lewtun HF staff commited on
Commit
f75c7ed
1 Parent(s): 5fb4856

Disable data saving:

Browse files
Files changed (1) hide show
  1. app.py +38 -38
app.py CHANGED
@@ -29,24 +29,24 @@ def randomize_seed_generator():
29
  return seed
30
 
31
 
32
- def save_inputs_and_outputs(now, inputs, outputs, generate_kwargs, model):
33
- buffer = StringIO()
34
- timestamp = datetime.datetime.now().strftime("%Y-%m-%dT%H:%M:%S.%f")
35
- file_name = f"prompts_{timestamp}.jsonl"
36
- data = {"model": model, "inputs": inputs, "outputs": outputs, "generate_kwargs": generate_kwargs}
37
- pd.DataFrame([data]).to_json(buffer, orient="records", lines=True)
38
-
39
- # Push to Hub
40
- upload_file(
41
- path_in_repo=f"{now.date()}/{now.hour}/{file_name}",
42
- path_or_fileobj=buffer.getvalue().encode(),
43
- repo_id=DIALOGUES_DATASET,
44
- token=HF_TOKEN,
45
- repo_type="dataset",
46
- )
47
-
48
- # Clean and rerun
49
- buffer.close()
50
 
51
 
52
  def get_total_inputs(inputs, chatbot, preprompt, user_name, assistant_name, sep):
@@ -94,7 +94,7 @@ def generate(
94
  top_p,
95
  max_new_tokens,
96
  repetition_penalty,
97
- do_save=True,
98
  ):
99
  client = Client(
100
  model2endpoint[model_name],
@@ -177,14 +177,14 @@ def generate(
177
 
178
  yield chat, history, user_message, ""
179
 
180
- if HF_TOKEN and do_save:
181
- try:
182
- now = datetime.datetime.now()
183
- current_time = now.strftime("%Y-%m-%d %H:%M:%S")
184
- print(f"[{current_time}] Pushing prompt and completion to the Hub")
185
- save_inputs_and_outputs(now, prompt, output, generate_kwargs, model_name)
186
- except Exception as e:
187
- print(e)
188
 
189
  return chat, history, user_message, ""
190
 
@@ -231,7 +231,7 @@ def retry_last_answer(
231
  top_p,
232
  max_new_tokens,
233
  repetition_penalty,
234
- do_save,
235
  ):
236
  if chat and history:
237
  # Removing the previous conversation from chat
@@ -255,7 +255,7 @@ def retry_last_answer(
255
  top_p,
256
  max_new_tokens,
257
  repetition_penalty,
258
- do_save,
259
  )
260
 
261
 
@@ -294,12 +294,12 @@ with gr.Blocks(analytics_enabled=False, css=custom_css) as demo:
294
  """
295
  )
296
 
297
- with gr.Row():
298
- do_save = gr.Checkbox(
299
- value=True,
300
- label="Store data",
301
- info="You agree to the storage of your prompt and generated text for research and development purposes:",
302
- )
303
 
304
  with gr.Row():
305
  selected_model = gr.Radio(choices=model_names, value=model_names[1], label="Select a model")
@@ -406,7 +406,7 @@ with gr.Blocks(analytics_enabled=False, css=custom_css) as demo:
406
  top_p,
407
  max_new_tokens,
408
  repetition_penalty,
409
- do_save,
410
  ],
411
  outputs=[chatbot, history, last_user_message, user_message],
412
  )
@@ -425,7 +425,7 @@ with gr.Blocks(analytics_enabled=False, css=custom_css) as demo:
425
  top_p,
426
  max_new_tokens,
427
  repetition_penalty,
428
- do_save,
429
  ],
430
  outputs=[chatbot, history, last_user_message, user_message],
431
  )
@@ -443,7 +443,7 @@ with gr.Blocks(analytics_enabled=False, css=custom_css) as demo:
443
  top_p,
444
  max_new_tokens,
445
  repetition_penalty,
446
- do_save,
447
  ],
448
  outputs=[chatbot, history, last_user_message, user_message],
449
  )
29
  return seed
30
 
31
 
32
+ # def save_inputs_and_outputs(now, inputs, outputs, generate_kwargs, model):
33
+ # buffer = StringIO()
34
+ # timestamp = datetime.datetime.now().strftime("%Y-%m-%dT%H:%M:%S.%f")
35
+ # file_name = f"prompts_{timestamp}.jsonl"
36
+ # data = {"model": model, "inputs": inputs, "outputs": outputs, "generate_kwargs": generate_kwargs}
37
+ # pd.DataFrame([data]).to_json(buffer, orient="records", lines=True)
38
+
39
+ # # Push to Hub
40
+ # upload_file(
41
+ # path_in_repo=f"{now.date()}/{now.hour}/{file_name}",
42
+ # path_or_fileobj=buffer.getvalue().encode(),
43
+ # repo_id=DIALOGUES_DATASET,
44
+ # token=HF_TOKEN,
45
+ # repo_type="dataset",
46
+ # )
47
+
48
+ # # Clean and rerun
49
+ # buffer.close()
50
 
51
 
52
  def get_total_inputs(inputs, chatbot, preprompt, user_name, assistant_name, sep):
94
  top_p,
95
  max_new_tokens,
96
  repetition_penalty,
97
+ # do_save=True,
98
  ):
99
  client = Client(
100
  model2endpoint[model_name],
177
 
178
  yield chat, history, user_message, ""
179
 
180
+ # if HF_TOKEN and do_save:
181
+ # try:
182
+ # now = datetime.datetime.now()
183
+ # current_time = now.strftime("%Y-%m-%d %H:%M:%S")
184
+ # print(f"[{current_time}] Pushing prompt and completion to the Hub")
185
+ # save_inputs_and_outputs(now, prompt, output, generate_kwargs, model_name)
186
+ # except Exception as e:
187
+ # print(e)
188
 
189
  return chat, history, user_message, ""
190
 
231
  top_p,
232
  max_new_tokens,
233
  repetition_penalty,
234
+ # do_save,
235
  ):
236
  if chat and history:
237
  # Removing the previous conversation from chat
255
  top_p,
256
  max_new_tokens,
257
  repetition_penalty,
258
+ # do_save,
259
  )
260
 
261
 
294
  """
295
  )
296
 
297
+ # with gr.Row():
298
+ # do_save = gr.Checkbox(
299
+ # value=True,
300
+ # label="Store data",
301
+ # info="You agree to the storage of your prompt and generated text for research and development purposes:",
302
+ # )
303
 
304
  with gr.Row():
305
  selected_model = gr.Radio(choices=model_names, value=model_names[1], label="Select a model")
406
  top_p,
407
  max_new_tokens,
408
  repetition_penalty,
409
+ # do_save,
410
  ],
411
  outputs=[chatbot, history, last_user_message, user_message],
412
  )
425
  top_p,
426
  max_new_tokens,
427
  repetition_penalty,
428
+ # do_save,
429
  ],
430
  outputs=[chatbot, history, last_user_message, user_message],
431
  )
443
  top_p,
444
  max_new_tokens,
445
  repetition_penalty,
446
+ # do_save,
447
  ],
448
  outputs=[chatbot, history, last_user_message, user_message],
449
  )