Spaces:
Running
Running
Upload event_detection_model.py
Browse files- event_detection_model.py +29 -0
event_detection_model.py
ADDED
@@ -0,0 +1,29 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
import torch
|
2 |
+
from transformers import AutoModel
|
3 |
+
from transformers import AutoModelForMaskedLM
|
4 |
+
|
5 |
+
class DistillBERTClass(torch.nn.Module):
|
6 |
+
def __init__(self, checkpoint_model):
|
7 |
+
#the super class is not important here!
|
8 |
+
super(DistillBERTClass, self).__init__()
|
9 |
+
#check the rmodel used here !
|
10 |
+
self.pre_trained_model = AutoModelForMaskedLM.from_pretrained(checkpoint_model,output_hidden_states=True)
|
11 |
+
self.linear = torch.nn.Linear(768, 768)
|
12 |
+
self.relu = torch.nn.ReLU()
|
13 |
+
self.dropout = torch.nn.Dropout(0.3)
|
14 |
+
self.classifier = torch.nn.Linear(768, 12)
|
15 |
+
|
16 |
+
def forward(self, input_ids, attention_mask):
|
17 |
+
pre_trained_output = self.pre_trained_model(input_ids=input_ids, attention_mask=attention_mask)
|
18 |
+
hidden_state = pre_trained_output.hidden_states[-1]
|
19 |
+
|
20 |
+
hidden_state = hidden_state[:, 0, :]
|
21 |
+
output = self.linear(hidden_state)
|
22 |
+
output = self.relu(output)
|
23 |
+
output = self.dropout(output)
|
24 |
+
output = self.classifier(output)
|
25 |
+
return output
|
26 |
+
|
27 |
+
|
28 |
+
|
29 |
+
|