Update model.py
Browse files
model.py
CHANGED
@@ -0,0 +1,32 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
import torch.nn as nn
|
2 |
+
import torch
|
3 |
+
|
4 |
+
class LSTMClassifier(nn.Module):
|
5 |
+
def __init__(self, input_size=1, hidden_size=32, num_layers=1,
|
6 |
+
bidirectional=True, dropout=0.0, num_classes=2):
|
7 |
+
super(LSTMClassifier, self).__init__()
|
8 |
+
self.hidden_size = hidden_size
|
9 |
+
self.num_layers = num_layers
|
10 |
+
self.bidirectional = bidirectional
|
11 |
+
|
12 |
+
self.lstm = nn.LSTM(
|
13 |
+
input_size=input_size,
|
14 |
+
hidden_size=hidden_size,
|
15 |
+
num_layers=num_layers,
|
16 |
+
batch_first=True,
|
17 |
+
dropout=dropout if num_layers > 1 else 0.0,
|
18 |
+
bidirectional=bidirectional
|
19 |
+
)
|
20 |
+
|
21 |
+
direction_factor = 2 if bidirectional else 1
|
22 |
+
self.fc = nn.Linear(hidden_size * direction_factor, num_classes)
|
23 |
+
|
24 |
+
def forward(self, x):
|
25 |
+
_, (hn, _) = self.lstm(x)
|
26 |
+
if self.bidirectional:
|
27 |
+
forward = hn[-2]
|
28 |
+
backward = hn[-1]
|
29 |
+
combined = torch.cat((forward, backward), dim=1)
|
30 |
+
else:
|
31 |
+
combined = hn[-1]
|
32 |
+
return self.fc(combined)
|