Spaces:
Sleeping
Sleeping
Commit
·
48a61f4
1
Parent(s):
0964dff
Update app.py (#3)
Browse files- Update app.py (3f6e3a1d72badba0df3deca3b864d08d9f48a42e)
Co-authored-by: Kim Hyomin <hyomin@users.noreply.huggingface.co>
app.py
CHANGED
|
@@ -20,7 +20,7 @@ def is_false_alarm(code_text):
|
|
| 20 |
code_text = re.sub('(\\\\n)+', '\\n', code_text)
|
| 21 |
|
| 22 |
# 1. CFA-CodeBERTa-small.pt -> CodeBERTa-small-v1 finetunig model
|
| 23 |
-
path = os.getcwd() + '
|
| 24 |
tokenizer = AutoTokenizer.from_pretrained("huggingface/CodeBERTa-small-v1")
|
| 25 |
input_ids = tokenizer.encode(
|
| 26 |
code_text, max_length=512, truncation=True, padding='max_length')
|
|
@@ -32,7 +32,7 @@ def is_false_alarm(code_text):
|
|
| 32 |
# model(input_ids)[0].argmax().detach().cpu().numpy().item()
|
| 33 |
|
| 34 |
# 2. CFA-codebert-c.pt -> codebert-c finetuning model
|
| 35 |
-
path = os.getcwd() + '
|
| 36 |
tokenizer = AutoTokenizer.from_pretrained(path)
|
| 37 |
input_ids = tokenizer(code_text, padding=True, max_length=512,
|
| 38 |
truncation=True, return_token_type_ids=True)['input_ids']
|
|
@@ -43,7 +43,7 @@ def is_false_alarm(code_text):
|
|
| 43 |
pred_2 = model(input_ids)[0].detach().cpu().numpy()[0]
|
| 44 |
|
| 45 |
# 3. CFA-codebert-c-v2.pt -> undersampling + codebert-c finetuning model
|
| 46 |
-
path = os.getcwd() + '
|
| 47 |
tokenizer = RobertaTokenizer.from_pretrained(path)
|
| 48 |
input_ids = tokenizer(code_text, padding=True, max_length=512,
|
| 49 |
truncation=True, return_token_type_ids=True)['input_ids']
|
|
@@ -54,7 +54,7 @@ def is_false_alarm(code_text):
|
|
| 54 |
pred_3 = model(input_ids)[0].detach().cpu().numpy()
|
| 55 |
|
| 56 |
# 4. codeT5 finetuning model
|
| 57 |
-
path = os.getcwd() + '
|
| 58 |
model_params = {
|
| 59 |
# model_type: t5-base/t5-large
|
| 60 |
"MODEL": path,
|
|
|
|
| 20 |
code_text = re.sub('(\\\\n)+', '\\n', code_text)
|
| 21 |
|
| 22 |
# 1. CFA-CodeBERTa-small.pt -> CodeBERTa-small-v1 finetunig model
|
| 23 |
+
path = os.getcwd() + '/models/CFA-CodeBERTa-small.pt'
|
| 24 |
tokenizer = AutoTokenizer.from_pretrained("huggingface/CodeBERTa-small-v1")
|
| 25 |
input_ids = tokenizer.encode(
|
| 26 |
code_text, max_length=512, truncation=True, padding='max_length')
|
|
|
|
| 32 |
# model(input_ids)[0].argmax().detach().cpu().numpy().item()
|
| 33 |
|
| 34 |
# 2. CFA-codebert-c.pt -> codebert-c finetuning model
|
| 35 |
+
path = os.getcwd() + '/models/CFA-codebert-c.pt'
|
| 36 |
tokenizer = AutoTokenizer.from_pretrained(path)
|
| 37 |
input_ids = tokenizer(code_text, padding=True, max_length=512,
|
| 38 |
truncation=True, return_token_type_ids=True)['input_ids']
|
|
|
|
| 43 |
pred_2 = model(input_ids)[0].detach().cpu().numpy()[0]
|
| 44 |
|
| 45 |
# 3. CFA-codebert-c-v2.pt -> undersampling + codebert-c finetuning model
|
| 46 |
+
path = os.getcwd() + '/models/CFA-codebert-c-v2.pt'
|
| 47 |
tokenizer = RobertaTokenizer.from_pretrained(path)
|
| 48 |
input_ids = tokenizer(code_text, padding=True, max_length=512,
|
| 49 |
truncation=True, return_token_type_ids=True)['input_ids']
|
|
|
|
| 54 |
pred_3 = model(input_ids)[0].detach().cpu().numpy()
|
| 55 |
|
| 56 |
# 4. codeT5 finetuning model
|
| 57 |
+
path = os.getcwd() + '/models/CFA-codeT5'
|
| 58 |
model_params = {
|
| 59 |
# model_type: t5-base/t5-large
|
| 60 |
"MODEL": path,
|