priyasaravana commited on
Commit
d1231be
1 Parent(s): ae5bfc1

Update gradio_app.py

Browse files
Files changed (1) hide show
  1. gradio_app.py +18 -17
gradio_app.py CHANGED
@@ -8,18 +8,10 @@ from gradio import utils
8
  import huggingface_hub
9
  from pathlib import Path
10
  from src.utils.utilities import Utility
11
-
12
- model = BERTClassifier(model_name='jeevavijay10/nlp-goemotions-bert')
13
-
14
- classes = Utility().read_emotion_list()
15
-
16
- hf_token = os.getenv("HF_TOKEN")
17
 
18
  dataset_dir = "logs"
19
-
20
  headers = ["input", "output", "timestamp", "elapsed"]
21
-
22
-
23
  repo = huggingface_hub.Repository(
24
  local_dir=dataset_dir,
25
  clone_from="https://huggingface.co/spaces/priyasaravana/CodeSummarization",
@@ -38,21 +30,30 @@ def log_record(vals):
38
  schedule.run_pending()
39
  print(f"Last Sync: {job.last_run}")
40
 
41
- def predict(sentence):
42
-
 
 
 
 
 
 
 
 
 
 
 
 
43
  timestamp = datetime.datetime.now().isoformat()
44
  start_time = time.time()
45
- predictions = model.evaluate([sentence])
46
- elapsed_time = time.time() - start_time
47
-
48
- output = classes[predictions[0]]
49
-
50
  print(f"Sentence: {sentence} \nPrediction: {predictions[0]} - {output}")
51
  log_record([sentence, output, timestamp, str(elapsed_time)])
52
 
53
  return output
54
 
55
-
56
  def sync_logs():
57
  print(f"Repo Clean: {repo.is_repo_clean()}")
58
  if not repo.is_repo_clean():
 
8
  import huggingface_hub
9
  from pathlib import Path
10
  from src.utils.utilities import Utility
11
+ from transformers import RobertaTokenizer, T5ForConditionalGeneration
 
 
 
 
 
12
 
13
  dataset_dir = "logs"
 
14
  headers = ["input", "output", "timestamp", "elapsed"]
 
 
15
  repo = huggingface_hub.Repository(
16
  local_dir=dataset_dir,
17
  clone_from="https://huggingface.co/spaces/priyasaravana/CodeSummarization",
 
30
  schedule.run_pending()
31
  print(f"Last Sync: {job.last_run}")
32
 
33
+ def evaluate(sentence):
34
+ tokenizer = RobertaTokenizer.from_pretrained('Salesforce/codet5-base')
35
+ model = T5ForConditionalGeneration.from_pretrained('Salesforce/codet5-base-multi-sum')
36
+
37
+ # Prepare the input text
38
+ input_text = code_snippet.strip()
39
+ input_ids = tokenizer.encode(input_text, return_tensors='pt')
40
+ # Generate a summary
41
+ generated_ids = model.generate(input_ids, max_length=20)
42
+ summary = tokenizer.decode(generated_ids[0], skip_special_tokens=True)
43
+
44
+ return summary
45
+
46
+ def predict(sentence):
47
  timestamp = datetime.datetime.now().isoformat()
48
  start_time = time.time()
49
+ predictions = evaluate([sentence])
50
+ elapsed_time = time.time() - start_time
51
+ output = predictions
 
 
52
  print(f"Sentence: {sentence} \nPrediction: {predictions[0]} - {output}")
53
  log_record([sentence, output, timestamp, str(elapsed_time)])
54
 
55
  return output
56
 
 
57
  def sync_logs():
58
  print(f"Repo Clean: {repo.is_repo_clean()}")
59
  if not repo.is_repo_clean():