-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathmodel.py
140 lines (110 loc) · 7.34 KB
/
model.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
# encoding=utf-8
import torch
import torch.nn as nn
import logging
import numpy as np
from net.LSTMEncoder import LSTMEncoder
from net.BertEncoder import WordBertEncoder
from net.Attention import Attention
from net.TextCNNEncoder import TextCNNEncoder
class Model(nn.Module):
def __init__(self, vocab, bert_path, max_sent_len, dropout,
sent_rep_size, sent_hidden_size, sent_num_layers, use_cuda, device):
super(Model, self).__init__()
self.sent_rep_size = sent_rep_size # 256
self.doc_rep_size = sent_hidden_size * 2
self.all_parameters = {}
parameters = []
self.word_encoder = WordBertEncoder(bert_path, max_sent_len, dropout)
bert_parameters = self.word_encoder.get_bert_parameters()
self.sent_encoder = LSTMEncoder(self.sent_rep_size, sent_hidden_size, sent_num_layers, dropout)
self.sent_attention = Attention(self.doc_rep_size)
parameters.extend(list(filter(lambda p: p.requires_grad, self.sent_encoder.parameters())))
parameters.extend(list(filter(lambda p: p.requires_grad, self.sent_attention.parameters())))
self.out = nn.Linear(self.doc_rep_size, vocab.label_size, bias=True)
parameters.extend(list(filter(lambda p: p.requires_grad, self.out.parameters())))
if torch.cuda.is_available():
device = torch.device("cuda")
self.to(device)
else:
device = torch.device("cpu")
self.to(device)
if len(parameters) > 0:
self.all_parameters["basic_parameters"] = parameters
self.all_parameters["bert_parameters"] = bert_parameters
logging.info('Build model with bert word encoder, lstm sent encoder.')
para_num = sum([np.prod(list(p.size())) for p in self.parameters()]) # 运行到这后，cuda内存拉爆了，把别的代码弄崩了
logging.info('Model param num: %.2f M.' % (para_num / 1e6))
def forward(self, batch_inputs):
# batch_inputs(batch_inputs1, batch_inputs2): b x doc_len x sent_len
# batch_masks : b x doc_len x sent_len
batch_inputs1, batch_inputs2, batch_masks = batch_inputs
batch_size, max_doc_len, max_sent_len = batch_inputs1.shape[0], batch_inputs1.shape[1], batch_inputs1.shape[2]
batch_inputs1 = batch_inputs1.view(batch_size * max_doc_len, max_sent_len) # sen_num x sent_len
batch_inputs2 = batch_inputs2.view(batch_size * max_doc_len, max_sent_len) # sen_num x sent_len
batch_masks = batch_masks.view(batch_size * max_doc_len, max_sent_len) # sen_num x sent_len
sent_reps = self.word_encoder(batch_inputs1, batch_inputs2) # sen_num x sent_rep_size
sent_reps = sent_reps.view(batch_size, max_doc_len, self.sent_rep_size) # b x doc_len x sent_rep_size
batch_masks = batch_masks.view(batch_size, max_doc_len, max_sent_len) # b x doc_len x max_sent_len
sent_masks = batch_masks.bool().any(2).float() # b x doc_len
sent_hiddens = self.sent_encoder(sent_reps, sent_masks) # b x doc_len x doc_rep_size
doc_reps, atten_scores = self.sent_attention(sent_hiddens, sent_masks) # b x doc_rep_size
batch_outputs = self.out(doc_reps) # b x num_labels
return batch_outputs
class TextCNNModel(nn.Module):
def __init__(self, vocab, bert_path, max_sent_len, dropout,
sent_rep_size, sent_hidden_size, sent_num_layers, use_cuda, device):
super(TextCNNModel, self).__init__()
self.sent_rep_size = 768 # 256
self.doc_rep_size = sent_hidden_size * 2
self.all_parameters = {}
parameters = []
self.word_encoder = WordBertEncoder(bert_path, max_sent_len, dropout)
bert_parameters = self.word_encoder.get_bert_parameters()
self.textcnn_encoder = TextCNNEncoder(dropout)
parameters.extend(list(filter(lambda p: p.requires_grad, self.textcnn_encoder.parameters())))
self.sent_encoder = LSTMEncoder(self.sent_rep_size, sent_hidden_size, sent_num_layers, dropout)
self.sent_attention = Attention(self.sent_rep_size) # Attention(self.doc_rep_size) LSTM是doc_rep_size
parameters.extend(list(filter(lambda p: p.requires_grad, self.sent_encoder.parameters())))
parameters.extend(list(filter(lambda p: p.requires_grad, self.sent_attention.parameters())))
self.out = nn.Linear(self.sent_rep_size, vocab.label_size, bias=True) # self.doc_rep_size (LSTM)
parameters.extend(list(filter(lambda p: p.requires_grad, self.out.parameters())))
if use_cuda:
self.to(device)
if len(parameters) > 0:
self.all_parameters["basic_parameters"] = parameters
self.all_parameters["bert_parameters"] = bert_parameters
logging.info('Build model with bert word encoder, lstm sent encoder.')
para_num = sum([np.prod(list(p.size())) for p in self.parameters()])
logging.info('Model param num: %.2f M.' % (para_num / 1e6))
def forward(self, batch_inputs):
# batch_inputs(batch_inputs1, batch_inputs2): b x doc_len x sent_len
# batch_masks : b x doc_len x sent_len
batch_inputs1, batch_inputs2, batch_masks = batch_inputs
batch_size, max_doc_len, max_sent_len = batch_inputs1.shape[0], batch_inputs1.shape[1], batch_inputs1.shape[2]
# shape:[batch_size_ , seq_length]
batch_inputs1 = batch_inputs1.view(batch_size * max_doc_len, max_sent_len) # sen_num x sent_len
batch_inputs2 = batch_inputs2.view(batch_size * max_doc_len, max_sent_len) # sen_num x sent_len
batch_masks = batch_masks.view(batch_size * max_doc_len, max_sent_len) # sen_num x sent_len
# sent_reps = self.word_encoder(batch_inputs1, batch_inputs2) # sen_num x sent_rep_size 当pooled=True
sent_reps = self.word_encoder(batch_inputs1, batch_inputs2) # sen_num x seq_len x sent_rep_size
# print("sent_reps1:{}".format(sent_reps.shape)) # sent_reps1:torch.Size([1, 281, 768])
sent_reps = self.textcnn_encoder(sent_reps)
# print("sent_reps2:{}".format(sent_reps.shape)) # sent_reps2:torch.Size([1, 768])
sent_reps = sent_reps.view(batch_size, max_doc_len, self.sent_rep_size) # b x doc_len x sent_rep_size
batch_masks = batch_masks.view(batch_size, max_doc_len, max_sent_len) # b x doc_len x max_sent_len
sent_masks = batch_masks.bool().any(2).float() # b x doc_len
# 下面这三行是使用了LSTM来作为句子的编码方式
# sent_hiddens = self.sent_encoder(sent_reps, sent_masks) # b x doc_len x doc_rep_size
# doc_reps, atten_scores = self.sent_attention(sent_hiddens, sent_masks) # b x doc_rep_size
# batch_outputs = self.out(doc_reps) # b x num_labels
# 下面这三行是使用了TextCNN来作为句子的编码方式
# sen_num, sent_len, embedding_size TextCNN输入需要的格式
# sent_hiddens = self.textcnn_encoder(sent_reps) # b x doc_len x doc_rep_size
# 上面这个textcnn貌似还缺乏一个mask
doc_reps, atten_scores = self.sent_attention(sent_reps, sent_masks) # b x doc_rep_size
batch_outputs = self.out(doc_reps) # b x num_labels
# 下面这两行是直接使用bert后的向量进行attention（暂未成功）
# doc_reps, atten_scores = self.sent_attention(sent_reps, sent_masks) # b x doc_rep_size
# batch_outputs = self.out(doc_reps) # b x num_labels
return batch_outputs