracy:  0.967, Loss:  0.048
Epoch   7 Batch  109/134 - Train Accuracy:  0.963, Validation Accuracy:  0.966, Loss:  0.043
Epoch   7 Batch  110/134 - Train Accuracy:  0.967, Validation Accuracy:  0.968, Loss:  0.040
Epoch   7 Batch  111/134 - Train Accuracy:  0.959, Validation Accuracy:  0.966, Loss:  0.048
Epoch   7 Batch  112/134 - Train Accuracy:  0.960, Validation Accuracy:  0.966, Loss:  0.045
Epoch   7 Batch  113/134 - Train Accuracy:  0.963, Validation Accuracy:  0.966, Loss:  0.048
Epoch   7 Batch  114/134 - Train Accuracy:  0.967, Validation Accuracy:  0.968, Loss:  0.045
Epoch   7 Batch  115/134 - Train Accuracy:  0.959, Validation Accuracy:  0.969, Loss:  0.046
Epoch   7 Batch  116/134 - Train Accuracy:  0.967, Validation Accuracy:  0.969, Loss:  0.044
Epoch   7 Batch  117/134 - Train Accuracy:  0.970, Validation Accuracy:  0.969, Loss:  0.039
Epoch   7 Batch  118/134 - Train Accuracy:  0.971, Validation Accuracy:  0.968, Loss:  0.044
Epoch   7 Batch  119/134 - Train Accuracy:  0.967, Validation Accuracy:  0.965, Loss:  0.040
Epoch   7 Batch  120/134 - Train Accuracy:  0.960, Validation Accuracy:  0.965, Loss:  0.046
Epoch   7 Batch  121/134 - Train Accuracy:  0.966, Validation Accuracy:  0.967, Loss:  0.042
Epoch   7 Batch  122/134 - Train Accuracy:  0.960, Validation Accuracy:  0.967, Loss:  0.046
Epoch   7 Batch  123/134 - Train Accuracy:  0.969, Validation Accuracy:  0.966, Loss:  0.041
Epoch   7 Batch  124/134 - Train Accuracy:  0.966, Validation Accuracy:  0.969, Loss:  0.044
Epoch   7 Batch  125/134 - Train Accuracy:  0.971, Validation Accuracy:  0.966, Loss:  0.036
Epoch   7 Batch  126/134 - Train Accuracy:  0.957, Validation Accuracy:  0.965, Loss:  0.047
Epoch   7 Batch  127/134 - Train Accuracy:  0.961, Validation Accuracy:  0.968, Loss:  0.043
Epoch   7 Batch  128/134 - Train Accuracy:  0.962, Validation Accuracy:  0.961, Loss:  0.047
Epoch   7 Batch  129/134 - Train Accuracy:  0.970, Validation Accuracy:  0.966, Loss:  0.042
Epoch   7 Batch  130/134 - Train Accuracy:  0.962, Validation Accuracy:  0.968, Loss:  0.044
Epoch   7 Batch  131/134 - Train Accuracy:  0.963, Validation Accuracy:  0.968, Loss:  0.046
Epoch   7 Batch  132/134 - Train Accuracy:  0.962, Validation Accuracy:  0.966, Loss:  0.040
Epoch   8 Batch    0/134 - Train Accuracy:  0.966, Validation Accuracy:  0.969, Loss:  0.043
Epoch   8 Batch    1/134 - Train Accuracy:  0.959, Validation Accuracy:  0.969, Loss:  0.044
Epoch   8 Batch    2/134 - Train Accuracy:  0.972, Validation Accuracy:  0.970, Loss:  0.040
Epoch   8 Batch    3/134 - Train Accuracy:  0.968, Validation Accuracy:  0.970, Loss:  0.045
Epoch   8 Batch    4/134 - Train Accuracy:  0.968, Validation Accuracy:  0.966, Loss:  0.043
Epoch   8 Batch    5/134 - Train Accuracy:  0.961, Validation Accuracy:  0.966, Loss:  0.048
Epoch   8 Batch    6/134 - Train Accuracy:  0.967, Validation Accuracy:  0.967, Loss:  0.039
Epoch   8 Batch    7/134 - Train Accuracy:  0.964, Validation Accuracy:  0.964, Loss:  0.040
Epoch   8 Batch    8/134 - Train Accuracy:  0.962, Validation Accuracy:  0.961, Loss:  0.040
Epoch   8 Batch    9/134 - Train Accuracy:  0.965, Validation Accuracy:  0.961, Loss:  0.039
Epoch   8 Batch   10/134 - Train Accuracy:  0.959, Validation Accuracy:  0.965, Loss:  0.043
Epoch   8 Batch   11/134 - Train Accuracy:  0.965, Validation Accuracy:  0.964, Loss:  0.041
Epoch   8 Batch   12/134 - Train Accuracy:  0.960, Validation Accuracy:  0.964, Loss:  0.043
Epoch   8 Batch   13/134 - Train Accuracy:  0.955, Validation Accuracy:  0.964, Loss:  0.045
Epoch   8 Batch   14/134 - Train Accuracy:  0.956, Validation Accuracy:  0.962, Loss:  0.046
Epoch   8 Batch   15/134 - Train Accuracy:  0.963, Validation Accuracy:  0.964, Loss:  0.042
Epoch   8 Batch   16/134 - Train Accuracy:  0.964, Validation Accuracy:  0.968, Loss:  0.037
Epoch   8 Batch   17/134 - Train Accuracy:  0.959, Validation Accuracy:  0.967, Loss:  0.048
Epoch   8 Batch   18/134 - Train Accuracy:  0.965, Validation Accuracy:  0.968, Loss:  0.043
Epoch   8 Batch   19/134 - Train Accuracy:  0.961, Validation Accuracy:  0.963, Loss:  0.045
Epoch   8 Batch   20/134 - Train Accuracy:  0.970, Validation Accuracy:  0.965, Loss:  0.042
Epoch   8 Batch   21/134 - Train Accuracy:  0.965, Validation Accuracy:  0.970, Loss:  0.046
Epoch   8 Batch   22/134 - Train Accuracy:  0.965, Validation Accuracy:  0.970, Loss:  0.042
Epoch   8 Batch   23/134 - Train Accuracy:  0.964, Validation Accuracy:  0.970, Loss:  0.039
Epoch   8 Batch   24/134 - Train Accuracy:  0.961, Validation Accuracy:  0.967, Loss:  0.046
Epoch   8 Batch   25/134 - Train Accuracy:  0.966, Validation Accuracy:  0.967, Loss:  0.036
Epoch   8 Batch   26/134 - Train Accuracy:  0.967, Validation Accuracy:  0.965, Loss:  0.043
Epoch   8 Batch   27/134 - Train Accuracy:  0.962, Validation Accuracy:  0.963, Loss:  0.039
Epoch   8 Batch   28/134 - Train Accuracy:  0.963, Validation Accuracy:  0.965, Loss:  0.046
Epoch   8 Batch   29/134 - Train Accuracy:  0.972, Validation Accuracy:  0.964, Loss:  0.040
Epoch   8 Batch   30/134 - Train Accuracy:  0.956, Validation Accuracy:  0.962, Loss:  0.044
Epoch   8 Batch   31/134 - Train Accuracy:  0.960, Validation Accuracy:  0.964, Loss:  0.046
Epoch   8 Batch   32/134 - Train Accuracy:  0.960, Validation Accuracy:  0.966, Loss:  0.043
Epoch   8 Batch   33/134 - Train Accuracy:  0.962, Validation Accuracy:  0.966, Loss:  0.050
Epoch   8 Batch   34/134 - Train Accuracy:  0.962, Validation Accuracy:  0.961, Loss:  0.047
Epoch   8 Batch   35/134 - Train Accuracy:  0.961, Validation Accuracy:  0.966, Loss:  0.048
Epoch   8 Batch   36/134 - Train Accuracy:  0.965, Validation Accuracy:  0.968, Loss:  0.043
Epoch   8 Batch   37/134 - Train Accuracy:  0.963, Validation Accuracy:  0.968, Loss:  0.043
Epoch   8 Batch   38/134 - Train Accuracy:  0.970, Validation Accuracy:  0.967, Loss:  0.043
Epoch   8 Batch   39/134 - Train Accuracy:  0.964, Validation Accuracy:  0.968, Loss:  0.042
Epoch   8 Batch   40/134 - Train Accuracy:  0.963, Validation Accuracy:  0.966, Loss:  0.040
Epoch   8 Batch   41/134 - Train Accuracy:  0.957, Validation Accuracy:  0.966, Loss:  0.050
Epoch   8 Batch   42/134 - Train Accuracy:  0.963, Validation Accuracy:  0.964, Loss:  0.039
Epoch   8 Batch   43/134 - Train Accuracy:  0.960, Validation Accuracy:  0.965, Loss:  0.046
Epoch   8 Batch   44/134 - Train Accuracy:  0.964, Validation Accuracy:  0.965, Loss:  0.043
Epoch   8 Batch   45/134 - Train Accuracy:  0.974, Validation Accuracy:  0.964, Loss:  0.039
Epoch   8 Batch   46/134 - Train Accuracy:  0.959, Validation Accuracy:  0.963, Loss:  0.045
Epoch   8 Batch   47/134 - Train Accuracy:  0.964, Validation Accuracy:  0.965, Loss:  0.042
Epoch   8 Batch   48/134 - Train Accuracy:  0.968, Validation Accuracy:  0.967, Loss:  0.046
Epoch   8 Batch   49/134 - Train Accuracy:  0.968, Validation Accuracy:  0.967, Loss:  0.040
Epoch   8 Batch   50/134 - Train Accuracy:  0.966, Validation Accuracy:  0.968, Loss:  0.045
Epoch   8 Batch   51/134 - Train Accuracy:  0.967, Validation Accuracy:  0.968, Loss:  0.046
Epoch   8 Batch   52/134 - Train Accuracy:  0.971, Validation Accuracy:  0.970, Loss:  0.040
Epoch   8 Batch   53/134 - Train Accuracy:  0.966, Validation Accuracy:  0.968, Loss:  0.039
Epoch   8 Batch   54/134 - Train Accuracy:  0.959, Validation Accuracy:  0.970, Loss:  0.043
Epoch   8 Batch   55/134 - Train Accuracy:  0.965, Validation Accuracy:  0.968, Loss:  0.041
Epoch   8 Batch   56/134 - Train Accuracy:  0.963, Validation Accuracy:  0.968, Loss:  0.040
Epoch   8 Batch   57/134 - Train Accuracy:  0.966, Validation Accuracy:  0.969, Loss:  0.041
Epoch   8 Batch   58/134 - Train Accuracy:  0.967, Validation Accuracy:  0.969, Loss:  0.037
Epoch   8 Batch   59/134 - Train Accuracy:  0.958, Validation Accuracy:  0.967, Loss:  0.045
Epoch   8 Batch   60/134 - Train Accuracy:  0.967, Validation Accuracy:  0.967, Loss:  0.041
Epoch   8 Batch   61/134 - Train Accuracy:  0.965, Validation Accuracy:  0.966, Loss:  0.036
Epoch   8 Batch   62/134 - Train Accuracy:  0.966, Validation Accuracy:  0.966, Loss:  0.038
Epoch   8 Batch   63/134 - Train Accuracy:  0.966, Validation Accuracy:  0.969, Loss:  0.038
Epoch   8 Batch   64/134 - Train Accuracy:  0.961, Validation Accuracy:  0.967, Loss:  0.041
Epoch   8 Batch   65/134 - Train Accuracy:  0.954, Validation Accuracy:  0.965, Loss:  0.043
Epoch   8 Batch   66/134 - Train Accuracy:  0.962, Validation Accuracy:  0.964, Loss:  0.037
Epoch   8 Batch   67/134 - Train Accuracy:  0.953, Validation Accuracy:  0.962, Loss:  0.045
Epoch   8 Batch   68/134 - Train Accuracy:  0.958, Validation Accuracy:  0.966, Loss:  0.046
Epoch   8 Batch   69/134 - Train Accuracy:  0.963, Validation Accuracy:  0.970, Loss:  0.041
Epoch   8 Batch   70/134 - Train Accuracy:  0.963, Validation Accuracy:  0.969, Loss:  0.040
Epoch   8 Batch   71/134 - Train Accuracy:  0.971, Validation Accuracy:  0.968, Loss:  0.034
Epoch   8 Batch   72/134 - Train Accuracy:  0.967, Validation Accuracy:  0.966, Loss:  0.039
Epoch   8 Batch   73/134 - Train Accuracy:  0.971, Validation Accuracy:  0.962, Loss:  0.041
Epoch   8 Batch   74/134 - Train Accuracy:  0.955, Validation Accuracy:  0.964, Loss:  0.046
Epoch   8 Batch   75/134 - Train Accuracy:  0.964, Validation Accuracy:  0.967, Loss:  0.042
Epoch   8 Batch   76/134 - Train Accuracy:  0.970, Validation Accuracy:  0.968, Loss:  0.040
Epoch   8 Batch   77/134 - Train Accuracy:  0.965, Validation Accuracy:  0.968, Loss:  0.040
Epoch   8 Batch   78/134 - Train Accuracy:  0.958, Validation Accuracy:  0.969, Loss:  0.041
Epoch   8 Batch   79/134 - Train Accuracy:  0.961, Validation Accuracy:  0.966, Loss:  0.042
Epoch   8 Batch   80/134 - Train Accuracy:  0.969, Validation Accuracy:  0.966, Loss:  0.039
Epoch   8 Batch   81/134 - Train Accuracy:  0.969, Validation Accuracy:  0.969, Loss:  0.037
Epoch   8 Batch   82/134 - Train Accuracy:  0.969, Validation Accuracy:  0.967, Loss:  0.037
Epoch   8 Batch   83/134 - Train Accuracy:  0.969, Validation Accuracy:  0.966, Loss:  0.041
Epoch   8 Batch   84/134 - Train Accuracy:  0.960, Validation Accuracy:  0.966, Loss:  0.039
Epoch   8 Batch   85/134 - Train Accuracy:  0.967, Validation Accuracy:  0.968, Loss:  0.044
Epoch   8 Batch   86/134 - Train Accuracy:  0.965, Validation Accuracy:  0.966, Loss:  0.039
Epoch   8 Batch   87/134 - Train Accuracy:  0.961, Validation Accuracy:  0.965, Loss:  0.048
Epoch   8 Batch   88/134 - Train Accuracy:  0.972, Validation Accuracy:  0.965, Loss:  0.036
Epoch   8 Batch   89/134 - Train Accuracy:  0.966, Validation Accuracy:  0.967, Loss:  0.036
Epoch   8 Batch   90/134 - Train Accuracy:  0.961, Validation Accuracy:  0.965, Loss:  0.045
Epoch   8 Batch   91/134 - Train Accuracy:  0.973, Validation Accuracy:  0.964, Loss:  0.032
Epoch   8 Batch   92/134 - Train Accuracy:  0.975, Validation Accuracy:  0.965, Loss:  0.037
Epoch   8 Batch   93/134 - Train Accuracy:  0.970, Validation Accuracy:  0.966, Loss:  0.039
Epoch   8 Batch   94/134 - Train Accuracy:  0.966, Validation Accuracy:  0.968, Loss:  0.041
Epoch   8 Batch   95/134 - Train Accuracy:  0.968, Validation Accuracy:  0.969, Loss:  0.039
Epoch   8 Batch   96/134 - Train Accuracy:  0.956, Validation Accuracy:  0.968, Loss:  0.040
Epoch   8 Batch   97/134 - Train Accuracy:  0.959, Validation Accuracy:  0.968, Loss:  0.040
Epoch   8 Batch   98/134 - Train Accuracy:  0.968, Validation Accuracy:  0.967, Loss:  0.039
Epoch   8 Batch   99/134 - Train Accuracy:  0.974, Validation Accuracy:  0.967, Loss:  0.038
Epoch   8 Batch  100/134 - Train Accuracy:  0.968, Validation Accuracy:  0.972, Loss:  0.042
Epoch   8 Batch  101/134 - Train Accuracy:  0.964, Validation Accuracy:  0.971, Loss:  0.040
Epoch   8 Batch  102/134 - Train Accuracy:  0.966, Validation Accuracy:  0.970, Loss:  0.042
Epoch   8 Batch  103/134 - Train Accuracy:  0.968, Validation Accuracy:  0.971, Loss:  0.040
Epoch   8 Batch  104/134 - Train Accuracy:  0.970, Validation Accuracy:  0.973, Loss:  0.035
Epoch   8 Batch  105/134 - Train Accuracy:  0.963, Validation Accuracy:  0.972, Loss:  0.045
Epoch   8 Batch  106/134 - Train Accuracy:  0.961, Validation Accuracy:  0.972, Loss:  0.041
Epoch   8 Batch  107/134 - Train Accuracy:  0.961, Validation Accuracy:  0.970, Loss:  0.045
Epoch   8 Batch  108/134 - Train Accuracy:  0.962, Validation Accuracy:  0.969, Loss:  0.042
Epoch   8 Batch  109/134 - Train Accuracy:  0.965, Validation Accuracy:  0.967, Loss:  0.038
Epoch   8 Batch  110/134 - Train Accuracy:  0.967, Validation Accuracy:  0.966, Loss:  0.039
Epoch   8 Batch  111/134 - Train Accuracy:  0.963, Validation Accuracy:  0.967, Loss:  0.040
Epoch   8 Batch  112/134 - Train Accuracy:  0.964, Validation Accuracy:  0.970, Loss:  0.040
Epoch   8 Batch  113/134 - Train Accuracy:  0.969, Validation Accuracy:  0.969, Loss:  0.038
Epoch   8 Batch  114/134 - Train Accuracy:  0.966, Validation Accuracy:  0.969, Loss:  0.036
Epoch   8 Batch  115/134 - Train Accuracy:  0.961, Validation Accuracy:  0.969, Loss:  0.037
Epoch   8 Batch  116/134 - Train Accuracy:  0.969, Validation Accuracy:  0.971, Loss:  0.040
Epoch   8 Batch  117/134 - Train Accuracy:  0.978, Validation Accuracy:  0.971, Loss:  0.034
Epoch   8 Batch  118/134 - Train Accuracy:  0.967, Validation Accuracy:  0.973, Loss:  0.038
Epoch   8 Batch  119/134 - Train Accuracy:  0.969, Validation Accuracy:  0.974, Loss:  0.037
Epoch   8 Batch  120/134 - Train Accuracy:  0.962, Validation Accuracy:  0.973, Loss:  0.042
Epoch   8 Batch  121/134 - Train Accuracy:  0.964, Validation Accuracy:  0.973, Loss:  0.037
Epoch   8 Batch  122/134 - Train Accuracy:  0.962, Validation Accuracy:  0.972, Loss:  0.039
Epoch   8 Batch  123/134 - Train Accuracy:  0.970, Validation Accuracy:  0.973, Loss:  0.035
Epoch   8 Batch  124/134 - Train Accuracy:  0.964, Validation Accuracy:  0.971, Loss:  0.036
Epoch   8 Batch  125/134 - Train Accuracy:  0.971, Validation Accuracy:  0.970, Loss:  0.035
Epoch   8 Batch  126/134 - Train Accuracy:  0.965, Validation Accuracy:  0.971, Loss:  0.039
Epoch   8 Batch  127/134 - Train Accuracy:  0.962, Validation Accuracy:  0.965, Loss:  0.037
Epoch   8 Batch  128/134 - Train Accuracy:  0.961, Validation Accuracy:  0.966, Loss:  0.041
Epoch   8 Batch  129/134 - Train Accuracy:  0.966, Validation Accuracy:  0.967, Loss:  0.041
Epoch   8 Batch  130/134 - Train Accuracy:  0.966, Validation Accuracy:  0.967, Loss:  0.040
Epoch   8 Batch  131/134 - Train Accuracy:  0.971, Validation Accuracy:  0.973, Loss:  0.040
Epoch   8 Batch  132/134 - Train Accuracy:  0.967, Validation Accuracy:  0.974, Loss:  0.034
Epoch   9 Batch    0/134 - Train Accuracy:  0.971, Validation Accuracy:  0.972, Loss:  0.036
Epoch   9 Batch    1/134 - Train Accuracy:  0.965, Validation Accuracy:  0.973, Loss:  0.038
Epoch   9 Batch    2/134 - Train Accuracy:  0.967, Validation Accuracy:  0.971, Loss:  0.036
Epoch   9 Batch    3/134 - Train Accuracy:  0.969, Validation Accuracy:  0.971, Loss:  0.041
Epoch   9 Batch    4/134 - Train Accuracy:  0.972, Validation Accuracy:  0.970, Loss:  0.036
Epoch   9 Batch    5/134 - Train Accuracy:  0.964, Validation Accuracy:  0.970, Loss:  0.044
Epoch   9 Batch    6/134 - Train Accuracy:  0.972, Validation Accuracy:  0.969, Loss:  0.035
Epoch   9 Batch    7/134 - Train Accuracy:  0.969, Validation Accuracy:  0.970, Loss:  0.038
Epoch   9 Batch    8/134 - Train Accuracy:  0.965, Validation Accuracy:  0.966, Loss:  0.035
Epoch   9 Batch    9/134 - Train Accuracy:  0.967, Validation Accuracy:  0.967, Loss:  0.033
Epoch   9 Batch   10/134 - Train Accuracy:  0.961, Validation Accuracy:  0.970, Loss:  0.038
Epoch   9 Batch   11/134 - Train Accuracy:  0.969, Validation Accuracy:  0.971, Loss:  0.040
Epoch   9 Batch   12/134 - Train Accuracy:  0.965, Validation Accuracy:  0.970, Loss:  0.036
Epoch   9 Batch   13/134 - Train Accuracy:  0.962, Validation Accuracy:  0.971, Loss:  0.040
Epoch   9 Batch   14/134 - Train Accuracy:  0.966, Validation Accuracy:  0.971, Loss:  0.039
Epoch   9 Batch   15/134 - Train Accuracy:  0.968, Validation Accuracy:  0.970, Loss:  0.037
Epoch   9 Batch   16/134 - Train Accuracy:  0.965, Validation Accuracy:  0.966, Loss:  0.035
Epoch   9 Batch   17/134 - Train Accuracy:  0.961, Validation Accuracy:  0.967, Loss:  0.044
Epoch   9 Batch   18/134 - Train Accuracy:  0.966, Validation Accuracy:  0.971, Loss:  0.037
Epoch   9 Batch   19/134 - Train Accuracy:  0.961, Validation Accuracy:  0.969, Loss:  0.038
Epoch   9 Batch   20/134 - Train Accuracy:  0.971, Validation Accuracy:  0.965, Loss:  0.037
Epoch   9 Batch   21/134 - Train Accuracy:  0.965, Validation Accuracy:  0.965, Loss:  0.041
Epoch   9 Batch   22/134 - Train Accuracy:  0.959, Validation Accuracy:  0.965, Loss:  0.041
Epoch   9 Batch   23/134 - Train Accuracy:  0.965, Validation Accuracy:  0.969, Loss:  0.034
Epoch   9 Batch   24/134 - Train Accuracy:  0.964, Validation Accuracy:  0.969, Loss:  0.042
Epoch   9 Batch   25/134 - Train Accuracy:  0.966, Validation Accuracy:  0.970, Loss:  0.032
Epoch   9 Batch   26/134 - Train Accuracy:  0.966, Validation Accuracy:  0.967, Loss:  0.039
Epoch   9 Batch   27/134 - Train Accuracy:  0.964, Validation Accuracy:  0.967, Loss:  0.038
Epoch   9 Batch   28/134 - Train Accuracy:  0.965, Validation Accuracy:  0.969, Loss:  0.042
Epoch   9 Batch   29/134 - Train Accuracy:  0.971, Validation Accuracy:  0.969, Loss:  0.032
Epoch   9 Batch   30/134 - Train Accuracy:  0.955, Validation Accuracy:  0.968, Loss:  0.041
Epoch   9 Batch   31/134 - Train Accuracy:  0.965, Validation Accuracy:  0.968, Loss:  0.037
Epoch   9 Batch   32/134 - Train Accuracy:  0.967, Validation Accuracy:  0.969, Loss:  0.037
Epoch   9 Batch   33/134 - Train Accuracy:  0.963, Validation Accuracy:  0.968, Loss:  0.044
Epoch   9 Batch   34/134 - Train Accuracy:  0.965, Validation Accuracy:  0.969, Loss:  0.040
Epoch   9 Batch   35/134 - Train Accuracy:  0.967, Validation Accuracy:  0.969, Loss:  0.040
Epoch   9 Batch   36/134 - Train Accuracy:  0.965, Validation Accuracy:  0.968, Loss:  0.040
Epoch   9 Batch   37/134 - Train Accuracy:  0.963, Validation Accuracy:  0.964, Loss:  0.038
Epoch   9 Batch   38/134 - Train Accuracy:  0.975, Validation Accuracy:  0.966, Loss:  0.037
Epoch   9 Batch   39/134 - Train Accuracy:  0.972, Validation Accuracy:  0.966, Loss:  0.036
Epoch   9 Batch   40/134 - Train Accuracy:  0.967, Validation Accuracy:  0.967, Loss:  0.035
Epoch   9 Batch   41/134 - Train Accuracy:  0.963, Validation Accuracy:  0.970, Loss:  0.042
Epoch   9 Batch   42/134 - Train Accuracy:  0.964, Validation Accuracy:  0.970, Loss:  0.033
Epoch   9 Batch   43/134 - Train Accuracy:  0.965, Validation Accuracy:  0.968, Loss:  0.038
Epoch   9 Batch   44/134 - Train Accuracy:  0.965, Validation Accuracy:  0.971, Loss:  0.037
Epoch   9 Batch   45/134 - Train Accuracy:  0.972, Validation Accuracy:  0.972, Loss:  0.035
Epoch   9 Batch   46/134 - Train Accuracy:  0.967, Validation Accuracy:  0.972, Loss:  0.040
Epoch   9 Batch   47/134 - Train Accuracy:  0.965, Validation Accuracy:  0.972, Loss:  0.033
Epoch   9 Batch   48/134 - Train Accuracy:  0.965, Validation Accuracy:  0.972, Loss:  0.040
Epoch   9 Batch   49/134 - Train Accuracy:  0.964, Validation Accuracy:  0.972, Loss:  0.034
Epoch   9 Batch   50/134 - Train Accuracy:  0.963, Validation Accuracy:  0.970, Loss:  0.040
Epoch   9 Batch   51/134 - Train Accuracy:  0.970, Validation Accuracy:  0.969, Loss:  0.038
Epoch   9 Batch   52/134 - Train Accuracy:  0.969, Validation Accuracy:  0.970, Loss:  0.035
Epoch   9 Batch   53/134 - Train Accuracy:  0.969, Validation Accuracy:  0.968, Loss:  0.034
Epoch   9 Batch   54/134 - Train Accuracy:  0.966, Validation Accuracy:  0.967, Loss:  0.036
Epoch   9 Batch   55/134 - Train Accuracy:  0.963, Validation Accuracy:  0.969, Loss:  0.041
Epoch   9 Batch   56/134 - Train Accuracy:  0.964, Validation Accuracy:  0.968, Loss:  0.036
Epoch   9 Batch   57/134 - Train Accuracy:  0.967, Validation Accuracy:  0.967, Loss:  0.037
Epoch   9 Batch   58/134 - Train Accuracy:  0.968, Validation Accuracy:  0.968, Loss:  0.032
Epoch   9 Batch   59/134 - Train Accuracy:  0.964, Validation Accuracy:  0.970, Loss:  0.040
Epoch   9 Batch   60/134 - Train Accuracy:  0.968, Validation Accuracy:  0.971, Loss:  0.035
Epoch   9 Batch   61/134 - Train Accuracy:  0.969, Validation Accuracy:  0.972, Loss:  0.032
Epoch   9 Batch   62/134 - Train Accuracy:  0.964, Validation Accuracy:  0.971, Loss:  0.034
Epoch   9 Batch   63/134 - Train Accuracy:  0.972, Validation Accuracy:  0.971, Loss:  0.035
Epoch   9 Batch   64/134 - Train Accuracy:  0.963, Validation Accuracy:  0.970, Loss:  0.039
Epoch   9 Batch   65/134 - Train Accuracy:  0.962, Validation Accuracy:  0.969, Loss:  0.035
Epoch   9 Batch   66/134 - Train Accuracy:  0.971, Validation Accuracy:  0.970, Loss:  0.033
Epoch   9 Batch   67/134 - Train Accuracy:  0.962, Validation Accuracy:  0.972, Loss:  0.037
Epoch   9 Batch   68/134 - Train Accuracy:  0.959, Validation Accuracy:  0.970, Loss:  0.039
Epoch   9 Batch   69/134 - Train Accuracy:  0.964, Validation Accuracy:  0.973, Loss:  0.037
Epoch   9 Batch   70/134 - Train Accuracy:  0.971, Validation Accuracy:  0.973, Loss:  0.040
Epoch   9 Batch   71/134 - Train Accuracy:  0.963, Validation Accuracy:  0.971, Loss:  0.031
Epoch   9 Batch   72/134 - Train Accuracy:  0.969, Validation Accuracy:  0.972, Loss:  0.034
Epoch   9 Batch   73/134 - Train Accuracy:  0.972, Validation Accuracy:  0.972, Loss:  0.039
Epoch   9 Batch   74/134 - Train Accuracy:  0.960, Validation Accuracy:  0.970, Loss:  0.040
Epoch   9 Batch   75/134 - Train Accuracy:  0.971, Validation Accuracy:  0.970, Loss:  0.038
Epoch   9 Batch   76/134 - Train Accuracy:  0.972, Validation Accuracy:  0.972, Loss:  0.034
Epoch   9 Batch   77/134 - Train Accuracy:  0.972, Validation Accuracy:  0.971, Loss:  0.034
Epoch   9 Batch   78/134 - Train Accuracy:  0.966, Validation Accuracy:  0.969, Loss:  0.032
Epoch   9 Batch   79/134 - Train Accuracy:  0.967, Validation Accuracy:  0.970, Loss:  0.033
Epoch   9 Batch   80/134 - Train Accuracy:  0.972, Validation Accuracy:  0.971, Loss:  0.034
Epoch   9 Batch   81/134 - Train Accuracy:  0.974, Validation Accuracy:  0.974, Loss:  0.034
Epoch   9 Batch   82/134 - Train Accuracy:  0.973, Validation Accuracy:  0.971, Loss:  0.031
Epoch   9 Batch   83/134 - Train Accuracy:  0.971, Validation Accuracy:  0.972, Loss:  0.036
Epoch   9 Batch   84/134 - Train Accuracy:  0.963, Validation Accuracy:  0.970, Loss:  0.035
Epoch   9 Batch   85/134 - Train Accuracy:  0.966, Validation Accuracy:  0.970, Loss:  0.035
Epoch   9 Batch   86/134 - Train Accuracy:  0.968, Validation Accuracy:  0.970, Loss:  0.036
Epoch   9 Batch   87/134 - Train Accuracy:  0.964, Validation Accuracy:  0.972, Loss:  0.043
Epoch   9 Batch   88/134 - Train Accuracy:  0.973, Validation Accuracy:  0.973, Loss:  0.033
Epoch   9 Batch   89/134 - Train Accuracy:  0.965, Validation Accuracy:  0.975, Loss:  0.034
Epoch   9 Batch   90/134 - Train Accuracy:  0.960, Validation Accuracy:  0.974, Loss:  0.040
Epoch   9 Batch   91/134 - Train Accuracy:  0.971, Validation Accuracy:  0.975, Loss:  0.027
Epoch   9 Batch   92/134 - Train Accuracy:  0.976, Validation Accuracy:  0.975, Loss:  0.033
Epoch   9 Batch   93/134 - Train Accuracy:  0.969, Validation Accuracy:  0.973, Loss:  0.034
Epoch   9 Batch   94/134 - Train Accuracy:  0.963, Validation Accuracy:  0.970, Loss:  0.037
Epoch   9 Batch   95/134 - Train Accuracy:  0.971, Validation Accuracy:  0.969, Loss:  0.037
Epoch   9 Batch   96/134 - Train Accuracy:  0.962, Validation Accuracy:  0.973, Loss:  0.038
Epoch   9 Batch   97/134 - Train Accuracy:  0.963, Validation Accuracy:  0.968, Loss:  0.033
Epoch   9 Batch   98/134 - Train Accuracy:  0.971, Validation Accuracy:  0.971, Loss:  0.037
Epoch   9 Batch   99/134 - Train Accuracy:  0.973, Validation Accuracy:  0.972, Loss:  0.035
Epoch   9 Batch  100/134 - Train Accuracy:  0.969, Validation Accuracy:  0.971, Loss:  0.038
Epoch   9 Batch  101/134 - Train Accuracy:  0.966, Validation Accuracy:  0.969, Loss:  0.035
Epoch   9 Batch  102/134 - Train Accuracy:  0.968, Validation Accuracy:  0.969, Loss:  0.035
Epoch   9 Batch  103/134 - Train Accuracy:  0.965, Validation Accuracy:  0.969, Loss:  0.033
Epoch   9 Batch  104/134 - Train Accuracy:  0.972, Validation Accuracy:  0.971, Loss:  0.033
Epoch   9 Batch  105/134 - Train Accuracy:  0.958, Validation Accuracy:  0.972, Loss:  0.043
Epoch   9 Batch  106/134 - Train Accuracy:  0.966, Validation Accuracy:  0.972, Loss:  0.034
Epoch   9 Batch  107/134 - Train Accuracy:  0.966, Validation Accuracy:  0.972, Loss:  0.039
Epoch   9 Batch  108/134 - Train Accuracy:  0.963, Validation Accuracy:  0.973, Loss:  0.034
Epoch   9 Batch  109/134 - Train Accuracy:  0.966, Validation Accuracy:  0.974, Loss:  0.034
Epoch   9 Batch  110/134 - Train Accuracy:  0.969, Validation Accuracy:  0.974, Loss:  0.035
Epoch   9 Batch  111/134 - Train Accuracy:  0.965, Validation Accuracy:  0.970, Loss:  0.040
Epoch   9 Batch  112/134 - Train Accuracy:  0.967, Validation Accuracy:  0.965, Loss:  0.035
Epoch   9 Batch  113/134 - Train Accuracy:  0.969, Validation Accuracy:  0.966, Loss:  0.040
Epoch   9 Batch  114/134 - Train Accuracy:  0.969, Validation Accuracy:  0.971, Loss:  0.037
Epoch   9 Batch  115/134 - Train Accuracy:  0.961, Validation Accuracy:  0.972, Loss:  0.036
Epoch   9 Batch  116/134 - Train Accuracy:  0.973, Validation Accuracy:  0.967, Loss:  0.037
Epoch   9 Batch  117/134 - Train Accuracy:  0.974, Validation Accuracy:  0.968, Loss:  0.030
Epoch   9 Batch  118/134 - Train Accuracy:  0.975, Validation Accuracy:  0.969, Loss:  0.039
Epoch   9 Batch  119/134 - Train Accuracy:  0.970, Validation Accuracy:  0.967, Loss:  0.034
Epoch   9 Batch  120/134 - Train Accuracy:  0.963, Validation Accuracy:  0.970, Loss:  0.041
Epoch   9 Batch  121/134 - Train Accuracy:  0.969, Validation Accuracy:  0.973, Loss:  0.034
Epoch   9 Batch  122/134 - Train Accuracy:  0.963, Validation Accuracy:  0.974, Loss:  0.036
Epoch   9 Batch  123/134 - Train Accuracy:  0.973, Validation Accuracy:  0.972, Loss:  0.034
Epoch   9 Batch  124/134 - Train Accuracy:  0.969, Validation Accuracy:  0.971, Loss:  0.033
Epoch   9 Batch  125/134 - Train Accuracy:  0.972, Validation Accuracy:  0.969, Loss:  0.032
Epoch   9 Batch  126/134 - Train Accuracy:  0.966, Validation Accuracy:  0.972, Loss:  0.037
Epoch   9 Batch  127/134 - Train Accuracy:  0.966, Validation Accuracy:  0.971, Loss:  0.036
Epoch   9 Batch  128/134 - Train Accuracy:  0.960, Validation Accuracy:  0.968, Loss:  0.039
Epoch   9 Batch  129/134 - Train Accuracy:  0.965, Validation Accuracy:  0.967, Loss:  0.037
Epoch   9 Batch  130/134 - Train Accuracy:  0.968, Validation Accuracy:  0.968, Loss:  0.034
Epoch   9 Batch  131/134 - Train Accuracy:  0.963, Validation Accuracy:  0.973, Loss:  0.039
Epoch   9 Batch  132/134 - Train Accuracy:  0.971, Validation Accuracy:  0.973, Loss:  0.032
Model Trained and Saved

Save Parameters

Save the batch_size and save_path parameters for inference.





In [36]:



    


"""
DON'T MODIFY ANYTHING IN THIS CELL
"""
# Save parameters for checkpoint
helper.save_params(save_path)

Checkpoint





In [37]:



    


"""
DON'T MODIFY ANYTHING IN THIS CELL
"""
import tensorflow as tf
import numpy as np
import helper
import problem_unittests as tests

_, (source_vocab_to_int, target_vocab_to_int), (source_int_to_vocab, target_int_to_vocab) = helper.load_preprocess()
load_path = helper.load_params()

Sentence to Sequence

To feed a sentence into the model for translation, you first need to preprocess it. Implement the function sentence_to_seq() to preprocess new sentences.

  • Convert the sentence to lowercase
  • Convert words into ids using vocab_to_int
    • Convert words not in the vocabulary, to the <UNK> word id.




In [40]:



    


def sentence_to_seq(sentence, vocab_to_int):
    """
    Convert a sentence to a sequence of ids
    :param sentence: String
    :param vocab_to_int: Dictionary to go from the words to an id
    :return: List of word ids
    """
    
    sentence_lower = sentence.lower()
    word_ids = []
    
    for word in sentence_lower.split():
        
        word_id = vocab_to_int.get(word, vocab_to_int['<UNK>'])
        word_ids.append(word_id)
     
    return word_ids


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



    


















    






Tests Passed

Translate

This will translate translate_sentence from English to French.





In [54]:



    


translate_sentence = 'he saw a old yellow truck .'

# translate_sentence = "New Jersey is usually chilly during july , and it is usually freezing in november"

"""
DON'T MODIFY ANYTHING IN THIS CELL
"""
translate_sentence = sentence_to_seq(translate_sentence, source_vocab_to_int)

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

    input_data = loaded_graph.get_tensor_by_name('input:0')
    logits = loaded_graph.get_tensor_by_name('predictions:0')
    target_sequence_length = loaded_graph.get_tensor_by_name('target_sequence_length:0')
    source_sequence_length = loaded_graph.get_tensor_by_name('source_sequence_length:0')
    keep_prob = loaded_graph.get_tensor_by_name('keep_prob:0')

    translate_logits = sess.run(logits, {input_data: [translate_sentence]*batch_size,
                                         target_sequence_length: [len(translate_sentence)*2]*batch_size,
                                         source_sequence_length: [len(translate_sentence)]*batch_size,
                                         keep_prob: 1.0})[0]

print('Input')
print('  Word Ids:      {}'.format([i for i in translate_sentence]))
print('  English Words: {}'.format([source_int_to_vocab[i] for i in translate_sentence]))

print('\nPrediction')
print('  Word Ids:      {}'.format([i for i in translate_logits]))
print('  French Words: {}'.format(" ".join([target_int_to_vocab[i] for i in translate_logits])))



    


















    






Input
  Word Ids:      [103, 219, 135, 211, 136, 92, 56]
  English Words: ['he', 'saw', 'a', 'old', 'yellow', 'truck', '.']

Prediction
  Word Ids:      [316, 321, 115, 276, 269, 341, 281, 112, 1]
  French Words: ['il', 'a', 'vu', 'un', 'vieux', 'camion', 'jaune', '.', '<EOS>']

Imperfect Translation

You might notice that some sentences translate better than others. Since the dataset you're using only has a vocabulary of 227 English words of the thousands that you use, you're only going to see good results using these words. For this project, you don't need a perfect translation. However, if you want to create a better translation model, you'll need better data.

You can train on the WMT10 French-English corpus. This dataset has more vocabulary and richer in topics discussed. However, this will take you days to train, so make sure you've a GPU and the neural network is performing well on dataset we provided. Just make sure you play with the WMT10 corpus after you've submitted this project.

Submitting This Project

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

Content source: lukechen526/deep-learning

Similar notebooks:

notebook.community | gallery | about