-
Notifications
You must be signed in to change notification settings - Fork 0
/
gpt-3_openai_terminal_validation_model_testing.txt
354 lines (325 loc) · 17.1 KB
/
gpt-3_openai_terminal_validation_model_testing.txt
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
(gpt3) ThinkPad-T560:openai_dataset $ ls
example_train.json example_validation.json
(gpt3) ThinkPad-T560:openai_dataset $ openai tools fine_tunes.prepare_data -f example_train.json
Analyzing...
- Your file contains 500 prompt-completion pairs
- All prompts end with suffix `}`
- All prompts start with prefix `{'title': '`
- All completions end with suffix `]`
- The completion should start with a whitespace character (` `). This tends to produce better results due to the tokenization we use. See https://beta.openai.com/docs/guides/fine-tuning/preparing-your-dataset for more details
Based on the analysis we will perform the following actions:
- [Recommended] Add a whitespace character to the beginning of the completion [Y/n]: y
Your data will be written to a new JSONL file. Proceed [Y/n]: y
Wrote modified file to `example_train_prepared.jsonl`
Feel free to take a look!
Now use that file when fine-tuning:
> openai api fine_tunes.create -t "example_train_prepared.jsonl"
After you’ve fine-tuned a model, remember that your prompt has to end with the indicator string `}` for the model to start generating completions, rather than continuing with the prompt. Make sure to include `stop=["]"]` so that the generated texts ends at the expected place.
Once your model starts training, it'll approximately take 9.31 minutes to train a `curie` model, and less for `ada` and `babbage`. Queue will approximately take half an hour per job ahead of you.
(gpt3) ThinkPad-T560:openai_dataset $ openai tools fine_tunes.prepare_data -f example_validation.json
Analyzing...
- Your file contains 500 prompt-completion pairs
- There are 8 duplicated prompt-completion sets. These are rows: [17, 24, 162, 325, 340, 370, 430, 468]
- All prompts end with suffix `}`
- All prompts start with prefix `{'title': 'Super_Bowl_50', 'context': `. Fine-tuning doesn't require the instruction specifying the task, or a few-shot example scenario. Most of the time you should only add the input data into the prompt, and the desired output into the completion
- All completions end with suffix `]`
- The completion should start with a whitespace character (` `). This tends to produce better results due to the tokenization we use. See https://beta.openai.com/docs/guides/fine-tuning/preparing-your-dataset for more details
Based on the analysis we will perform the following actions:
- [Recommended] Remove 8 duplicate rows [Y/n]: y
- [Recommended] Remove prefix `{'title': 'Super_Bowl_50', 'context': ` from all prompts [Y/n]: y
/home/rhakbari/.virtualenvs/gpt3/lib/python3.10/site-packages/openai/validators.py:279: SettingWithCopyWarning:
A value is trying to be set on a copy of a slice from a DataFrame.
Try using .loc[row_indexer,col_indexer] = value instead
See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy
x["prompt"] = x["prompt"].str[len(prefix) :]
- [Recommended] Add a whitespace character to the beginning of the completion [Y/n]: y
/home/rhakbari/.virtualenvs/gpt3/lib/python3.10/site-packages/openai/validators.py:421: SettingWithCopyWarning:
A value is trying to be set on a copy of a slice from a DataFrame.
Try using .loc[row_indexer,col_indexer] = value instead
See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy
x["completion"] = x["completion"].apply(
Your data will be written to a new JSONL file. Proceed [Y/n]: y
Wrote modified file to `example_validation_prepared.jsonl`
Feel free to take a look!
Now use that file when fine-tuning:
> openai api fine_tunes.create -t "example_validation_prepared.jsonl"
After you’ve fine-tuned a model, remember that your prompt has to end with the indicator string `}` for the model to start generating completions, rather than continuing with the prompt. Make sure to include `stop=["]"]` so that the generated texts ends at the expected place.
Once your model starts training, it'll approximately take 26.68 minutes to train a `curie` model, and less for `ada` and `babbage`. Queue will approximately take half an hour per job ahead of you.
(gpt3) ThinkPad-T560:openai_dataset $ openai api fine_tunes.create -t example_train_prepared.jsonl -v example_validation_prepared.jsonl -m davinci
Upload progress: 100%|███████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 569k/569k [00:00<00:00, 379Mit/s]
Uploaded file from example_train_prepared.jsonl: file-BaSgvdFFgpJVlaGtNdlHklKo
Upload progress: 100%|███████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 401k/401k [00:00<00:00, 124Mit/s]
Uploaded file from example_validation_prepared.jsonl: file-I3JjtwFBkFZMoCTflxfwWxGt
Created fine-tune: ft-Gm0iAk5ZaexwCLcup9rwWCS2
Streaming events until fine-tuning is complete...
(Ctrl-C will interrupt the stream, but not cancel the fine-tune)
[2022-11-22 23:49:09] Created fine-tune: ft-Gm0iAk5ZaexwCLcup9rwWCS2
[2022-11-22 23:49:17] Fine-tune failed. Fine-tune will exceed billing hard limit
Job failed. Please contact [email protected] if you need assistance.
(gpt3) ThinkPad-T560:openai_dataset $ openai api fine_tunes.create -t example_train_prepared.jsonl -v example_validation_prepared.jsonl -m davinci
Found potentially duplicated files with name 'example_train_prepared.jsonl', purpose 'fine-tune' and size 568893 bytes
file-BaSgvdFFgpJVlaGtNdlHklKo
Enter file ID to reuse an already uploaded file, or an empty string to upload this file anyway:
Upload progress: 100%|███████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 569k/569k [00:00<00:00, 377Mit/s]
Uploaded file from example_train_prepared.jsonl: file-mu5L0OpcXddxMac2qCN9d9Ng
Found potentially duplicated files with name 'example_validation_prepared.jsonl', purpose 'fine-tune' and size 401392 bytes
file-I3JjtwFBkFZMoCTflxfwWxGt
Enter file ID to reuse an already uploaded file, or an empty string to upload this file anyway:
Upload progress: 100%|███████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 401k/401k [00:00<00:00, 165Mit/s]
Uploaded file from example_validation_prepared.jsonl: file-USYcMDA0ZTIB61AmjHTcguVl
Created fine-tune: ft-UohoIaNdc6yZW9dziJyznUkQ
Streaming events until fine-tuning is complete...
(Ctrl-C will interrupt the stream, but not cancel the fine-tune)
[2022-11-22 23:50:10] Created fine-tune: ft-UohoIaNdc6yZW9dziJyznUkQ
[2022-11-22 23:50:15] Fine-tune costs $14.73
[2022-11-22 23:50:16] Fine-tune enqueued. Queue number: 3
[2022-11-22 23:51:35] Fine-tune is in the queue. Queue number: 2
Stream interrupted (client disconnected).
To resume the stream, run:
openai api fine_tunes.follow -i ft-UohoIaNdc6yZW9dziJyznUkQ
(gpt3) ThinkPad-T560:openai_dataset $
(gpt3) ThinkPad-T560:openai_dataset $ openai api fine_tunes.create -t example_train_prepared.jsonl -v example_validation_prepared.jsonl -m davinci
Found potentially duplicated files with name 'example_train_prepared.jsonl', purpose 'fine-tune' and size 568893 bytes
file-mu5L0OpcXddxMac2qCN9d9Ng
file-BaSgvdFFgpJVlaGtNdlHklKo
Enter file ID to reuse an already uploaded file, or an empty string to upload this file anyway:
Upload progress: 100%|███████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 569k/569k [00:00<00:00, 594Mit/s]
Uploaded file from example_train_prepared.jsonl: file-nH36yEO3lsVTEoofQ1QmjMqp
Found potentially duplicated files with name 'example_validation_prepared.jsonl', purpose 'fine-tune' and size 401392 bytes
file-I3JjtwFBkFZMoCTflxfwWxGt
file-USYcMDA0ZTIB61AmjHTcguVl
Enter file ID to reuse an already uploaded file, or an empty string to upload this file anyway:
Upload progress: 100%|███████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 401k/401k [00:00<00:00, 125Mit/s]
Uploaded file from example_validation_prepared.jsonl: file-NSHsCurfdf7ihZmNdJCmdXmT
Created fine-tune: ft-RGN7XHkQPVmuUSPbb7YG4zda
Streaming events until fine-tuning is complete...
(Ctrl-C will interrupt the stream, but not cancel the fine-tune)
[2022-11-23 00:02:07] Created fine-tune: ft-RGN7XHkQPVmuUSPbb7YG4zda
(gpt3) ThinkPad-T560:openai_dataset $ openai api fine_tunes.list
{
"data": [
{
"created_at": 1669105199,
"fine_tuned_model": "curie:ft-personal-2022-11-22-08-21-19",
"hyperparams": {
"batch_size": 1,
"learning_rate_multiplier": 0.1,
"n_epochs": 4,
"prompt_loss_weight": 0.01
},
"id": "ft-Uh58zHfBzCMhD34Bnzxn05jY",
"model": "curie",
"object": "fine-tune",
"organization_id": "org-UA0xqHRa8sJZiblAwTQ6C0kt",
"result_files": [
{
"bytes": 1869,
"created_at": 1669105280,
"filename": "compiled_results.csv",
"id": "file-yqY3UikoeY4A1nMehDLyMUiI",
"object": "file",
"purpose": "fine-tune-results",
"status": "processed",
"status_details": null
}
],
"status": "succeeded",
"training_files": [
{
"bytes": 1110,
"created_at": 1669105198,
"filename": "example-test-1_prepared.jsonl",
"id": "file-N5LVL48NBDVFvPWj1iqlIDqn",
"object": "file",
"purpose": "fine-tune",
"status": "processed",
"status_details": null
}
],
"updated_at": 1669105280,
"validation_files": []
},
{
"created_at": 1669142376,
"fine_tuned_model": null,
"hyperparams": {
"batch_size": null,
"learning_rate_multiplier": null,
"n_epochs": 4,
"prompt_loss_weight": 0.01
},
"id": "ft-3IWX8OSkvVfYoREBN1nHHoVm",
"model": "curie",
"object": "fine-tune",
"organization_id": "org-UA0xqHRa8sJZiblAwTQ6C0kt",
"result_files": [],
"status": "failed",
"training_files": [
{
"bytes": 81556009,
"created_at": 1669142376,
"filename": "example_train_prepared.jsonl",
"id": "file-GbCmPfDb4IViGC3E5rZoDdHQ",
"object": "file",
"purpose": "fine-tune",
"status": "processed",
"status_details": null
}
],
"updated_at": 1669142415,
"validation_files": []
},
{
"created_at": 1669142949,
"fine_tuned_model": null,
"hyperparams": {
"batch_size": null,
"learning_rate_multiplier": null,
"n_epochs": 4,
"prompt_loss_weight": 0.01
},
"id": "ft-Gm0iAk5ZaexwCLcup9rwWCS2",
"model": "davinci",
"object": "fine-tune",
"organization_id": "org-UA0xqHRa8sJZiblAwTQ6C0kt",
"result_files": [],
"status": "failed",
"training_files": [
{
"bytes": 568893,
"created_at": 1669142946,
"filename": "example_train_prepared.jsonl",
"id": "file-BaSgvdFFgpJVlaGtNdlHklKo",
"object": "file",
"purpose": "fine-tune",
"status": "processed",
"status_details": null
}
],
"updated_at": 1669142957,
"validation_files": [
{
"bytes": 401392,
"created_at": 1669142949,
"filename": "example_validation_prepared.jsonl",
"id": "file-I3JjtwFBkFZMoCTflxfwWxGt",
"object": "file",
"purpose": "fine-tune",
"status": "processed",
"status_details": null
}
]
},
{
"created_at": 1669143010,
"fine_tuned_model": "davinci:ft-personal-2022-11-22-20-19-27",
"hyperparams": {
"batch_size": 1,
"learning_rate_multiplier": 0.1,
"n_epochs": 4,
"prompt_loss_weight": 0.01
},
"id": "ft-UohoIaNdc6yZW9dziJyznUkQ",
"model": "davinci",
"object": "fine-tune",
"organization_id": "org-UA0xqHRa8sJZiblAwTQ6C0kt",
"result_files": [
{
"bytes": 108010,
"created_at": 1669148368,
"filename": "compiled_results.csv",
"id": "file-q587lqpCBWv6WGJRGaTGcvBa",
"object": "file",
"purpose": "fine-tune-results",
"status": "processed",
"status_details": null
}
],
"status": "succeeded",
"training_files": [
{
"bytes": 568893,
"created_at": 1669143005,
"filename": "example_train_prepared.jsonl",
"id": "file-mu5L0OpcXddxMac2qCN9d9Ng",
"object": "file",
"purpose": "fine-tune",
"status": "processed",
"status_details": null
}
],
"updated_at": 1669148368,
"validation_files": [
{
"bytes": 401392,
"created_at": 1669143010,
"filename": "example_validation_prepared.jsonl",
"id": "file-USYcMDA0ZTIB61AmjHTcguVl",
"object": "file",
"purpose": "fine-tune",
"status": "processed",
"status_details": null
}
]
},
{
"created_at": 1669143727,
"fine_tuned_model": "davinci:ft-personal-2022-11-22-21-04-53",
"hyperparams": {
"batch_size": 1,
"learning_rate_multiplier": 0.1,
"n_epochs": 4,
"prompt_loss_weight": 0.01
},
"id": "ft-RGN7XHkQPVmuUSPbb7YG4zda",
"model": "davinci",
"object": "fine-tune",
"organization_id": "org-UA0xqHRa8sJZiblAwTQ6C0kt",
"result_files": [
{
"bytes": 107796,
"created_at": 1669151095,
"filename": "compiled_results.csv",
"id": "file-xbSCXJFWJTEr5Ne2QbfwMUN2",
"object": "file",
"purpose": "fine-tune-results",
"status": "processed",
"status_details": null
}
],
"status": "succeeded",
"training_files": [
{
"bytes": 568893,
"created_at": 1669143723,
"filename": "example_train_prepared.jsonl",
"id": "file-nH36yEO3lsVTEoofQ1QmjMqp",
"object": "file",
"purpose": "fine-tune",
"status": "processed",
"status_details": null
}
],
"updated_at": 1669151095,
"validation_files": [
{
"bytes": 401392,
"created_at": 1669143727,
"filename": "example_validation_prepared.jsonl",
"id": "file-NSHsCurfdf7ihZmNdJCmdXmT",
"object": "file",
"purpose": "fine-tune",
"status": "processed",
"status_details": null
}
]
}
],
"object": "list"
}
$ python model_testing.py
[<OpenAIObject at 0x7f00854e15d0> JSON: {
" The": -2.7414877,
"\u00a0": -3.0495987
}]
Amazonia