kevinjesse
commited on
Commit
·
646aebf
1
Parent(s):
bb0def0
Upload 3 files
Browse files- RefactorModels.py +47 -0
- pytorch_model_RP.bin +3 -0
- pytorch_model_RSP.bin +3 -0
RefactorModels.py
ADDED
@@ -0,0 +1,47 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
import torch.nn as nn
|
2 |
+
import torch
|
3 |
+
from transformers import AutoModel, AutoConfig
|
4 |
+
|
5 |
+
class RefactorSpanModel(nn.Module):
|
6 |
+
def __init__(self):
|
7 |
+
super().__init__()
|
8 |
+
base_model_path = 'microsoft/codebert-base'
|
9 |
+
self.base_config = AutoConfig.from_pretrained(base_model_path)
|
10 |
+
self.base_model = AutoModel.from_config(self.base_config)
|
11 |
+
self.dropout = nn.Dropout(0.5)
|
12 |
+
self.classifier = nn.Linear(768, 1)
|
13 |
+
self.start_span = nn.Linear(768, 1)
|
14 |
+
|
15 |
+
def forward(self, input_ids):
|
16 |
+
outputs = self.base_model(input_ids)
|
17 |
+
outputs_pool = self.dropout(outputs[1]) #use pooler output...
|
18 |
+
outputs_hidden = self.dropout(outputs[0])
|
19 |
+
refactor = self.classifier(outputs_pool)
|
20 |
+
span = self.start_span(outputs_hidden)
|
21 |
+
return refactor, span
|
22 |
+
|
23 |
+
class RefactorModel(nn.Module):
|
24 |
+
def __init__(self):
|
25 |
+
super().__init__()
|
26 |
+
base_model_path = 'microsoft/codebert-base'
|
27 |
+
self.base_config = AutoConfig.from_pretrained(base_model_path)
|
28 |
+
self.base_model = AutoModel.from_config(self.base_config)
|
29 |
+
self.dropout = nn.Dropout(0.5)
|
30 |
+
self.classifier = nn.Linear(768, 1)
|
31 |
+
|
32 |
+
def forward(self, input_ids):
|
33 |
+
outputs = self.base_model(input_ids)
|
34 |
+
outputs_pool = self.dropout(outputs[1]) #use pooler output...
|
35 |
+
refactor = self.classifier(outputs_pool)
|
36 |
+
return refactor
|
37 |
+
|
38 |
+
if __name__ == "__main__":
|
39 |
+
checkpoint = 'pytorch_model_RSP.bin'
|
40 |
+
model = RefactorSpanModel()
|
41 |
+
model.load_state_dict(torch.load(checkpoint), strict=True)
|
42 |
+
#print(model.base_model.embeddings.word_embeddings.weight)
|
43 |
+
|
44 |
+
checkpoint = 'pytorch_model_RP.bin'
|
45 |
+
model = RefactorModel()
|
46 |
+
model.load_state_dict(torch.load(checkpoint), strict=True)
|
47 |
+
#print(model.base_model.embeddings.word_embeddings.weight)
|
pytorch_model_RP.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:b2bf67df05f4bed987db3343a44d1c6a95eae260e557e8bfb3e05fffaf40ccc8
|
3 |
+
size 498678473
|
pytorch_model_RSP.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:a1b310c8c8915cf8e5d3c81dd47a1acccb93455202cc5de7a7141e62441cc649
|
3 |
+
size 498682203
|