 View on TensorFlow.org
|
 Run in Google Colab
|
 View source on GitHub
|
 Download notebook
|
This tutorial demonstrates how to generate text using a character-based RNN. We will work with a dataset of Shakespeare's writing from Andrej Karpathy's The Unreasonable Effectiveness of Recurrent Neural Networks. Given a sequence of characters from this data ("Shakespear"), train a model to predict the next character in the sequence ("e"). Longer sequences of text can be generated by calling the model repeatedly.
This tutorial includes runnable code implemented using tf.keras and eager execution. The following is sample output when the model in this tutorial trained for 30 epochs, and started with the string "Q":
QUEENE: I had thought thou hadst a Roman; for the oracle, Thus by All bids the man against the word, Which are so weak of care, by old care done; Your children were in your holy love, And the precipitation through the bleeding throne. BISHOP OF ELY: Marry, and will, my lord, to weep in such a one were prettiest; Yet now I was adopted heir Of the world's lamentable day, To watch the next way with his father with his face? ESCALUS: The cause why then we are all resolved more sons. VOLUMNIA: O, no, no, no, no, no, no, no, no, no, no, no, no, no, no, no, no, no, no, no, no, it is no sin it should be dead, And love and pale as any will to that word. QUEEN ELIZABETH: But how long have I heard the soul for this world, And show his hands of life be proved to stand. PETRUCHIO: I say he look'd on, if I must be content To stay him from the fatal of our country's bliss. His lordship pluck'd from this sentence then for prey, And then let us twain, being the moon, were she such a case as fills m
While some of the sentences are grammatical, most do not make sense. The model has not learned the meaning of words, but consider:
The model is character-based. When training started, the model did not know how to spell an English word, or that words were even a unit of text.
The structure of the output resembles a play—blocks of text generally begin with a speaker name, in all capital letters similar to the dataset.
As demonstrated below, the model is trained on small batches of text (100 characters each), and is still able to generate a longer sequence of text with coherent structure.
Setup
Import TensorFlow and other libraries
import tensorflow as tf
import numpy as np
import os
import time
Download the Shakespeare dataset
Change the following line to run this code on your own data.
path_to_file = tf.keras.utils.get_file('shakespeare.txt', 'https://storage.googleapis.com/download.tensorflow.org/data/shakespeare.txt')
Downloading data from https://storage.googleapis.com/download.tensorflow.org/data/shakespeare.txt 1122304/1115394 [==============================] - 0s 0us/step
Read the data
First, look in the text:
# Read, then decode for py2 compat.
text = open(path_to_file, 'rb').read().decode(encoding='utf-8')
# length of text is the number of characters in it
print ('Length of text: {} characters'.format(len(text)))
Length of text: 1115394 characters
# Take a look at the first 250 characters in text
print(text[:250])
First Citizen: Before we proceed any further, hear me speak. All: Speak, speak. First Citizen: You are all resolved rather to die than to famish? All: Resolved. resolved. First Citizen: First, you know Caius Marcius is chief enemy to the people.
# The unique characters in the file
vocab = sorted(set(text))
print ('{} unique characters'.format(len(vocab)))
65 unique characters
Process the text
Vectorize the text
Before training, we need to map strings to a numerical representation. Create two lookup tables: one mapping characters to numbers, and another for numbers to characters.
# Creating a mapping from unique characters to indices
char2idx = {u:i for i, u in enumerate(vocab)}
idx2char = np.array(vocab)
text_as_int = np.array([char2idx[c] for c in text])
Now we have an integer representation for each character. Notice that we mapped the character as indexes from 0 to len(unique).
print('{')
for char,_ in zip(char2idx, range(20)):
print(' {:4s}: {:3d},'.format(repr(char), char2idx[char]))
print(' ...\n}')
{
'\n': 0,
' ' : 1,
'!' : 2,
'$' : 3,
'&' : 4,
"'" : 5,
',' : 6,
'-' : 7,
'.' : 8,
'3' : 9,
':' : 10,
';' : 11,
'?' : 12,
'A' : 13,
'B' : 14,
'C' : 15,
'D' : 16,
'E' : 17,
'F' : 18,
'G' : 19,
...
}
# Show how the first 13 characters from the text are mapped to integers
print ('{} ---- characters mapped to int ---- > {}'.format(repr(text[:13]), text_as_int[:13]))
'First Citizen' ---- characters mapped to int ---- > [18 47 56 57 58 1 15 47 58 47 64 43 52]
The prediction task
Given a character, or a sequence of characters, what is the most probable next character? This is the task we're training the model to perform. The input to the model will be a sequence of characters, and we train the model to predict the output—the following character at each time step.
Since RNNs maintain an internal state that depends on the previously seen elements, given all the characters computed until this moment, what is the next character?
Create training examples and targets
Next divide the text into example sequences. Each input sequence will contain seq_length characters from the text.
For each input sequence, the corresponding targets contain the same length of text, except shifted one character to the right.
So break the text into chunks of seq_length+1. For example, say seq_length is 4 and our text is "Hello". The input sequence would be "Hell", and the target sequence "ello".
To do this first use the tf.data.Dataset.from_tensor_slices function to convert the text vector into a stream of character indices.
# The maximum length sentence we want for a single input in characters
seq_length = 100
examples_per_epoch = len(text)//(seq_length+1)
# Create training examples / targets
char_dataset = tf.data.Dataset.from_tensor_slices(text_as_int)
for i in char_dataset.take(5):
print(idx2char[i.numpy()])
F i r s t
The batch method lets us easily convert these individual characters to sequences of the desired size.
sequences = char_dataset.batch(seq_length+1, drop_remainder=True)
for item in sequences.take(5):
print(repr(''.join(idx2char[item.numpy()])))
'First Citizen:\nBefore we proceed any further, hear me speak.\n\nAll:\nSpeak, speak.\n\nFirst Citizen:\nYou ' 'are all resolved rather to die than to famish?\n\nAll:\nResolved. resolved.\n\nFirst Citizen:\nFirst, you k' "now Caius Marcius is chief enemy to the people.\n\nAll:\nWe know't, we know't.\n\nFirst Citizen:\nLet us ki" "ll him, and we'll have corn at our own price.\nIs't a verdict?\n\nAll:\nNo more talking on't; let it be d" 'one: away, away!\n\nSecond Citizen:\nOne word, good citizens.\n\nFirst Citizen:\nWe are accounted poor citi'
For each sequence, duplicate and shift it to form the input and target text by using the map method to apply a simple function to each batch:
def split_input_target(chunk):
input_text = chunk[:-1]
target_text = chunk[1:]
return input_text, target_text
dataset = sequences.map(split_input_target)
Print the first examples input and target values:
for input_example, target_example in dataset.take(1):
print ('Input data: ', repr(''.join(idx2char[input_example.numpy()])))
print ('Target data:', repr(''.join(idx2char[target_example.numpy()])))
Input data: 'First Citizen:\nBefore we proceed any further, hear me speak.\n\nAll:\nSpeak, speak.\n\nFirst Citizen:\nYou' Target data: 'irst Citizen:\nBefore we proceed any further, hear me speak.\n\nAll:\nSpeak, speak.\n\nFirst Citizen:\nYou '
Each index of these vectors are processed as one time step. For the input at time step 0, the model receives the index for "F" and trys to predict the index for "i" as the next character. At the next timestep, it does the same thing but the RNN considers the previous step context in addition to the current input character.
for i, (input_idx, target_idx) in enumerate(zip(input_example[:5], target_example[:5])):
print("Step {:4d}".format(i))
print(" input: {} ({:s})".format(input_idx, repr(idx2char[input_idx])))
print(" expected output: {} ({:s})".format(target_idx, repr(idx2char[target_idx])))
Step 0
input: 18 ('F')
expected output: 47 ('i')
Step 1
input: 47 ('i')
expected output: 56 ('r')
Step 2
input: 56 ('r')
expected output: 57 ('s')
Step 3
input: 57 ('s')
expected output: 58 ('t')
Step 4
input: 58 ('t')
expected output: 1 (' ')
Create training batches
We used tf.data to split the text into manageable sequences. But before feeding this data into the model, we need to shuffle the data and pack it into batches.
# Batch size
BATCH_SIZE = 64
# Buffer size to shuffle the dataset
# (TF data is designed to work with possibly infinite sequences,
# so it doesn't attempt to shuffle the entire sequence in memory. Instead,
# it maintains a buffer in which it shuffles elements).
BUFFER_SIZE = 10000
dataset = dataset.shuffle(BUFFER_SIZE).batch(BATCH_SIZE, drop_remainder=True)
dataset
<BatchDataset shapes: ((64, 100), (64, 100)), types: (tf.int64, tf.int64)>
Build The Model
Use tf.keras.Sequential to define the model. For this simple example three layers are used to define our model:
tf.keras.layers.Embedding: The input layer. A trainable lookup table that will map the numbers of each character to a vector withembedding_dimdimensions;tf.keras.layers.GRU: A type of RNN with sizeunits=rnn_units(You can also use a LSTM layer here.)tf.keras.layers.Dense: The output layer, withvocab_sizeoutputs.
# Length of the vocabulary in chars
vocab_size = len(vocab)
# The embedding dimension
embedding_dim = 256
# Number of RNN units
rnn_units = 1024
def build_model(vocab_size, embedding_dim, rnn_units, batch_size):
model = tf.keras.Sequential([
tf.keras.layers.Embedding(vocab_size, embedding_dim,
batch_input_shape=[batch_size, None]),
tf.keras.layers.GRU(rnn_units,
return_sequences=True,
stateful=True,
recurrent_initializer='glorot_uniform'),
tf.keras.layers.Dense(vocab_size)
])
return model
model = build_model(
vocab_size = len(vocab),
embedding_dim=embedding_dim,
rnn_units=rnn_units,
batch_size=BATCH_SIZE)
For each character the model looks up the embedding, runs the GRU one timestep with the embedding as input, and applies the dense layer to generate logits predicting the log-likelihood of the next character:
Try the model
Now run the model to see that it behaves as expected.
First check the shape of the output:
for input_example_batch, target_example_batch in dataset.take(1):
example_batch_predictions = model(input_example_batch)
print(example_batch_predictions.shape, "# (batch_size, sequence_length, vocab_size)")
(64, 100, 65) # (batch_size, sequence_length, vocab_size)
In the above example the sequence length of the input is 100 but the model can be run on inputs of any length:
model.summary()
Model: "sequential" _________________________________________________________________ Layer (type) Output Shape Param # ================================================================= embedding (Embedding) (64, None, 256) 16640 _________________________________________________________________ gru (GRU) (64, None, 1024) 3938304 _________________________________________________________________ dense (Dense) (64, None, 65) 66625 ================================================================= Total params: 4,021,569 Trainable params: 4,021,569 Non-trainable params: 0 _________________________________________________________________
To get actual predictions from the model we need to sample from the output distribution, to get actual character indices. This distribution is defined by the logits over the character vocabulary.
Try it for the first example in the batch:
sampled_indices = tf.random.categorical(example_batch_predictions[0], num_samples=1)
sampled_indices = tf.squeeze(sampled_indices,axis=-1).numpy()
This gives us, at each timestep, a prediction of the next character index:
sampled_indices
array([29, 12, 22, 51, 1, 39, 50, 47, 56, 23, 49, 39, 23, 6, 49, 3, 15,
25, 19, 1, 50, 48, 51, 0, 9, 0, 59, 24, 24, 63, 59, 17, 60, 20,
30, 37, 28, 31, 39, 15, 27, 57, 0, 33, 20, 28, 39, 44, 53, 9, 19,
10, 4, 54, 44, 30, 19, 24, 31, 64, 49, 60, 60, 19, 35, 37, 31, 46,
53, 50, 60, 3, 52, 29, 35, 56, 57, 59, 16, 51, 14, 2, 42, 33, 6,
7, 13, 7, 32, 1, 42, 61, 35, 47, 25, 17, 55, 41, 21, 59])
Decode these to see the text predicted by this untrained model:
print("Input: \n", repr("".join(idx2char[input_example_batch[0]])))
print()
print("Next Char Predictions: \n", repr("".join(idx2char[sampled_indices ])))
Input: "\n\nHORTENSIO:\nSir, let me be so bold as ask you,\nDid you yet ever see Baptista's daughter?\n\nTRANIO:\nN" Next Char Predictions: 'Q?Jm alirKkaK,k$CMG ljm\n3\nuLLyuEvHRYPSaCOs\nUHPafo3G:&pfRGLSzkvvGWYSholv$nQWrsuDmB!dU,-A-T dwWiMEqcIu'
Train the model
At this point the problem can be treated as a standard classification problem. Given the previous RNN state, and the input this time step, predict the class of the next character.
Attach an optimizer, and a loss function
The standard tf.keras.losses.sparse_categorical_crossentropy loss function works in this case because it is applied across the last dimension of the predictions.
Because our model returns logits, we need to set the from_logits flag.
def loss(labels, logits):
return tf.keras.losses.sparse_categorical_crossentropy(labels, logits, from_logits=True)
example_batch_loss = loss(target_example_batch, example_batch_predictions)
print("Prediction shape: ", example_batch_predictions.shape, " # (batch_size, sequence_length, vocab_size)")
print("scalar_loss: ", example_batch_loss.numpy().mean())
Prediction shape: (64, 100, 65) # (batch_size, sequence_length, vocab_size) scalar_loss: 4.175127
Configure the training procedure using the tf.keras.Model.compile method. We'll use tf.keras.optimizers.Adam with default arguments and the loss function.
model.compile(optimizer='adam', loss=loss)
Configure checkpoints
Use a tf.keras.callbacks.ModelCheckpoint to ensure that checkpoints are saved during training:
# Directory where the checkpoints will be saved
checkpoint_dir = './training_checkpoints'
# Name of the checkpoint files
checkpoint_prefix = os.path.join(checkpoint_dir, "ckpt_{epoch}")
checkpoint_callback=tf.keras.callbacks.ModelCheckpoint(
filepath=checkpoint_prefix,
save_weights_only=True)
Execute the training
To keep training time reasonable, use 10 epochs to train the model. In Colab, set the runtime to GPU for faster training.
EPOCHS=10
history = model.fit(dataset, epochs=EPOCHS, callbacks=[checkpoint_callback])
Train for 172 steps Epoch 1/10 172/172 [==============================] - 7s 40ms/step - loss: 2.6536 Epoch 2/10 172/172 [==============================] - 6s 34ms/step - loss: 1.9498 Epoch 3/10 172/172 [==============================] - 6s 34ms/step - loss: 1.6860 Epoch 4/10 172/172 [==============================] - 6s 35ms/step - loss: 1.5411 Epoch 5/10 172/172 [==============================] - 6s 34ms/step - loss: 1.4541 Epoch 6/10 172/172 [==============================] - 6s 34ms/step - loss: 1.3952 Epoch 7/10 172/172 [==============================] - 6s 34ms/step - loss: 1.3497 Epoch 8/10 172/172 [==============================] - 6s 34ms/step - loss: 1.3097 Epoch 9/10 172/172 [==============================] - 6s 34ms/step - loss: 1.2764 Epoch 10/10 172/172 [==============================] - 6s 35ms/step - loss: 1.2435
Generate text
Restore the latest checkpoint
To keep this prediction step simple, use a batch size of 1.
Because of the way the RNN state is passed from timestep to timestep, the model only accepts a fixed batch size once built.
To run the model with a different batch_size, we need to rebuild the model and restore the weights from the checkpoint.
tf.train.latest_checkpoint(checkpoint_dir)
'./training_checkpoints/ckpt_10'
model = build_model(vocab_size, embedding_dim, rnn_units, batch_size=1)
model.load_weights(tf.train.latest_checkpoint(checkpoint_dir))
model.build(tf.TensorShape([1, None]))
model.summary()
Model: "sequential_1" _________________________________________________________________ Layer (type) Output Shape Param # ================================================================= embedding_1 (Embedding) (1, None, 256) 16640 _________________________________________________________________ gru_1 (GRU) (1, None, 1024) 3938304 _________________________________________________________________ dense_1 (Dense) (1, None, 65) 66625 ================================================================= Total params: 4,021,569 Trainable params: 4,021,569 Non-trainable params: 0 _________________________________________________________________
The prediction loop
The following code block generates the text:
It Starts by choosing a start string, initializing the RNN state and setting the number of characters to generate.
Get the prediction distribution of the next character using the start string and the RNN state.
Then, use a categorical distribution to calculate the index of the predicted character. Use this predicted character as our next input to the model.
The RNN state returned by the model is fed back into the model so that it now has more context, instead than only one character. After predicting the next character, the modified RNN states are again fed back into the model, which is how it learns as it gets more context from the previously predicted characters.
Looking at the generated text, you'll see the model knows when to capitalize, make paragraphs and imitates a Shakespeare-like writing vocabulary. With the small number of training epochs, it has not yet learned to form coherent sentences.
def generate_text(model, start_string):
# Evaluation step (generating text using the learned model)
# Number of characters to generate
num_generate = 1000
# Converting our start string to numbers (vectorizing)
input_eval = [char2idx[s] for s in start_string]
input_eval = tf.expand_dims(input_eval, 0)
# Empty string to store our results
text_generated = []
# Low temperatures results in more predictable text.
# Higher temperatures results in more surprising text.
# Experiment to find the best setting.
temperature = 1.0
# Here batch size == 1
model.reset_states()
for i in range(num_generate):
predictions = model(input_eval)
# remove the batch dimension
predictions = tf.squeeze(predictions, 0)
# using a categorical distribution to predict the character returned by the model
predictions = predictions / temperature
predicted_id = tf.random.categorical(predictions, num_samples=1)[-1,0].numpy()
# We pass the predicted character as the next input to the model
# along with the previous hidden state
input_eval = tf.expand_dims([predicted_id], 0)
text_generated.append(idx2char[predicted_id])
return (start_string + ''.join(text_generated))
print(generate_text(model, start_string=u"ROMEO: "))
ROMEO: Hastings with Your two make gross my horses did umacten, or taumbted You af their youth; we must here prick? LUCENTIO: Alas! I have heard his cales in beasts to his heart. HENRY BOLINGBROKE: In entertain in marth and well-more than e. ROMEO: O, mute awhile; but Montague with his as the world green sensied treasures of their winded gone, but by both excel With sughts and vowbrave youildaress proclaim prove. Gallenderous Let more than heaven? point, though good neire of mine, Who wounds the act ordenly. TRANIO: In all quit isle, being rude awrieve to fear my heart, Yet need. God-den claim of Menusos; how such a man the second Kingsoo, Look'd how concludents. Proceed those earthly in the moour for all twice; if you part; And Waling helm, hastily have the mild, Seeming twites that got the child-but he that? Hook! O thou with two sheet particular sword; whose careguard not your son, each one that is a woman of my foot! Furse: I never say 'gaiding and My brits no hare, she are true To
The easiest thing you can do to improve the results it to train it for longer (try EPOCHS=30).
You can also experiment with a different start string, or try adding another RNN layer to improve the model's accuracy, or adjusting the temperature parameter to generate more or less random predictions.
Advanced: Customized Training
The above training procedure is simple, but does not give you much control.
So now that you've seen how to run the model manually let's unpack the training loop, and implement it ourselves. This gives a starting point if, for example, to implement curriculum learning to help stabilize the model's open-loop output.
We will use tf.GradientTape to track the gradients. You can learn more about this approach by reading the eager execution guide.
The procedure works as follows:
First, initialize the RNN state. We do this by calling the
tf.keras.Model.reset_statesmethod.Next, iterate over the dataset (batch by batch) and calculate the predictions associated with each.
Open a
tf.GradientTape, and calculate the predictions and loss in that context.Calculate the gradients of the loss with respect to the model variables using the
tf.GradientTape.gradsmethod.Finally, take a step downwards by using the optimizer's
tf.train.Optimizer.apply_gradientsmethod.
model = build_model(
vocab_size = len(vocab),
embedding_dim=embedding_dim,
rnn_units=rnn_units,
batch_size=BATCH_SIZE)
optimizer = tf.keras.optimizers.Adam()
@tf.function
def train_step(inp, target):
with tf.GradientTape() as tape:
predictions = model(inp)
loss = tf.reduce_mean(
tf.keras.losses.sparse_categorical_crossentropy(
target, predictions, from_logits=True))
grads = tape.gradient(loss, model.trainable_variables)
optimizer.apply_gradients(zip(grads, model.trainable_variables))
return loss
# Training step
EPOCHS = 10
for epoch in range(EPOCHS):
start = time.time()
# initializing the hidden state at the start of every epoch
# initally hidden is None
hidden = model.reset_states()
for (batch_n, (inp, target)) in enumerate(dataset):
loss = train_step(inp, target)
if batch_n % 100 == 0:
template = 'Epoch {} Batch {} Loss {}'
print(template.format(epoch+1, batch_n, loss))
# saving (checkpoint) the model every 5 epochs
if (epoch + 1) % 5 == 0:
model.save_weights(checkpoint_prefix.format(epoch=epoch))
print ('Epoch {} Loss {:.4f}'.format(epoch+1, loss))
print ('Time taken for 1 epoch {} sec\n'.format(time.time() - start))
model.save_weights(checkpoint_prefix.format(epoch=epoch))
Epoch 1 Batch 0 Loss 4.17379903793335 Epoch 1 Batch 100 Loss 2.3642020225524902 Epoch 1 Loss 2.1451 Time taken for 1 epoch 6.280483961105347 sec Epoch 2 Batch 0 Loss 2.144087791442871 Epoch 2 Batch 100 Loss 1.9279800653457642 Epoch 2 Loss 1.7839 Time taken for 1 epoch 5.3385679721832275 sec Epoch 3 Batch 0 Loss 1.797566533088684 Epoch 3 Batch 100 Loss 1.7153781652450562 Epoch 3 Loss 1.6565 Time taken for 1 epoch 5.347963571548462 sec Epoch 4 Batch 0 Loss 1.5770742893218994 Epoch 4 Batch 100 Loss 1.5482399463653564 Epoch 4 Loss 1.5179 Time taken for 1 epoch 5.357679128646851 sec Epoch 5 Batch 0 Loss 1.506169080734253 Epoch 5 Batch 100 Loss 1.4482747316360474 Epoch 5 Loss 1.4387 Time taken for 1 epoch 5.294915437698364 sec Epoch 6 Batch 0 Loss 1.3886843919754028 Epoch 6 Batch 100 Loss 1.3790349960327148 Epoch 6 Loss 1.3743 Time taken for 1 epoch 5.332662343978882 sec Epoch 7 Batch 0 Loss 1.3449599742889404 Epoch 7 Batch 100 Loss 1.3802368640899658 Epoch 7 Loss 1.3486 Time taken for 1 epoch 5.3653295040130615 sec Epoch 8 Batch 0 Loss 1.2723708152770996 Epoch 8 Batch 100 Loss 1.3609793186187744 Epoch 8 Loss 1.2874 Time taken for 1 epoch 5.3812360763549805 sec Epoch 9 Batch 0 Loss 1.2503230571746826 Epoch 9 Batch 100 Loss 1.2794026136398315 Epoch 9 Loss 1.2940 Time taken for 1 epoch 5.521594762802124 sec Epoch 10 Batch 0 Loss 1.2300844192504883 Epoch 10 Batch 100 Loss 1.2621116638183594 Epoch 10 Loss 1.2470 Time taken for 1 epoch 5.4489147663116455 sec