Sequence-to-sequence example in Keras (character-level).

This script demonstrates how to implement a basic character-level CNN sequence-to-sequence model. We apply it to translating short English sentences into short French sentences, character-by-character. Note that it is fairly unusual to do character-level machine translation, as word-level models are much more common in this domain. This example is for demonstration purposes only.

Summary of the algorithm

• We start with input sequences from a domain (e.g. English sentences) and corresponding target sequences from another domain (e.g. French sentences).
• An encoder CNN encodes the input character sequence.
• A decoder CNN is trained to turn the target sequences into the same sequence but offset by one timestep in the future, a training process called "teacher forcing" in this context. It uses the output from the encoder. Effectively, the decoder learns to generate targets[t+1...] given targets[...t], conditioned on the input sequence.
• In inference mode, when we want to decode unknown input sequences, we:
  • Encode the input sequence.
  • Start with a target sequence of size 1 (just the start-of-sequence character)
  • Feed the input sequence and 1-char target sequence to the decoder to produce predictions for the next character
  • Sample the next character using these predictions (we simply use argmax).
  • Append the sampled character to the target sequence
  • Repeat until we hit the character limit.

Data download

English to French sentence pairs.

Lots of neat sentence pairs datasets.

References

• lstm_seq2seq.py
• https://wanasit.github.io/attention-based-sequence-to-sequence-in-keras.html

from __future__ import print_function

import numpy as np

from keras.layers import Input, Convolution1D, Dot, Dense, Activation, Concatenate
from keras.models import Model

batch_size = 64  # Batch size for training.
epochs = 100  # Number of epochs to train for.
num_samples = 10000  # Number of samples to train on.
# Path to the data txt file on disk.
data_path = 'fra-eng/fra.txt'

# Vectorize the data.
input_texts = []
target_texts = []
input_characters = set()
target_characters = set()
with open(data_path, 'r', encoding='utf-8') as f:
    lines = f.read().split('\n')
for line in lines[: min(num_samples, len(lines) - 1)]:
    input_text, target_text = line.split('\t')
    # We use "tab" as the "start sequence" character
    # for the targets, and "\n" as "end sequence" character.
    target_text = '\t' + target_text + '\n'
    input_texts.append(input_text)
    target_texts.append(target_text)
    for char in input_text:
        if char not in input_characters:
            input_characters.add(char)
    for char in target_text:
        if char not in target_characters:
            target_characters.add(char)

input_characters = sorted(list(input_characters))
target_characters = sorted(list(target_characters))
num_encoder_tokens = len(input_characters)
num_decoder_tokens = len(target_characters)
max_encoder_seq_length = max([len(txt) for txt in input_texts])
max_decoder_seq_length = max([len(txt) for txt in target_texts])

print('Number of samples:', len(input_texts))
print('Number of unique input tokens:', num_encoder_tokens)
print('Number of unique output tokens:', num_decoder_tokens)
print('Max sequence length for inputs:', max_encoder_seq_length)
print('Max sequence length for outputs:', max_decoder_seq_length)

input_token_index = dict(
    [(char, i) for i, char in enumerate(input_characters)])
target_token_index = dict(
    [(char, i) for i, char in enumerate(target_characters)])

encoder_input_data = np.zeros(
    (len(input_texts), max_encoder_seq_length, num_encoder_tokens),
    dtype='float32')
decoder_input_data = np.zeros(
    (len(input_texts), max_decoder_seq_length, num_decoder_tokens),
    dtype='float32')
decoder_target_data = np.zeros(
    (len(input_texts), max_decoder_seq_length, num_decoder_tokens),
    dtype='float32')

for i, (input_text, target_text) in enumerate(zip(input_texts, target_texts)):
    for t, char in enumerate(input_text):
        encoder_input_data[i, t, input_token_index[char]] = 1.
    for t, char in enumerate(target_text):
        # decoder_target_data is ahead of decoder_input_data by one timestep
        decoder_input_data[i, t, target_token_index[char]] = 1.
        if t > 0:
            # decoder_target_data will be ahead by one timestep
            # and will not include the start character.
            decoder_target_data[i, t - 1, target_token_index[char]] = 1.

# Define an input sequence and process it.
encoder_inputs = Input(shape=(None, num_encoder_tokens))
# Encoder
x_encoder = Convolution1D(256, kernel_size=3, activation='relu',
                          padding='causal')(encoder_inputs)
x_encoder = Convolution1D(256, kernel_size=3, activation='relu',
                          padding='causal', dilation_rate=2)(x_encoder)
x_encoder = Convolution1D(256, kernel_size=3, activation='relu',
                          padding='causal', dilation_rate=4)(x_encoder)

decoder_inputs = Input(shape=(None, num_decoder_tokens))
# Decoder
x_decoder = Convolution1D(256, kernel_size=3, activation='relu',
                          padding='causal')(decoder_inputs)
x_decoder = Convolution1D(256, kernel_size=3, activation='relu',
                          padding='causal', dilation_rate=2)(x_decoder)
x_decoder = Convolution1D(256, kernel_size=3, activation='relu',
                          padding='causal', dilation_rate=4)(x_decoder)
# Attention
attention = Dot(axes=[2, 2])([x_decoder, x_encoder])
attention = Activation('softmax')(attention)

context = Dot(axes=[2, 1])([attention, x_encoder])
decoder_combined_context = Concatenate(axis=-1)([context, x_decoder])

decoder_outputs = Convolution1D(64, kernel_size=3, activation='relu',
                                padding='causal')(decoder_combined_context)
decoder_outputs = Convolution1D(64, kernel_size=3, activation='relu',
                                padding='causal')(decoder_outputs)
# Output
decoder_dense = Dense(num_decoder_tokens, activation='softmax')
decoder_outputs = decoder_dense(decoder_outputs)

# Define the model that will turn
# `encoder_input_data` & `decoder_input_data` into `decoder_target_data`
model = Model([encoder_inputs, decoder_inputs], decoder_outputs)
model.summary()

# Run training
model.compile(optimizer='adam', loss='categorical_crossentropy')
model.fit([encoder_input_data, decoder_input_data], decoder_target_data,
          batch_size=batch_size,
          epochs=epochs,
          validation_split=0.2)
# Save model
model.save('cnn_s2s.h5')

# Next: inference mode (sampling).

# Define sampling models
reverse_input_char_index = dict(
    (i, char) for char, i in input_token_index.items())
reverse_target_char_index = dict(
    (i, char) for char, i in target_token_index.items())

nb_examples = 100
in_encoder = encoder_input_data[:nb_examples]
in_decoder = np.zeros(
    (len(input_texts), max_decoder_seq_length, num_decoder_tokens),
    dtype='float32')

in_decoder[:, 0, target_token_index["\t"]] = 1

predict = np.zeros(
    (len(input_texts), max_decoder_seq_length),
    dtype='float32')

for i in range(max_decoder_seq_length - 1):
    predict = model.predict([in_encoder, in_decoder])
    predict = predict.argmax(axis=-1)
    predict_ = predict[:, i].ravel().tolist()
    for j, x in enumerate(predict_):
        in_decoder[j, i + 1, x] = 1

for seq_index in range(nb_examples):
    # Take one sequence (part of the training set)
    # for trying out decoding.
    output_seq = predict[seq_index, :].ravel().tolist()
    decoded = []
    for x in output_seq:
        if reverse_target_char_index[x] == "\n":
            break
        else:
            decoded.append(reverse_target_char_index[x])
    decoded_sentence = "".join(decoded)
    print('-')
    print('Input sentence:', input_texts[seq_index])
    print('Decoded sentence:', decoded_sentence)