simple_seq2seq.py 8.1 KB
Newer Older
Y
yangyaming 已提交
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
#   Copyright (c) 2018 PaddlePaddle Authors. All Rights Reserve.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
#     http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.

import numpy as np
16 17 18 19 20 21
import paddle.v2
import paddle.fluid as fluid
import paddle.fluid.core as core
import paddle.fluid.framework as framework
import paddle.fluid.layers as pd
from paddle.fluid.executor import Executor
Y
yangyaming 已提交
22
from beam_search_api import *
Y
yangyaming 已提交
23 24 25

dict_size = 30000
source_dict_dim = target_dict_dim = dict_size
26
src_dict, trg_dict = paddle.v2.dataset.wmt14.get_dict(dict_size)
Y
yangyaming 已提交
27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57
hidden_dim = 32
word_dim = 16
IS_SPARSE = True
batch_size = 2
max_length = 8
topk_size = 50
trg_dic_size = 10000
beam_size = 2

decoder_size = hidden_dim

place = core.CPUPlace()


def encoder():
    # encoder
    src_word_id = pd.data(
        name="src_word_id", shape=[1], dtype='int64', lod_level=1)
    src_embedding = pd.embedding(
        input=src_word_id,
        size=[dict_size, word_dim],
        dtype='float32',
        is_sparse=IS_SPARSE,
        param_attr=fluid.ParamAttr(name='vemb'))

    fc1 = pd.fc(input=src_embedding, size=hidden_dim * 4, act='tanh')
    lstm_hidden0, lstm_0 = pd.dynamic_lstm(input=fc1, size=hidden_dim * 4)
    encoder_out = pd.sequence_last_step(input=lstm_hidden0)
    return encoder_out


58
def decoder_state_cell(context):
Y
yangyaming 已提交
59 60 61
    h = InitState(init=context)
    state_cell = StateCell(
        cell_size=decoder_size, inputs={'x': None}, states={'h': h})
Y
yangyaming 已提交
62 63 64 65 66

    @state_cell.state_updater
    def updater(state_cell):
        current_word = state_cell.get_input('x')
        prev_h = state_cell.get_state('h')
67 68
        # make sure lod of h heritted from prev_h
        h = pd.fc(input=[prev_h, current_word], size=decoder_size, act='tanh')
Y
yangyaming 已提交
69
        state_cell.set_state('h', h)
Y
yangyaming 已提交
70

71 72 73 74
    return state_cell


def decoder_train(state_cell):
Y
yangyaming 已提交
75 76 77 78 79 80 81 82 83 84
    # decoder
    trg_language_word = pd.data(
        name="target_language_word", shape=[1], dtype='int64', lod_level=1)
    trg_embedding = pd.embedding(
        input=trg_language_word,
        size=[dict_size, word_dim],
        dtype='float32',
        is_sparse=IS_SPARSE,
        param_attr=fluid.ParamAttr(name='vemb'))

85
    decoder = TrainingDecoder(state_cell)
Y
yangyaming 已提交
86

87
    with decoder.block():
Y
yangyaming 已提交
88 89
        current_word = decoder.step_input(trg_embedding)
        decoder.state_cell.compute_state(inputs={'x': current_word})
90
        current_score = pd.fc(input=decoder.state_cell.get_state('h'),
Y
yangyaming 已提交
91 92
                              size=target_dict_dim,
                              act='softmax')
93
        decoder.state_cell.update_states()
Y
yangyaming 已提交
94
        decoder.output(current_score)
Y
yangyaming 已提交
95 96 97 98

    return decoder()


99
def decoder_decode(state_cell):
Y
yangyaming 已提交
100 101 102 103 104
    init_ids = pd.data(name="init_ids", shape=[1], dtype="int64", lod_level=2)
    init_scores = pd.data(
        name="init_scores", shape=[1], dtype="float32", lod_level=2)

    def embedding(input):
105
        return pd.embedding(
Y
yangyaming 已提交
106
            input=input,
107
            size=[dict_size, word_dim],
Y
yangyaming 已提交
108
            dtype='float32',
109 110
            is_sparse=IS_SPARSE,
            param_attr=fluid.ParamAttr('vemb'))
Y
yangyaming 已提交
111

112 113 114 115 116 117 118 119 120 121 122
    decoder = BeamSearchDecoder(state_cell, max_len=max_length)

    with decoder.block():
        prev_ids = decoder.read_array(init=init_ids, is_ids=True)
        prev_scores = decoder.read_array(init=init_scores, is_scores=True)
        prev_ids_embedding = embedding(prev_ids)
        prev_state = decoder.state_cell.get_state('h')
        prev_state_expanded = pd.sequence_expand(prev_state, prev_scores)
        decoder.state_cell.set_state('h', prev_state_expanded)
        decoder.state_cell.compute_state(inputs={'x': prev_ids_embedding})
        current_state = decoder.state_cell.get_state('h')
123
        # copy lod from prev_ids to current_state
124 125 126 127 128 129 130
        scores = pd.fc(input=current_state, size=target_dict_dim, act='softmax')
        topk_scores, topk_indices = pd.topk(scores, k=50)
        selected_ids, selected_scores = pd.beam_search(
            prev_ids, topk_indices, topk_scores, beam_size, end_id=10, level=0)
        decoder.state_cell.update_states()
        decoder.update_array(prev_ids, selected_ids)
        decoder.update_array(prev_scores, selected_scores)
Y
yangyaming 已提交
131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160

    translation_ids, translation_scores = decoder()

    return translation_ids, translation_scores


def set_init_lod(data, lod, place):
    res = core.LoDTensor()
    res.set(data, place)
    res.set_lod(lod)
    return res


def to_lodtensor(data, place):
    seq_lens = [len(seq) for seq in data]
    cur_len = 0
    lod = [cur_len]
    for l in seq_lens:
        cur_len += l
        lod.append(cur_len)
    flattened_data = np.concatenate(data, axis=0).astype("int64")
    flattened_data = flattened_data.reshape([len(flattened_data), 1])
    res = core.LoDTensor()
    res.set(flattened_data, place)
    res.set_lod([lod])
    return res


def train_main():
    context = encoder()
161 162
    state_cell = decoder_state_cell(context)
    rnn_out = decoder_train(state_cell)
Y
yangyaming 已提交
163 164 165 166 167 168 169 170
    label = pd.data(
        name="target_language_next_word", shape=[1], dtype='int64', lod_level=1)
    cost = pd.cross_entropy(input=rnn_out, label=label)
    avg_cost = pd.mean(x=cost)

    optimizer = fluid.optimizer.Adagrad(learning_rate=1e-4)
    optimizer.minimize(avg_cost)

171 172 173
    train_data = paddle.v2.batch(
        paddle.v2.reader.shuffle(
            paddle.v2.dataset.wmt14.train(dict_size), buf_size=1000),
Y
yangyaming 已提交
174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195
        batch_size=batch_size)

    exe = Executor(place)

    exe.run(framework.default_startup_program())

    batch_id = 0
    for pass_id in xrange(1):
        for data in train_data():
            word_data = to_lodtensor(map(lambda x: x[0], data), place)
            trg_word = to_lodtensor(map(lambda x: x[1], data), place)
            trg_word_next = to_lodtensor(map(lambda x: x[2], data), place)
            outs = exe.run(framework.default_main_program(),
                           feed={
                               'src_word_id': word_data,
                               'target_language_word': trg_word,
                               'target_language_next_word': trg_word_next
                           },
                           fetch_list=[avg_cost])
            avg_cost_val = np.array(outs[0])
            print('pass_id=' + str(pass_id) + ' batch=' + str(batch_id) +
                  " avg_cost=" + str(avg_cost_val))
196
            if batch_id > 3000:
Y
yangyaming 已提交
197 198 199 200 201 202
                break
            batch_id += 1


def decode_main():
    context = encoder()
203 204
    state_cell = decoder_state_cell(context)
    translation_ids, translation_scores = decoder_decode(state_cell)
Y
yangyaming 已提交
205 206 207 208 209 210 211 212 213 214 215 216

    exe = Executor(place)
    exe.run(framework.default_startup_program())

    init_ids_data = np.array([1 for _ in range(batch_size)], dtype='int64')
    init_scores_data = np.array(
        [1. for _ in range(batch_size)], dtype='float32')
    init_ids_data = init_ids_data.reshape((batch_size, 1))
    init_scores_data = init_scores_data.reshape((batch_size, 1))
    init_lod = [i for i in range(batch_size)] + [batch_size]
    init_lod = [init_lod, init_lod]

217 218 219
    train_data = paddle.v2.batch(
        paddle.v2.reader.shuffle(
            paddle.v2.dataset.wmt14.train(dict_size), buf_size=1000),
Y
yangyaming 已提交
220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236
        batch_size=batch_size)
    for _, data in enumerate(train_data()):
        init_ids = set_init_lod(init_ids_data, init_lod, place)
        init_scores = set_init_lod(init_scores_data, init_lod, place)

        src_word_data = to_lodtensor(map(lambda x: x[0], data), place)

        result_ids, result_scores = exe.run(
            framework.default_main_program(),
            feed={
                'src_word_id': src_word_data,
                'init_ids': init_ids,
                'init_scores': init_scores
            },
            fetch_list=[translation_ids, translation_scores],
            return_numpy=False)
        print result_ids.lod()
237
        #break
Y
yangyaming 已提交
238 239 240


if __name__ == '__main__':
241 242
    #train_main()
    decode_main()