forked from xuwenshen/Machine-Translation
-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathnet.py
157 lines (107 loc) · 6.07 KB
/
net.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
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
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
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
import os
import argparse
import logging
import torch
from torch.optim.lr_scheduler import StepLR
import torchtext
import seq2seq
from seq2seq.trainer import SupervisedTrainer
from seq2seq.models import EncoderRNN, DecoderRNN, Seq2seq, TopKDecoder
from seq2seq.loss import Perplexity
from seq2seq.optim import Optimizer
from seq2seq.dataset import SourceField, TargetField
from seq2seq.evaluator import Predictor
from seq2seq.util.checkpoint import Checkpoint
import numpy as np
import torch
from torch import nn
from torch import autograd
from torch.autograd import Variable
import json
import random
import torch.nn.utils.rnn as rnn_utils
from Transform import Transform
from Transform import Transform
transform = Transform(zh_voc_path='/data/xuwenshen/ai_challenge/data/train/train/zh_voc.json',
en_voc_path='/data/xuwenshen/ai_challenge/data/train/train/en_voc.json')
weight = [1 for i in range(len(transform.zh_voc))]
weight[transform.zh_pad_id] = 0
class EncLSTM(nn.Module):
def __init__(self, dropout_p, en_hidden, en_dims, enc_layers):
super(EncLSTM, self).__init__()
self.en_hidden = en_hidden
self.enc_lstm = torch.nn.LSTM(input_size=en_dims,
num_layers=enc_layers,
hidden_size=en_hidden,
dropout=dropout_p,
bidirectional=True,
batch_first=True)
def forward(self, inputs):
outputs, states = self.enc_lstm(inputs)
return outputs, states
class Dec(nn.Module):
def __init__(self, zh_max_len, zh_hidden, dec_layers, input_dropout_p, dropout_p, beam_size, zh_embedding_size):
super(Dec, self).__init__()
self.dec_rnn = DecoderRNN(vocab_size = len(transform.zh_voc),
max_len = zh_max_len,
embedding_size = zh_embedding_size,
hidden_size = zh_hidden,
sos_id = transform.zh_go_id,
eos_id = transform.zh_eos_id,
n_layers = dec_layers,
rnn_cell='lstm',
bidirectional=True,
input_dropout_p = input_dropout_p,
dropout_p=dropout_p,
use_attention=True)
self.beam_dec = TopKDecoder(self.dec_rnn, beam_size)
def forward(self, gtruths, encoder_hidden, encoder_outputs, teacher_forcing_ratio, is_train):
if is_train:
if teacher_forcing_ratio > 0:
gtruths = Variable(gtruths.long()).cuda()
decoder_outputs, decoder_hidden, ret_dict = self.dec_rnn(inputs = gtruths,
encoder_hidden = encoder_hidden,
encoder_outputs = encoder_outputs,
teacher_forcing_ratio = teacher_forcing_ratio)
else:
decoder_outputs, decoder_hidden, ret_dict = self.beam_dec(inputs = gtruths,
encoder_hidden = encoder_hidden,
encoder_outputs = encoder_outputs,
teacher_forcing_ratio = teacher_forcing_ratio)
return decoder_outputs, decoder_hidden, ret_dict
class Seq2Seq(nn.Module):
def __init__(self, en_dims, zh_dims, input_dropout_p, dropout_p, en_hidden, zh_hidden, enc_layers, dec_layers, en_max_len, zh_max_len, beam_size):
super(Seq2Seq, self).__init__()
self.weight = torch.Tensor(weight)
self.input_dropout = nn.Dropout(input_dropout_p)
self.en_embedding = torch.nn.Embedding(num_embeddings=len(transform.en_voc), embedding_dim=en_dims)
self.cost_func = nn.NLLLoss(weight=self.weight)
self.rnn_net = EncLSTM(dropout_p=dropout_p, en_dims=en_dims, en_hidden=en_hidden, enc_layers=enc_layers)
self.dec_net = Dec(zh_max_len=zh_max_len,
zh_hidden=zh_hidden,
zh_embedding_size=zh_dims,
dec_layers=dec_layers,
input_dropout_p=input_dropout_p,
dropout_p=dropout_p,
beam_size = beam_size)
def forward(self, inputs, gtruths, is_train, teacher_forcing_ratio):
inputs = Variable(inputs).long().cuda()
inputs = self.en_embedding(inputs)
inputs = self.input_dropout(inputs)
rnn_enc, enc_hidden = self.rnn_net(inputs)
decoder_outputs, decoder_hidden, ret_dict = self.dec_net(gtruths=gtruths,
encoder_hidden=enc_hidden,
encoder_outputs=rnn_enc,
teacher_forcing_ratio=teacher_forcing_ratio,
is_train=is_train)
return decoder_outputs, ret_dict
def get_loss(self, logits, labels):
labels = Variable(labels).long().cuda()
labels = labels.transpose(0, 1)
for i in range(len(logits)):
logits[i] = logits[i].contiguous().view(1, logits[i].size(0), logits[i].size(1))
logits = torch.cat(logits)
logits = logits.contiguous().view(-1, logits.size(-1))
labels = labels.contiguous().view(-1)
loss = torch.mean(self.cost_func(logits, labels))
return loss