-
Notifications
You must be signed in to change notification settings - Fork 526
/
aen.py
127 lines (103 loc) · 5.2 KB
/
aen.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
# -*- coding: utf-8 -*-
# file: aen.py
# author: songyouwei <[email protected]>
# Copyright (C) 2018. All Rights Reserved.
from layers.dynamic_rnn import DynamicLSTM
from layers.squeeze_embedding import SqueezeEmbedding
from layers.attention import Attention, NoQueryAttention
from layers.point_wise_feed_forward import PositionwiseFeedForward
import torch
import torch.nn as nn
import torch.nn.functional as F
# CrossEntropyLoss for Label Smoothing Regularization
class CrossEntropyLoss_LSR(nn.Module):
def __init__(self, device, para_LSR=0.2):
super(CrossEntropyLoss_LSR, self).__init__()
self.para_LSR = para_LSR
self.device = device
self.logSoftmax = nn.LogSoftmax(dim=-1)
def _toOneHot_smooth(self, label, batchsize, classes):
prob = self.para_LSR * 1.0 / classes
one_hot_label = torch.zeros(batchsize, classes) + prob
for i in range(batchsize):
index = label[i]
one_hot_label[i, index] += (1.0 - self.para_LSR)
return one_hot_label
def forward(self, pre, label, size_average=True):
b, c = pre.size()
one_hot_label = self._toOneHot_smooth(label, b, c).to(self.device)
loss = torch.sum(-one_hot_label * self.logSoftmax(pre), dim=1)
if size_average:
return torch.mean(loss)
else:
return torch.sum(loss)
'''
We hereby focus on the bert version.
class AEN_GloVe(nn.Module):
def __init__(self, embedding_matrix, opt):
super(AEN_GloVe, self).__init__()
self.opt = opt
self.embed = nn.Embedding.from_pretrained(torch.tensor(embedding_matrix, dtype=torch.float))
self.squeeze_embedding = SqueezeEmbedding()
self.attn_k = Attention(opt.embed_dim, out_dim=opt.hidden_dim, n_head=8, score_function='mlp', dropout=opt.dropout)
self.attn_q = Attention(opt.embed_dim, out_dim=opt.hidden_dim, n_head=8, score_function='mlp', dropout=opt.dropout)
self.ffn_c = PositionwiseFeedForward(opt.hidden_dim, dropout=opt.dropout)
self.ffn_t = PositionwiseFeedForward(opt.hidden_dim, dropout=opt.dropout)
self.attn_s1 = Attention(opt.hidden_dim, n_head=8, score_function='mlp', dropout=opt.dropout)
self.dense = nn.Linear(opt.hidden_dim*3, opt.polarities_dim)
def forward(self, inputs):
text_raw_indices, target_indices = inputs[0], inputs[1]
context_len = torch.sum(text_raw_indices != 0, dim=-1)
target_len = torch.sum(target_indices != 0, dim=-1)
context = self.embed(text_raw_indices)
context = self.squeeze_embedding(context, context_len)
target = self.embed(target_indices)
target = self.squeeze_embedding(target, target_len)
hc, _ = self.attn_k(context, context)
hc = self.ffn_c(hc)
ht, _ = self.attn_q(context, target)
ht = self.ffn_t(ht)
s1, _ = self.attn_s1(hc, ht)
context_len = torch.tensor(context_len, dtype=torch.float).to(self.opt.device)
target_len = torch.tensor(target_len, dtype=torch.float).to(self.opt.device)
hc_mean = torch.div(torch.sum(hc, dim=1), context_len.view(context_len.size(0), 1))
ht_mean = torch.div(torch.sum(ht, dim=1), target_len.view(target_len.size(0), 1))
s1_mean = torch.div(torch.sum(s1, dim=1), context_len.view(context_len.size(0), 1))
x = torch.cat((hc_mean, s1_mean, ht_mean), dim=-1)
out = self.dense(x)
return out
'''
class AEN_BERT(nn.Module):
def __init__(self, bert, opt):
super(AEN_BERT, self).__init__()
self.opt = opt
self.bert = bert
self.squeeze_embedding = SqueezeEmbedding()
self.dropout = nn.Dropout(opt.dropout)
self.attn_k = Attention(opt.bert_dim, out_dim=opt.hidden_dim, n_head=8, score_function='mlp', dropout=opt.dropout)
self.attn_q = Attention(opt.bert_dim, out_dim=opt.hidden_dim, n_head=8, score_function='mlp', dropout=opt.dropout)
self.ffn_c = PositionwiseFeedForward(opt.hidden_dim, dropout=opt.dropout)
self.ffn_t = PositionwiseFeedForward(opt.hidden_dim, dropout=opt.dropout)
self.attn_s1 = Attention(opt.hidden_dim, n_head=8, score_function='mlp', dropout=opt.dropout)
self.dense = nn.Linear(opt.hidden_dim*3, opt.polarities_dim)
def forward(self, inputs):
context, target = inputs[0], inputs[1]
context_len = torch.sum(context != 0, dim=-1)
target_len = torch.sum(target != 0, dim=-1)
context = self.squeeze_embedding(context, context_len)
context, _ = self.bert(context)
context = self.dropout(context)
target = self.squeeze_embedding(target, target_len)
target, _ = self.bert(target)
target = self.dropout(target)
hc, _ = self.attn_k(context, context)
hc = self.ffn_c(hc)
ht, _ = self.attn_q(context, target)
ht = self.ffn_t(ht)
s1, _ = self.attn_s1(hc, ht)
hc_mean = torch.div(torch.sum(hc, dim=1), context_len.unsqueeze(1).float())
ht_mean = torch.div(torch.sum(ht, dim=1), target_len.unsqueeze(1).float())
s1_mean = torch.div(torch.sum(s1, dim=1), context_len.unsqueeze(1).float())
x = torch.cat((hc_mean, s1_mean, ht_mean), dim=-1)
out = self.dense(x)
return out