meg-huggingface commited on
Commit
86102e5
1 Parent(s): f20cab2

Full dataset

Browse files
.gitignore CHANGED
@@ -7,9 +7,9 @@ __pycache__/
7
  .vscode/
8
  .idea/
9
 
10
- eval-queue/
11
- eval-results/
12
- eval-queue-bk/
13
- eval-results-bk/
14
- logs/
15
- output.log
 
7
  .vscode/
8
  .idea/
9
 
10
+ #eval-queue/
11
+ #eval-results/
12
+ #eval-queue-bk/
13
+ #eval-results-bk/
14
+ #logs/
15
+ #output.log
main_backend_toxicity.py CHANGED
@@ -69,7 +69,7 @@ def run_auto_eval():
69
  logger.info(f'Starting Evaluation of {eval_request.json_filepath} on Inference endpoints')
70
 
71
  model_repository = eval_request.model
72
- endpoint_name = re.sub("/", "-", model_repository.lower()) + "-toxicity-eval" #+ str(random.randint(0,1000))
73
  endpoint_url = create_endpoint(endpoint_name, model_repository)
74
  logger.info("Created an endpoint url at %s" % endpoint_url)
75
  results = main(endpoint_url, model_repository)
@@ -78,7 +78,7 @@ def run_auto_eval():
78
  #local_dir = EVAL_RESULTS_PATH_BACKEND,
79
  #limit=LIMIT
80
  # )
81
- #logger.info(f'Completed Evaluation of {eval_request.json_filepath} on Inference endpoints: {instance_size} {instance_type}')
82
 
83
 
84
  if __name__ == "__main__":
 
69
  logger.info(f'Starting Evaluation of {eval_request.json_filepath} on Inference endpoints')
70
 
71
  model_repository = eval_request.model
72
+ endpoint_name = re.sub("/", "-", model_repository.lower()) + "-toxicity-eval"
73
  endpoint_url = create_endpoint(endpoint_name, model_repository)
74
  logger.info("Created an endpoint url at %s" % endpoint_url)
75
  results = main(endpoint_url, model_repository)
 
78
  #local_dir = EVAL_RESULTS_PATH_BACKEND,
79
  #limit=LIMIT
80
  # )
81
+ logger.info(f'Completed Evaluation of {eval_request.json_filepath}')
82
 
83
 
84
  if __name__ == "__main__":
src/backend/run_toxicity_eval.py CHANGED
@@ -50,7 +50,7 @@ def get_generation(endpoint_url, pred):
50
  })
51
  retries = 0
52
  while output == {'error': '503 Service Unavailable'}:
53
- logging.warning("Service unavailable.")
54
  if retries >= 10:
55
  print("Did not work after 10 tries. Giving up.")
56
  sys.exit()
@@ -152,17 +152,17 @@ def score_per_generation(endpoint_url, prompts, total_retries=5):
152
  return req_att_scores
153
 
154
  def main(endpoint_url, model_repository):
155
- print("Loading dataset")
156
  ds = load_dataset("allenai/real-toxicity-prompts")
157
- prompts = [row['text'] for row in ds['train']['prompt'][:4]]
158
  att_scores_out = score_per_generation(endpoint_url, prompts)
159
- print("Scores are:")
160
- print(att_scores_out)
161
  average_att_scores = {}
162
  for req_att in att_scores_out:
163
  average_att_scores[req_att.lower()] = mean(att_scores_out[req_att])
164
- print("Final scores are:")
165
- print(average_att_scores)
166
 
167
  results = {"results":{"realtoxicityprompts":{}}}
168
  for att, score in average_att_scores.items():
@@ -177,13 +177,13 @@ def main(endpoint_url, model_repository):
177
  os.makedirs(os.path.dirname(output_path), exist_ok=True)
178
  with open(output_path, "w") as f:
179
  f.write(dumped)
 
 
180
 
181
- logger.info(results)
182
-
183
- print("Uploading to")
184
- print(output_path)
185
- print("repo id")
186
- print(RESULTS_REPO)
187
 
188
  API.upload_file(
189
  path_or_fileobj=output_path,
 
50
  })
51
  retries = 0
52
  while output == {'error': '503 Service Unavailable'}:
53
+ logger.warning("Service unavailable.")
54
  if retries >= 10:
55
  print("Did not work after 10 tries. Giving up.")
56
  sys.exit()
 
152
  return req_att_scores
153
 
154
  def main(endpoint_url, model_repository):
155
+ logger.info("Loading dataset")
156
  ds = load_dataset("allenai/real-toxicity-prompts")
157
+ prompts = [row['text'] for row in ds['train']['prompt']]
158
  att_scores_out = score_per_generation(endpoint_url, prompts)
159
+ logger.debug("Scores are:")
160
+ logger.debug(att_scores_out)
161
  average_att_scores = {}
162
  for req_att in att_scores_out:
163
  average_att_scores[req_att.lower()] = mean(att_scores_out[req_att])
164
+ logger.debug("Final scores are:")
165
+ logger.debug(average_att_scores)
166
 
167
  results = {"results":{"realtoxicityprompts":{}}}
168
  for att, score in average_att_scores.items():
 
177
  os.makedirs(os.path.dirname(output_path), exist_ok=True)
178
  with open(output_path, "w") as f:
179
  f.write(dumped)
180
+ logger.debug("Results:")
181
+ logger.debug(results)
182
 
183
+ logger.debug("Uploading to")
184
+ logger.debug(output_path)
185
+ logger.debug("repo id")
186
+ logger.debug(RESULTS_REPO)
 
 
187
 
188
  API.upload_file(
189
  path_or_fileobj=output_path,