tmnam20 commited on
Commit
886ac22
1 Parent(s): ddb2b9f

Training in progress, step 29000, checkpoint

Browse files
checkpoint-29000/added_tokens.json ADDED
@@ -0,0 +1,5 @@
 
 
 
 
 
 
1
+ {
2
+ "<bos>": 40478,
3
+ "<eos>": 40479,
4
+ "<pad>": 40480
5
+ }
checkpoint-29000/config.json ADDED
@@ -0,0 +1,46 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "openai-gpt",
3
+ "afn": "gelu",
4
+ "architectures": [
5
+ "OpenAIGPTForSequenceClassification"
6
+ ],
7
+ "attn_pdrop": 0.1,
8
+ "embd_pdrop": 0.1,
9
+ "id2label": {
10
+ "0": "entailment",
11
+ "1": "neutral",
12
+ "2": "contradiction"
13
+ },
14
+ "initializer_range": 0.02,
15
+ "label2id": {
16
+ "contradiction": 2,
17
+ "entailment": 0,
18
+ "neutral": 1
19
+ },
20
+ "layer_norm_epsilon": 1e-05,
21
+ "model_type": "openai-gpt",
22
+ "n_ctx": 512,
23
+ "n_embd": 768,
24
+ "n_head": 12,
25
+ "n_layer": 12,
26
+ "n_positions": 512,
27
+ "n_special": 0,
28
+ "pad_token_id": 40480,
29
+ "predict_special_tokens": true,
30
+ "problem_type": "single_label_classification",
31
+ "resid_pdrop": 0.1,
32
+ "summary_activation": null,
33
+ "summary_first_dropout": 0.1,
34
+ "summary_proj_to_labels": true,
35
+ "summary_type": "cls_index",
36
+ "summary_use_proj": true,
37
+ "task_specific_params": {
38
+ "text-generation": {
39
+ "do_sample": true,
40
+ "max_length": 50
41
+ }
42
+ },
43
+ "torch_dtype": "float32",
44
+ "transformers_version": "4.33.0.dev0",
45
+ "vocab_size": 40481
46
+ }
checkpoint-29000/merges.txt ADDED
The diff for this file is too large to render. See raw diff
 
checkpoint-29000/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4da02a47506ec3a388d39d478a1e6dcec3880af1f72cbd23e3914ea16ba85017
3
+ size 932401285
checkpoint-29000/pytorch_model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:46540af3350ba7c6120acf69f95b4aa519c492c8b45ebee6e5b8b6a4c600b59d
3
+ size 466205145
checkpoint-29000/rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:714c0df5dbc6b44c4599172a9bdb7c0174b513e9b44bf6820d1d1468e99048c2
3
+ size 14575
checkpoint-29000/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c9a13827704ca46ded964fe7fe701cd8df0d3078e8195d069fd52dff43a81fd6
3
+ size 627
checkpoint-29000/special_tokens_map.json ADDED
@@ -0,0 +1,6 @@
 
 
 
 
 
 
 
1
+ {
2
+ "bos_token": "<bos>",
3
+ "eos_token": "<eos>",
4
+ "pad_token": "<pad>",
5
+ "unk_token": "<unk>"
6
+ }
checkpoint-29000/tokenizer_config.json ADDED
@@ -0,0 +1,7 @@
 
 
 
 
 
 
 
 
1
+ {
2
+ "clean_up_tokenization_spaces": true,
3
+ "model_max_length": 512,
4
+ "padding_side": "left",
5
+ "tokenizer_class": "OpenAIGPTTokenizer",
6
+ "unk_token": "<unk>"
7
+ }
checkpoint-29000/trainer_state.json ADDED
@@ -0,0 +1,460 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": 1.0414146184921265,
3
+ "best_model_checkpoint": "/kaggle/output/checkpoint-29000",
4
+ "epoch": 1.1815514993481095,
5
+ "eval_steps": 1000,
6
+ "global_step": 29000,
7
+ "is_hyper_param_search": false,
8
+ "is_local_process_zero": true,
9
+ "is_world_process_zero": true,
10
+ "log_history": [
11
+ {
12
+ "epoch": 0.0,
13
+ "learning_rate": 2.7777777777777777e-11,
14
+ "loss": 1.1065,
15
+ "step": 1
16
+ },
17
+ {
18
+ "epoch": 0.04,
19
+ "learning_rate": 2.7750000000000004e-08,
20
+ "loss": 1.1586,
21
+ "step": 1000
22
+ },
23
+ {
24
+ "epoch": 0.04,
25
+ "eval_accuracy": 0.32954091816367265,
26
+ "eval_loss": 1.1065887212753296,
27
+ "eval_runtime": 54.4372,
28
+ "eval_samples_per_second": 92.033,
29
+ "eval_steps_per_second": 11.518,
30
+ "step": 1000
31
+ },
32
+ {
33
+ "epoch": 0.08,
34
+ "learning_rate": 5.5527777777777784e-08,
35
+ "loss": 1.1259,
36
+ "step": 2000
37
+ },
38
+ {
39
+ "epoch": 0.08,
40
+ "eval_accuracy": 0.3435129740518962,
41
+ "eval_loss": 1.1045255661010742,
42
+ "eval_runtime": 54.458,
43
+ "eval_samples_per_second": 91.997,
44
+ "eval_steps_per_second": 11.513,
45
+ "step": 2000
46
+ },
47
+ {
48
+ "epoch": 0.12,
49
+ "learning_rate": 8.327777777777778e-08,
50
+ "loss": 1.1229,
51
+ "step": 3000
52
+ },
53
+ {
54
+ "epoch": 0.12,
55
+ "eval_accuracy": 0.3441117764471058,
56
+ "eval_loss": 1.1050777435302734,
57
+ "eval_runtime": 54.349,
58
+ "eval_samples_per_second": 92.182,
59
+ "eval_steps_per_second": 11.537,
60
+ "step": 3000
61
+ },
62
+ {
63
+ "epoch": 0.16,
64
+ "learning_rate": 1.1105555555555557e-07,
65
+ "loss": 1.1237,
66
+ "step": 4000
67
+ },
68
+ {
69
+ "epoch": 0.16,
70
+ "eval_accuracy": 0.3411177644710579,
71
+ "eval_loss": 1.1015312671661377,
72
+ "eval_runtime": 54.4388,
73
+ "eval_samples_per_second": 92.03,
74
+ "eval_steps_per_second": 11.518,
75
+ "step": 4000
76
+ },
77
+ {
78
+ "epoch": 0.2,
79
+ "learning_rate": 1.3883333333333335e-07,
80
+ "loss": 1.1202,
81
+ "step": 5000
82
+ },
83
+ {
84
+ "epoch": 0.2,
85
+ "eval_accuracy": 0.3405189620758483,
86
+ "eval_loss": 1.1006112098693848,
87
+ "eval_runtime": 54.4109,
88
+ "eval_samples_per_second": 92.077,
89
+ "eval_steps_per_second": 11.523,
90
+ "step": 5000
91
+ },
92
+ {
93
+ "epoch": 0.24,
94
+ "learning_rate": 1.6658333333333335e-07,
95
+ "loss": 1.1182,
96
+ "step": 6000
97
+ },
98
+ {
99
+ "epoch": 0.24,
100
+ "eval_accuracy": 0.35489021956087824,
101
+ "eval_loss": 1.0982259511947632,
102
+ "eval_runtime": 54.3841,
103
+ "eval_samples_per_second": 92.123,
104
+ "eval_steps_per_second": 11.529,
105
+ "step": 6000
106
+ },
107
+ {
108
+ "epoch": 0.29,
109
+ "learning_rate": 1.9436111111111112e-07,
110
+ "loss": 1.1125,
111
+ "step": 7000
112
+ },
113
+ {
114
+ "epoch": 0.29,
115
+ "eval_accuracy": 0.35748502994011977,
116
+ "eval_loss": 1.0954252481460571,
117
+ "eval_runtime": 54.3975,
118
+ "eval_samples_per_second": 92.1,
119
+ "eval_steps_per_second": 11.526,
120
+ "step": 7000
121
+ },
122
+ {
123
+ "epoch": 0.33,
124
+ "learning_rate": 2.2213888888888891e-07,
125
+ "loss": 1.1096,
126
+ "step": 8000
127
+ },
128
+ {
129
+ "epoch": 0.33,
130
+ "eval_accuracy": 0.3664670658682635,
131
+ "eval_loss": 1.0949230194091797,
132
+ "eval_runtime": 54.3263,
133
+ "eval_samples_per_second": 92.221,
134
+ "eval_steps_per_second": 11.541,
135
+ "step": 8000
136
+ },
137
+ {
138
+ "epoch": 0.37,
139
+ "learning_rate": 2.4988888888888893e-07,
140
+ "loss": 1.1057,
141
+ "step": 9000
142
+ },
143
+ {
144
+ "epoch": 0.37,
145
+ "eval_accuracy": 0.34890219560878244,
146
+ "eval_loss": 1.1037237644195557,
147
+ "eval_runtime": 54.3657,
148
+ "eval_samples_per_second": 92.154,
149
+ "eval_steps_per_second": 11.533,
150
+ "step": 9000
151
+ },
152
+ {
153
+ "epoch": 0.41,
154
+ "learning_rate": 2.776666666666667e-07,
155
+ "loss": 1.1056,
156
+ "step": 10000
157
+ },
158
+ {
159
+ "epoch": 0.41,
160
+ "eval_accuracy": 0.3838323353293413,
161
+ "eval_loss": 1.0928014516830444,
162
+ "eval_runtime": 54.6518,
163
+ "eval_samples_per_second": 91.671,
164
+ "eval_steps_per_second": 11.473,
165
+ "step": 10000
166
+ },
167
+ {
168
+ "epoch": 0.45,
169
+ "learning_rate": 3.054444444444444e-07,
170
+ "loss": 1.1036,
171
+ "step": 11000
172
+ },
173
+ {
174
+ "epoch": 0.45,
175
+ "eval_accuracy": 0.3838323353293413,
176
+ "eval_loss": 1.0908663272857666,
177
+ "eval_runtime": 54.579,
178
+ "eval_samples_per_second": 91.794,
179
+ "eval_steps_per_second": 11.488,
180
+ "step": 11000
181
+ },
182
+ {
183
+ "epoch": 0.49,
184
+ "learning_rate": 3.3322222222222225e-07,
185
+ "loss": 1.1012,
186
+ "step": 12000
187
+ },
188
+ {
189
+ "epoch": 0.49,
190
+ "eval_accuracy": 0.3872255489021956,
191
+ "eval_loss": 1.0890672206878662,
192
+ "eval_runtime": 54.5139,
193
+ "eval_samples_per_second": 91.903,
194
+ "eval_steps_per_second": 11.502,
195
+ "step": 12000
196
+ },
197
+ {
198
+ "epoch": 0.53,
199
+ "learning_rate": 3.609722222222222e-07,
200
+ "loss": 1.103,
201
+ "step": 13000
202
+ },
203
+ {
204
+ "epoch": 0.53,
205
+ "eval_accuracy": 0.34311377245508984,
206
+ "eval_loss": 1.0942790508270264,
207
+ "eval_runtime": 54.3951,
208
+ "eval_samples_per_second": 92.104,
209
+ "eval_steps_per_second": 11.527,
210
+ "step": 13000
211
+ },
212
+ {
213
+ "epoch": 0.57,
214
+ "learning_rate": 3.8875e-07,
215
+ "loss": 1.0997,
216
+ "step": 14000
217
+ },
218
+ {
219
+ "epoch": 0.57,
220
+ "eval_accuracy": 0.3712574850299401,
221
+ "eval_loss": 1.0931994915008545,
222
+ "eval_runtime": 54.3749,
223
+ "eval_samples_per_second": 92.138,
224
+ "eval_steps_per_second": 11.531,
225
+ "step": 14000
226
+ },
227
+ {
228
+ "epoch": 0.61,
229
+ "learning_rate": 4.1650000000000006e-07,
230
+ "loss": 1.0993,
231
+ "step": 15000
232
+ },
233
+ {
234
+ "epoch": 0.61,
235
+ "eval_accuracy": 0.3872255489021956,
236
+ "eval_loss": 1.0887969732284546,
237
+ "eval_runtime": 54.4149,
238
+ "eval_samples_per_second": 92.07,
239
+ "eval_steps_per_second": 11.523,
240
+ "step": 15000
241
+ },
242
+ {
243
+ "epoch": 0.65,
244
+ "learning_rate": 4.4427777777777783e-07,
245
+ "loss": 1.0983,
246
+ "step": 16000
247
+ },
248
+ {
249
+ "epoch": 0.65,
250
+ "eval_accuracy": 0.3980039920159681,
251
+ "eval_loss": 1.0836474895477295,
252
+ "eval_runtime": 54.3692,
253
+ "eval_samples_per_second": 92.148,
254
+ "eval_steps_per_second": 11.532,
255
+ "step": 16000
256
+ },
257
+ {
258
+ "epoch": 0.69,
259
+ "learning_rate": 4.7202777777777785e-07,
260
+ "loss": 1.0954,
261
+ "step": 17000
262
+ },
263
+ {
264
+ "epoch": 0.69,
265
+ "eval_accuracy": 0.3996007984031936,
266
+ "eval_loss": 1.0822749137878418,
267
+ "eval_runtime": 54.3313,
268
+ "eval_samples_per_second": 92.212,
269
+ "eval_steps_per_second": 11.54,
270
+ "step": 17000
271
+ },
272
+ {
273
+ "epoch": 0.73,
274
+ "learning_rate": 4.998055555555556e-07,
275
+ "loss": 1.097,
276
+ "step": 18000
277
+ },
278
+ {
279
+ "epoch": 0.73,
280
+ "eval_accuracy": 0.4033932135728543,
281
+ "eval_loss": 1.0845097303390503,
282
+ "eval_runtime": 54.3565,
283
+ "eval_samples_per_second": 92.169,
284
+ "eval_steps_per_second": 11.535,
285
+ "step": 18000
286
+ },
287
+ {
288
+ "epoch": 0.77,
289
+ "learning_rate": 5.275555555555556e-07,
290
+ "loss": 1.0942,
291
+ "step": 19000
292
+ },
293
+ {
294
+ "epoch": 0.77,
295
+ "eval_accuracy": 0.4039920159680639,
296
+ "eval_loss": 1.0809950828552246,
297
+ "eval_runtime": 54.3973,
298
+ "eval_samples_per_second": 92.1,
299
+ "eval_steps_per_second": 11.526,
300
+ "step": 19000
301
+ },
302
+ {
303
+ "epoch": 0.81,
304
+ "learning_rate": 5.553333333333334e-07,
305
+ "loss": 1.0921,
306
+ "step": 20000
307
+ },
308
+ {
309
+ "epoch": 0.81,
310
+ "eval_accuracy": 0.4055888223552894,
311
+ "eval_loss": 1.0818568468093872,
312
+ "eval_runtime": 54.8395,
313
+ "eval_samples_per_second": 91.358,
314
+ "eval_steps_per_second": 11.433,
315
+ "step": 20000
316
+ },
317
+ {
318
+ "epoch": 0.86,
319
+ "learning_rate": 5.830833333333334e-07,
320
+ "loss": 1.0914,
321
+ "step": 21000
322
+ },
323
+ {
324
+ "epoch": 0.86,
325
+ "eval_accuracy": 0.38163672654690617,
326
+ "eval_loss": 1.08183753490448,
327
+ "eval_runtime": 54.4076,
328
+ "eval_samples_per_second": 92.083,
329
+ "eval_steps_per_second": 11.524,
330
+ "step": 21000
331
+ },
332
+ {
333
+ "epoch": 0.9,
334
+ "learning_rate": 6.108611111111111e-07,
335
+ "loss": 1.091,
336
+ "step": 22000
337
+ },
338
+ {
339
+ "epoch": 0.9,
340
+ "eval_accuracy": 0.40658682634730536,
341
+ "eval_loss": 1.0757951736450195,
342
+ "eval_runtime": 54.4113,
343
+ "eval_samples_per_second": 92.077,
344
+ "eval_steps_per_second": 11.523,
345
+ "step": 22000
346
+ },
347
+ {
348
+ "epoch": 0.94,
349
+ "learning_rate": 6.386388888888889e-07,
350
+ "loss": 1.0856,
351
+ "step": 23000
352
+ },
353
+ {
354
+ "epoch": 0.94,
355
+ "eval_accuracy": 0.43313373253493015,
356
+ "eval_loss": 1.0659974813461304,
357
+ "eval_runtime": 54.3958,
358
+ "eval_samples_per_second": 92.103,
359
+ "eval_steps_per_second": 11.527,
360
+ "step": 23000
361
+ },
362
+ {
363
+ "epoch": 0.98,
364
+ "learning_rate": 6.663888888888889e-07,
365
+ "loss": 1.0794,
366
+ "step": 24000
367
+ },
368
+ {
369
+ "epoch": 0.98,
370
+ "eval_accuracy": 0.43952095808383235,
371
+ "eval_loss": 1.056550145149231,
372
+ "eval_runtime": 54.6408,
373
+ "eval_samples_per_second": 91.69,
374
+ "eval_steps_per_second": 11.475,
375
+ "step": 24000
376
+ },
377
+ {
378
+ "epoch": 1.02,
379
+ "learning_rate": 6.941666666666667e-07,
380
+ "loss": 1.0714,
381
+ "step": 25000
382
+ },
383
+ {
384
+ "epoch": 1.02,
385
+ "eval_accuracy": 0.4411177644710579,
386
+ "eval_loss": 1.0531866550445557,
387
+ "eval_runtime": 54.5122,
388
+ "eval_samples_per_second": 91.906,
389
+ "eval_steps_per_second": 11.502,
390
+ "step": 25000
391
+ },
392
+ {
393
+ "epoch": 1.06,
394
+ "learning_rate": 7.219166666666666e-07,
395
+ "loss": 1.0703,
396
+ "step": 26000
397
+ },
398
+ {
399
+ "epoch": 1.06,
400
+ "eval_accuracy": 0.43273453093812375,
401
+ "eval_loss": 1.0532232522964478,
402
+ "eval_runtime": 54.5098,
403
+ "eval_samples_per_second": 91.91,
404
+ "eval_steps_per_second": 11.503,
405
+ "step": 26000
406
+ },
407
+ {
408
+ "epoch": 1.1,
409
+ "learning_rate": 7.496944444444444e-07,
410
+ "loss": 1.0682,
411
+ "step": 27000
412
+ },
413
+ {
414
+ "epoch": 1.1,
415
+ "eval_accuracy": 0.4353293413173653,
416
+ "eval_loss": 1.050220012664795,
417
+ "eval_runtime": 54.3702,
418
+ "eval_samples_per_second": 92.146,
419
+ "eval_steps_per_second": 11.532,
420
+ "step": 27000
421
+ },
422
+ {
423
+ "epoch": 1.14,
424
+ "learning_rate": 7.774444444444445e-07,
425
+ "loss": 1.0671,
426
+ "step": 28000
427
+ },
428
+ {
429
+ "epoch": 1.14,
430
+ "eval_accuracy": 0.43333333333333335,
431
+ "eval_loss": 1.0470824241638184,
432
+ "eval_runtime": 54.4316,
433
+ "eval_samples_per_second": 92.042,
434
+ "eval_steps_per_second": 11.519,
435
+ "step": 28000
436
+ },
437
+ {
438
+ "epoch": 1.18,
439
+ "learning_rate": 8.052222222222223e-07,
440
+ "loss": 1.0584,
441
+ "step": 29000
442
+ },
443
+ {
444
+ "epoch": 1.18,
445
+ "eval_accuracy": 0.43972055888223555,
446
+ "eval_loss": 1.0414146184921265,
447
+ "eval_runtime": 54.4332,
448
+ "eval_samples_per_second": 92.039,
449
+ "eval_steps_per_second": 11.519,
450
+ "step": 29000
451
+ }
452
+ ],
453
+ "logging_steps": 1000,
454
+ "max_steps": 10000000,
455
+ "num_train_epochs": 408,
456
+ "save_steps": 1000,
457
+ "total_flos": 6.06200374862807e+16,
458
+ "trial_name": null,
459
+ "trial_params": null
460
+ }
checkpoint-29000/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:801932cde7104ce385ac1157cb55104d9d1819c965a9990b2a3d508b76c6c359
3
+ size 4091
checkpoint-29000/vocab.json ADDED
The diff for this file is too large to render. See raw diff