-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathmodel.py
194 lines (146 loc) · 7.31 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
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
import torch
import torch.nn as nn
import torch.nn.functional as F
import torch.nn.utils.rnn as rnn
from torch.autograd import Variable
from torch_geometric.nn import GCNConv
class PHVModel(nn.Module):
def __init__(self, max_word, word_embedding_size=32, lstm_hidden_size=32, lstm_layers=3, hidden_size=32):
super(PHVModel, self).__init__()
self.embedding_layer = nn.Embedding(max_word, word_embedding_size)
self.lstm_layer = nn.LSTM(word_embedding_size, lstm_hidden_size, batch_first=True, bidirectional=True, num_layers=lstm_layers)
self.linear_layer = nn.Linear(lstm_hidden_size * 2, hidden_size)
self.output_layer = nn.Sequential(
nn.Linear(hidden_size * 2, hidden_size),
nn.ReLU(),
nn.Linear(hidden_size, 2)
)
def forward(self, seq1, seq2, len1, len2):
seq1_embedding = self.embedding_layer(seq1)
seq2_embedding = self.embedding_layer(seq2)
seq1_embedding = rnn.pack_padded_sequence(seq1_embedding, len1, batch_first=True, enforce_sorted=False)
seq2_embedding = rnn.pack_padded_sequence(seq2_embedding, len2, batch_first=True, enforce_sorted=False)
seq1_embedding, _ = self.lstm_layer(seq1_embedding)
seq2_embedding, _ = self.lstm_layer(seq2_embedding)
seq1_lstm_out, lens_unpacked1 = rnn.pad_packed_sequence(seq1_embedding, batch_first=True)
seq2_lstm_out, lens_unpacked2 = rnn.pad_packed_sequence(seq2_embedding, batch_first=True)
seq1_lstm_out = torch.cat([seq1_lstm_out[i:i+1, lens_unpacked1[i]-1, :] for i in range(lens_unpacked1.size()[0])], dim=0)
seq2_lstm_out = torch.cat([seq2_lstm_out[i:i+1, lens_unpacked2[i]-1, :] for i in range(lens_unpacked2.size()[0])], dim=0)
seq1_lstm_out = seq1_lstm_out.view(seq1_lstm_out.size()[0], -1)
seq2_lstm_out = seq2_lstm_out.view(seq2_lstm_out.size()[0], -1)
seq1_embedded = self.linear_layer(seq1_lstm_out)
seq2_embedded = self.linear_layer(seq2_lstm_out)
out = torch.cat([seq1_embedded, seq2_embedded], dim=1)
out = self.output_layer(out)
return out
class PHVGNNModel(nn.Module):
def __init__(self, max_word, word_embedding_size=32, lstm_hidden_size=32, lstm_layers=3, hidden_size=32):
super(PHVGNNModel, self).__init__()
self.embedding_layer = nn.Embedding(max_word, word_embedding_size)
self.lstm_layer = nn.LSTM(word_embedding_size, lstm_hidden_size, batch_first=True, bidirectional=True, num_layers=lstm_layers)
self.linear_layer = nn.Linear(lstm_hidden_size * 2, hidden_size)
self.conv1 = GCNConv(hidden_size, 128)
self.conv2 = GCNConv(128, hidden_size)
self.output_layer = nn.Sequential(
nn.Linear(hidden_size * 2, hidden_size),
nn.ReLU(),
nn.Linear(hidden_size, 2)
)
def forward(self, x, pos_edge_index, neg_edge_index, seq_len):
seq_embedding = self.embedding_layer(x)
seq_embedding = rnn.pack_padded_sequence(seq_embedding, seq_len, batch_first=True, enforce_sorted=False)
seq_embedding, _ = self.lstm_layer(seq_embedding)
seq_lstm_out, lens_unpacked = rnn.pad_packed_sequence(seq_embedding, batch_first=True)
seq_lstm_out = torch.cat([seq_lstm_out[i:i+1, lens_unpacked[i]-1, :] for i in range(lens_unpacked.size()[0])], dim=0)
seq_lstm_out = seq_lstm_out.view(seq_lstm_out.size()[0], -1)
seq_embedded = self.linear_layer(seq_lstm_out)
seq_embedded = F.relu(self.conv1(seq_embedded, pos_edge_index))
seq_embedded = F.relu(self.conv2(seq_embedded, neg_edge_index))
edge_index = torch.cat([pos_edge_index,neg_edge_index], dim=-1)
out = torch.cat([seq_embedded[edge_index[0]], seq_embedded[edge_index[1]]], dim=1)
out = self.output_layer(out)
return out
class FocalLoss(nn.Module):
r"""
This criterion is a implemenation of Focal Loss, which is proposed in
Focal Loss for Dense Object Detection.
Loss(x, class) = - \alpha (1-softmax(x)[class])^gamma \log(softmax(x)[class])
The losses are averaged across observations for each minibatch.
Args:
alpha(1D Tensor, Variable) : the scalar factor for this criterion
gamma(float, double) : gamma > 0; reduces the relative loss for well-classified examples (p > .5),
putting more focus on hard, misclassified examples
size_average(bool): By default, the losses are averaged over observations for each minibatch.
However, if the field size_average is set to False, the losses are
instead summed for each minibatch.
"""
def __init__(self, class_num, alpha=None, gamma=2, size_average=True):
super(FocalLoss, self).__init__()
if alpha is None:
self.alpha = Variable(torch.ones(class_num, 1))
else:
if isinstance(alpha, Variable):
self.alpha = alpha
else:
self.alpha = Variable(alpha)
self.gamma = gamma
self.class_num = class_num
self.size_average = size_average
def forward(self, inputs, targets):
N = inputs.size(0)
C = inputs.size(1)
P = F.softmax(inputs)
class_mask = inputs.data.new(N, C).fill_(0)
class_mask = Variable(class_mask)
ids = targets.view(-1, 1)
class_mask.scatter_(1, ids.data, 1.)
if inputs.is_cuda and not self.alpha.is_cuda:
self.alpha = self.alpha.cuda()
alpha = self.alpha[ids.data.view(-1)]
probs = (P*class_mask).sum(1).view(-1,1)
log_p = probs.log()
batch_loss = -alpha*(torch.pow((1-probs), self.gamma))*log_p
if self.size_average:
loss = batch_loss.mean()
else:
loss = batch_loss.sum()
return loss
# from torch.utils.data import Dataset
# seq1 = [torch.tensor([1, 2, 3, 4, 5, 6, 7]),
# torch.tensor([2, 3, 4, 5, 6, 7]),
# torch.tensor([3, 4, 5, 6, 7]),
# torch.tensor([4, 5, 6, 7]),
# torch.tensor([5, 6, 7]),
# torch.tensor([6, 7]),
# torch.tensor([7])]
# class MyData(Dataset):
# def __init__(self, seq1, seq2):
# self.seq1 = seq1
# self.seq2 = seq2
# def __len__(self):
# return len(self.seq1)
# def __getitem__(self, item):
# return self.seq1[item], self.seq2[item]
# def collate_fn(train_data):
# train_data.sort(key=lambda data: len(data), reverse=True)
# data_length1 = [len(data[0]) for data in train_data]
# data_length2 = [len(data[1]) for data in train_data]
# seq1 = []
# seq2 = []
# for data in train_data:
# seq1.append(data[0])
# seq2.append(data[1])
# seq1 = rnn.pad_sequence(seq1, batch_first=True)
# seq2 = rnn.pad_sequence(seq2, batch_first=True)
# return seq1, seq2, torch.tensor(data_length1), torch.tensor(data_length2)
# train_data = MyData(seq1, seq1)
# train_loader = torch.utils.data.DataLoader(train_data, batch_size=2, shuffle=True, collate_fn=collate_fn)
# model = PHVModel(max_word=10)
# for data in train_loader:
# seq1, seq2, len1, len2 = data
# print(seq1.size())
# print(seq2.size())
# print(len1.size())
# print(len2.size())
# print(model(seq1, seq2, len1, len2))
# break