tmnam20 commited on
Commit
0595ecf
1 Parent(s): 5ca646e

Training in progress, step 51000, checkpoint

Browse files
checkpoint-51000/added_tokens.json ADDED
@@ -0,0 +1,5 @@
 
 
 
 
 
 
1
+ {
2
+ "<bos>": 40478,
3
+ "<eos>": 40479,
4
+ "<pad>": 40480
5
+ }
checkpoint-51000/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": "last",
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-51000/merges.txt ADDED
The diff for this file is too large to render. See raw diff
 
checkpoint-51000/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c527df24ae3337c755c1d4604c61402516265f19c1c2a727670bc03eba1d3c4b
3
+ size 932401285
checkpoint-51000/pytorch_model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:895df0d4b1b8f5d7a626da0d3d39cf038cdb6a6c2c6a98951f151b3a9ebfd3b2
3
+ size 466205145
checkpoint-51000/rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b1bb99071e41172280a64fff7fbf50e6be32f670a9e3be090dea1f35d91d5fff
3
+ size 14575
checkpoint-51000/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f69207795e3dfd61b31c6c1792032c7341d949373a800c6d6baa267357f4136f
3
+ size 627
checkpoint-51000/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-51000/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-51000/trainer_state.json ADDED
@@ -0,0 +1,790 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": 0.9929116368293762,
3
+ "best_model_checkpoint": "/kaggle/output/checkpoint-51000",
4
+ "epoch": 2.077900912646675,
5
+ "eval_steps": 1000,
6
+ "global_step": 51000,
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": 1.1383,
15
+ "step": 1
16
+ },
17
+ {
18
+ "epoch": 0.04,
19
+ "learning_rate": 2.7750000000000004e-08,
20
+ "loss": 1.1424,
21
+ "step": 1000
22
+ },
23
+ {
24
+ "epoch": 0.04,
25
+ "eval_accuracy": 0.32375249500998005,
26
+ "eval_loss": 1.1077626943588257,
27
+ "eval_runtime": 54.8633,
28
+ "eval_samples_per_second": 91.318,
29
+ "eval_steps_per_second": 11.428,
30
+ "step": 1000
31
+ },
32
+ {
33
+ "epoch": 0.08,
34
+ "learning_rate": 5.5527777777777784e-08,
35
+ "loss": 1.1244,
36
+ "step": 2000
37
+ },
38
+ {
39
+ "epoch": 0.08,
40
+ "eval_accuracy": 0.33652694610778444,
41
+ "eval_loss": 1.1080161333084106,
42
+ "eval_runtime": 54.7384,
43
+ "eval_samples_per_second": 91.526,
44
+ "eval_steps_per_second": 11.454,
45
+ "step": 2000
46
+ },
47
+ {
48
+ "epoch": 0.12,
49
+ "learning_rate": 8.327777777777778e-08,
50
+ "loss": 1.1228,
51
+ "step": 3000
52
+ },
53
+ {
54
+ "epoch": 0.12,
55
+ "eval_accuracy": 0.34331337325349304,
56
+ "eval_loss": 1.1084064245224,
57
+ "eval_runtime": 54.7948,
58
+ "eval_samples_per_second": 91.432,
59
+ "eval_steps_per_second": 11.443,
60
+ "step": 3000
61
+ },
62
+ {
63
+ "epoch": 0.16,
64
+ "learning_rate": 1.1105555555555557e-07,
65
+ "loss": 1.1216,
66
+ "step": 4000
67
+ },
68
+ {
69
+ "epoch": 0.16,
70
+ "eval_accuracy": 0.3385229540918164,
71
+ "eval_loss": 1.1014840602874756,
72
+ "eval_runtime": 54.8508,
73
+ "eval_samples_per_second": 91.339,
74
+ "eval_steps_per_second": 11.431,
75
+ "step": 4000
76
+ },
77
+ {
78
+ "epoch": 0.2,
79
+ "learning_rate": 1.3880555555555558e-07,
80
+ "loss": 1.1181,
81
+ "step": 5000
82
+ },
83
+ {
84
+ "epoch": 0.2,
85
+ "eval_accuracy": 0.33073852295409184,
86
+ "eval_loss": 1.1008135080337524,
87
+ "eval_runtime": 54.8304,
88
+ "eval_samples_per_second": 91.373,
89
+ "eval_steps_per_second": 11.435,
90
+ "step": 5000
91
+ },
92
+ {
93
+ "epoch": 0.24,
94
+ "learning_rate": 1.6658333333333335e-07,
95
+ "loss": 1.1132,
96
+ "step": 6000
97
+ },
98
+ {
99
+ "epoch": 0.24,
100
+ "eval_accuracy": 0.3520958083832335,
101
+ "eval_loss": 1.0993762016296387,
102
+ "eval_runtime": 54.8804,
103
+ "eval_samples_per_second": 91.289,
104
+ "eval_steps_per_second": 11.425,
105
+ "step": 6000
106
+ },
107
+ {
108
+ "epoch": 0.29,
109
+ "learning_rate": 1.9433333333333334e-07,
110
+ "loss": 1.1113,
111
+ "step": 7000
112
+ },
113
+ {
114
+ "epoch": 0.29,
115
+ "eval_accuracy": 0.3530938123752495,
116
+ "eval_loss": 1.0965770483016968,
117
+ "eval_runtime": 54.8881,
118
+ "eval_samples_per_second": 91.277,
119
+ "eval_steps_per_second": 11.423,
120
+ "step": 7000
121
+ },
122
+ {
123
+ "epoch": 0.33,
124
+ "learning_rate": 2.2211111111111114e-07,
125
+ "loss": 1.1111,
126
+ "step": 8000
127
+ },
128
+ {
129
+ "epoch": 0.33,
130
+ "eval_accuracy": 0.35708582834331337,
131
+ "eval_loss": 1.094658613204956,
132
+ "eval_runtime": 54.8233,
133
+ "eval_samples_per_second": 91.384,
134
+ "eval_steps_per_second": 11.437,
135
+ "step": 8000
136
+ },
137
+ {
138
+ "epoch": 0.37,
139
+ "learning_rate": 2.4986111111111113e-07,
140
+ "loss": 1.109,
141
+ "step": 9000
142
+ },
143
+ {
144
+ "epoch": 0.37,
145
+ "eval_accuracy": 0.34191616766467064,
146
+ "eval_loss": 1.106990933418274,
147
+ "eval_runtime": 54.9095,
148
+ "eval_samples_per_second": 91.241,
149
+ "eval_steps_per_second": 11.419,
150
+ "step": 9000
151
+ },
152
+ {
153
+ "epoch": 0.41,
154
+ "learning_rate": 2.776388888888889e-07,
155
+ "loss": 1.1036,
156
+ "step": 10000
157
+ },
158
+ {
159
+ "epoch": 0.41,
160
+ "eval_accuracy": 0.37584830339321357,
161
+ "eval_loss": 1.0930211544036865,
162
+ "eval_runtime": 54.9067,
163
+ "eval_samples_per_second": 91.246,
164
+ "eval_steps_per_second": 11.419,
165
+ "step": 10000
166
+ },
167
+ {
168
+ "epoch": 0.45,
169
+ "learning_rate": 3.0541666666666667e-07,
170
+ "loss": 1.1045,
171
+ "step": 11000
172
+ },
173
+ {
174
+ "epoch": 0.45,
175
+ "eval_accuracy": 0.3652694610778443,
176
+ "eval_loss": 1.092846393585205,
177
+ "eval_runtime": 54.8964,
178
+ "eval_samples_per_second": 91.263,
179
+ "eval_steps_per_second": 11.422,
180
+ "step": 11000
181
+ },
182
+ {
183
+ "epoch": 0.49,
184
+ "learning_rate": 3.3319444444444444e-07,
185
+ "loss": 1.1024,
186
+ "step": 12000
187
+ },
188
+ {
189
+ "epoch": 0.49,
190
+ "eval_accuracy": 0.39261477045908183,
191
+ "eval_loss": 1.089038372039795,
192
+ "eval_runtime": 54.9763,
193
+ "eval_samples_per_second": 91.13,
194
+ "eval_steps_per_second": 11.405,
195
+ "step": 12000
196
+ },
197
+ {
198
+ "epoch": 0.53,
199
+ "learning_rate": 3.6094444444444446e-07,
200
+ "loss": 1.1007,
201
+ "step": 13000
202
+ },
203
+ {
204
+ "epoch": 0.53,
205
+ "eval_accuracy": 0.34311377245508984,
206
+ "eval_loss": 1.0933948755264282,
207
+ "eval_runtime": 54.9285,
208
+ "eval_samples_per_second": 91.209,
209
+ "eval_steps_per_second": 11.415,
210
+ "step": 13000
211
+ },
212
+ {
213
+ "epoch": 0.57,
214
+ "learning_rate": 3.8872222222222223e-07,
215
+ "loss": 1.0985,
216
+ "step": 14000
217
+ },
218
+ {
219
+ "epoch": 0.57,
220
+ "eval_accuracy": 0.36367265469061877,
221
+ "eval_loss": 1.09434974193573,
222
+ "eval_runtime": 54.8032,
223
+ "eval_samples_per_second": 91.418,
224
+ "eval_steps_per_second": 11.441,
225
+ "step": 14000
226
+ },
227
+ {
228
+ "epoch": 0.61,
229
+ "learning_rate": 4.1650000000000006e-07,
230
+ "loss": 1.0988,
231
+ "step": 15000
232
+ },
233
+ {
234
+ "epoch": 0.61,
235
+ "eval_accuracy": 0.39481037924151696,
236
+ "eval_loss": 1.0886671543121338,
237
+ "eval_runtime": 54.9221,
238
+ "eval_samples_per_second": 91.22,
239
+ "eval_steps_per_second": 11.416,
240
+ "step": 15000
241
+ },
242
+ {
243
+ "epoch": 0.65,
244
+ "learning_rate": 4.4425e-07,
245
+ "loss": 1.0965,
246
+ "step": 16000
247
+ },
248
+ {
249
+ "epoch": 0.65,
250
+ "eval_accuracy": 0.3916167664670659,
251
+ "eval_loss": 1.0834949016571045,
252
+ "eval_runtime": 54.5628,
253
+ "eval_samples_per_second": 91.821,
254
+ "eval_steps_per_second": 11.491,
255
+ "step": 16000
256
+ },
257
+ {
258
+ "epoch": 0.69,
259
+ "learning_rate": 4.7202777777777785e-07,
260
+ "loss": 1.0926,
261
+ "step": 17000
262
+ },
263
+ {
264
+ "epoch": 0.69,
265
+ "eval_accuracy": 0.4239520958083832,
266
+ "eval_loss": 1.079688310623169,
267
+ "eval_runtime": 54.6989,
268
+ "eval_samples_per_second": 91.592,
269
+ "eval_steps_per_second": 11.463,
270
+ "step": 17000
271
+ },
272
+ {
273
+ "epoch": 0.73,
274
+ "learning_rate": 4.998055555555556e-07,
275
+ "loss": 1.0956,
276
+ "step": 18000
277
+ },
278
+ {
279
+ "epoch": 0.73,
280
+ "eval_accuracy": 0.4219560878243513,
281
+ "eval_loss": 1.080493688583374,
282
+ "eval_runtime": 54.6863,
283
+ "eval_samples_per_second": 91.613,
284
+ "eval_steps_per_second": 11.465,
285
+ "step": 18000
286
+ },
287
+ {
288
+ "epoch": 0.77,
289
+ "learning_rate": 5.275555555555556e-07,
290
+ "loss": 1.0878,
291
+ "step": 19000
292
+ },
293
+ {
294
+ "epoch": 0.77,
295
+ "eval_accuracy": 0.4343313373253493,
296
+ "eval_loss": 1.0664235353469849,
297
+ "eval_runtime": 54.7843,
298
+ "eval_samples_per_second": 91.45,
299
+ "eval_steps_per_second": 11.445,
300
+ "step": 19000
301
+ },
302
+ {
303
+ "epoch": 0.81,
304
+ "learning_rate": 5.553333333333334e-07,
305
+ "loss": 1.0793,
306
+ "step": 20000
307
+ },
308
+ {
309
+ "epoch": 0.81,
310
+ "eval_accuracy": 0.4365269461077844,
311
+ "eval_loss": 1.06390380859375,
312
+ "eval_runtime": 54.7978,
313
+ "eval_samples_per_second": 91.427,
314
+ "eval_steps_per_second": 11.442,
315
+ "step": 20000
316
+ },
317
+ {
318
+ "epoch": 0.86,
319
+ "learning_rate": 5.830833333333334e-07,
320
+ "loss": 1.0746,
321
+ "step": 21000
322
+ },
323
+ {
324
+ "epoch": 0.86,
325
+ "eval_accuracy": 0.4311377245508982,
326
+ "eval_loss": 1.0611063241958618,
327
+ "eval_runtime": 54.6084,
328
+ "eval_samples_per_second": 91.744,
329
+ "eval_steps_per_second": 11.482,
330
+ "step": 21000
331
+ },
332
+ {
333
+ "epoch": 0.9,
334
+ "learning_rate": 6.108611111111111e-07,
335
+ "loss": 1.0757,
336
+ "step": 22000
337
+ },
338
+ {
339
+ "epoch": 0.9,
340
+ "eval_accuracy": 0.43253493013972055,
341
+ "eval_loss": 1.0579031705856323,
342
+ "eval_runtime": 54.7147,
343
+ "eval_samples_per_second": 91.566,
344
+ "eval_steps_per_second": 11.459,
345
+ "step": 22000
346
+ },
347
+ {
348
+ "epoch": 0.94,
349
+ "learning_rate": 6.386111111111112e-07,
350
+ "loss": 1.0712,
351
+ "step": 23000
352
+ },
353
+ {
354
+ "epoch": 0.94,
355
+ "eval_accuracy": 0.43353293413173655,
356
+ "eval_loss": 1.0545520782470703,
357
+ "eval_runtime": 54.8205,
358
+ "eval_samples_per_second": 91.389,
359
+ "eval_steps_per_second": 11.437,
360
+ "step": 23000
361
+ },
362
+ {
363
+ "epoch": 0.98,
364
+ "learning_rate": 6.663888888888889e-07,
365
+ "loss": 1.0703,
366
+ "step": 24000
367
+ },
368
+ {
369
+ "epoch": 0.98,
370
+ "eval_accuracy": 0.43872255489021955,
371
+ "eval_loss": 1.0489881038665771,
372
+ "eval_runtime": 54.6899,
373
+ "eval_samples_per_second": 91.607,
374
+ "eval_steps_per_second": 11.465,
375
+ "step": 24000
376
+ },
377
+ {
378
+ "epoch": 1.02,
379
+ "learning_rate": 6.941666666666667e-07,
380
+ "loss": 1.0673,
381
+ "step": 25000
382
+ },
383
+ {
384
+ "epoch": 1.02,
385
+ "eval_accuracy": 0.4467065868263473,
386
+ "eval_loss": 1.0496938228607178,
387
+ "eval_runtime": 54.8292,
388
+ "eval_samples_per_second": 91.375,
389
+ "eval_steps_per_second": 11.436,
390
+ "step": 25000
391
+ },
392
+ {
393
+ "epoch": 1.06,
394
+ "learning_rate": 7.219166666666666e-07,
395
+ "loss": 1.0672,
396
+ "step": 26000
397
+ },
398
+ {
399
+ "epoch": 1.06,
400
+ "eval_accuracy": 0.4297405189620758,
401
+ "eval_loss": 1.0511012077331543,
402
+ "eval_runtime": 54.8046,
403
+ "eval_samples_per_second": 91.416,
404
+ "eval_steps_per_second": 11.441,
405
+ "step": 26000
406
+ },
407
+ {
408
+ "epoch": 1.1,
409
+ "learning_rate": 7.496944444444444e-07,
410
+ "loss": 1.0658,
411
+ "step": 27000
412
+ },
413
+ {
414
+ "epoch": 1.1,
415
+ "eval_accuracy": 0.43912175648702595,
416
+ "eval_loss": 1.0465787649154663,
417
+ "eval_runtime": 54.7507,
418
+ "eval_samples_per_second": 91.506,
419
+ "eval_steps_per_second": 11.452,
420
+ "step": 27000
421
+ },
422
+ {
423
+ "epoch": 1.14,
424
+ "learning_rate": 7.774444444444445e-07,
425
+ "loss": 1.0638,
426
+ "step": 28000
427
+ },
428
+ {
429
+ "epoch": 1.14,
430
+ "eval_accuracy": 0.43333333333333335,
431
+ "eval_loss": 1.0430774688720703,
432
+ "eval_runtime": 54.6439,
433
+ "eval_samples_per_second": 91.685,
434
+ "eval_steps_per_second": 11.474,
435
+ "step": 28000
436
+ },
437
+ {
438
+ "epoch": 1.18,
439
+ "learning_rate": 8.052222222222223e-07,
440
+ "loss": 1.0602,
441
+ "step": 29000
442
+ },
443
+ {
444
+ "epoch": 1.18,
445
+ "eval_accuracy": 0.4479041916167665,
446
+ "eval_loss": 1.0387687683105469,
447
+ "eval_runtime": 54.8009,
448
+ "eval_samples_per_second": 91.422,
449
+ "eval_steps_per_second": 11.441,
450
+ "step": 29000
451
+ },
452
+ {
453
+ "epoch": 1.22,
454
+ "learning_rate": 8.329722222222223e-07,
455
+ "loss": 1.0567,
456
+ "step": 30000
457
+ },
458
+ {
459
+ "epoch": 1.22,
460
+ "eval_accuracy": 0.4485029940119761,
461
+ "eval_loss": 1.0339411497116089,
462
+ "eval_runtime": 54.6621,
463
+ "eval_samples_per_second": 91.654,
464
+ "eval_steps_per_second": 11.47,
465
+ "step": 30000
466
+ },
467
+ {
468
+ "epoch": 1.26,
469
+ "learning_rate": 8.607500000000001e-07,
470
+ "loss": 1.0611,
471
+ "step": 31000
472
+ },
473
+ {
474
+ "epoch": 1.26,
475
+ "eval_accuracy": 0.4465069860279441,
476
+ "eval_loss": 1.0385600328445435,
477
+ "eval_runtime": 54.7106,
478
+ "eval_samples_per_second": 91.573,
479
+ "eval_steps_per_second": 11.46,
480
+ "step": 31000
481
+ },
482
+ {
483
+ "epoch": 1.3,
484
+ "learning_rate": 8.885e-07,
485
+ "loss": 1.0555,
486
+ "step": 32000
487
+ },
488
+ {
489
+ "epoch": 1.3,
490
+ "eval_accuracy": 0.4377245508982036,
491
+ "eval_loss": 1.0331332683563232,
492
+ "eval_runtime": 54.7554,
493
+ "eval_samples_per_second": 91.498,
494
+ "eval_steps_per_second": 11.451,
495
+ "step": 32000
496
+ },
497
+ {
498
+ "epoch": 1.34,
499
+ "learning_rate": 9.162777777777779e-07,
500
+ "loss": 1.0512,
501
+ "step": 33000
502
+ },
503
+ {
504
+ "epoch": 1.34,
505
+ "eval_accuracy": 0.44610778443113774,
506
+ "eval_loss": 1.028577446937561,
507
+ "eval_runtime": 54.7331,
508
+ "eval_samples_per_second": 91.535,
509
+ "eval_steps_per_second": 11.456,
510
+ "step": 33000
511
+ },
512
+ {
513
+ "epoch": 1.39,
514
+ "learning_rate": 9.440277777777779e-07,
515
+ "loss": 1.048,
516
+ "step": 34000
517
+ },
518
+ {
519
+ "epoch": 1.39,
520
+ "eval_accuracy": 0.4532934131736527,
521
+ "eval_loss": 1.0286486148834229,
522
+ "eval_runtime": 54.7585,
523
+ "eval_samples_per_second": 91.493,
524
+ "eval_steps_per_second": 11.45,
525
+ "step": 34000
526
+ },
527
+ {
528
+ "epoch": 1.43,
529
+ "learning_rate": 9.718055555555557e-07,
530
+ "loss": 1.0524,
531
+ "step": 35000
532
+ },
533
+ {
534
+ "epoch": 1.43,
535
+ "eval_accuracy": 0.449500998003992,
536
+ "eval_loss": 1.0262507200241089,
537
+ "eval_runtime": 54.1381,
538
+ "eval_samples_per_second": 92.541,
539
+ "eval_steps_per_second": 11.582,
540
+ "step": 35000
541
+ },
542
+ {
543
+ "epoch": 1.47,
544
+ "learning_rate": 9.995555555555557e-07,
545
+ "loss": 1.0472,
546
+ "step": 36000
547
+ },
548
+ {
549
+ "epoch": 1.47,
550
+ "eval_accuracy": 0.4552894211576846,
551
+ "eval_loss": 1.022876501083374,
552
+ "eval_runtime": 54.2077,
553
+ "eval_samples_per_second": 92.422,
554
+ "eval_steps_per_second": 11.567,
555
+ "step": 36000
556
+ },
557
+ {
558
+ "epoch": 1.51,
559
+ "learning_rate": 1.0273333333333335e-06,
560
+ "loss": 1.0454,
561
+ "step": 37000
562
+ },
563
+ {
564
+ "epoch": 1.51,
565
+ "eval_accuracy": 0.4477045908183633,
566
+ "eval_loss": 1.0219467878341675,
567
+ "eval_runtime": 54.4001,
568
+ "eval_samples_per_second": 92.095,
569
+ "eval_steps_per_second": 11.526,
570
+ "step": 37000
571
+ },
572
+ {
573
+ "epoch": 1.55,
574
+ "learning_rate": 1.0550833333333334e-06,
575
+ "loss": 1.0473,
576
+ "step": 38000
577
+ },
578
+ {
579
+ "epoch": 1.55,
580
+ "eval_accuracy": 0.46107784431137727,
581
+ "eval_loss": 1.0189749002456665,
582
+ "eval_runtime": 54.291,
583
+ "eval_samples_per_second": 92.28,
584
+ "eval_steps_per_second": 11.549,
585
+ "step": 38000
586
+ },
587
+ {
588
+ "epoch": 1.59,
589
+ "learning_rate": 1.0828611111111111e-06,
590
+ "loss": 1.0465,
591
+ "step": 39000
592
+ },
593
+ {
594
+ "epoch": 1.59,
595
+ "eval_accuracy": 0.4500998003992016,
596
+ "eval_loss": 1.0226292610168457,
597
+ "eval_runtime": 54.6634,
598
+ "eval_samples_per_second": 91.652,
599
+ "eval_steps_per_second": 11.47,
600
+ "step": 39000
601
+ },
602
+ {
603
+ "epoch": 1.63,
604
+ "learning_rate": 1.1106111111111112e-06,
605
+ "loss": 1.0408,
606
+ "step": 40000
607
+ },
608
+ {
609
+ "epoch": 1.63,
610
+ "eval_accuracy": 0.45229540918163674,
611
+ "eval_loss": 1.0191301107406616,
612
+ "eval_runtime": 54.6371,
613
+ "eval_samples_per_second": 91.696,
614
+ "eval_steps_per_second": 11.476,
615
+ "step": 40000
616
+ },
617
+ {
618
+ "epoch": 1.67,
619
+ "learning_rate": 1.138388888888889e-06,
620
+ "loss": 1.0433,
621
+ "step": 41000
622
+ },
623
+ {
624
+ "epoch": 1.67,
625
+ "eval_accuracy": 0.4620758483033932,
626
+ "eval_loss": 1.0231131315231323,
627
+ "eval_runtime": 54.2747,
628
+ "eval_samples_per_second": 92.308,
629
+ "eval_steps_per_second": 11.552,
630
+ "step": 41000
631
+ },
632
+ {
633
+ "epoch": 1.71,
634
+ "learning_rate": 1.166138888888889e-06,
635
+ "loss": 1.0392,
636
+ "step": 42000
637
+ },
638
+ {
639
+ "epoch": 1.71,
640
+ "eval_accuracy": 0.46726546906187627,
641
+ "eval_loss": 1.0097001791000366,
642
+ "eval_runtime": 54.4967,
643
+ "eval_samples_per_second": 91.932,
644
+ "eval_steps_per_second": 11.505,
645
+ "step": 42000
646
+ },
647
+ {
648
+ "epoch": 1.75,
649
+ "learning_rate": 1.1939166666666668e-06,
650
+ "loss": 1.0381,
651
+ "step": 43000
652
+ },
653
+ {
654
+ "epoch": 1.75,
655
+ "eval_accuracy": 0.4483033932135729,
656
+ "eval_loss": 1.032842755317688,
657
+ "eval_runtime": 54.2658,
658
+ "eval_samples_per_second": 92.323,
659
+ "eval_steps_per_second": 11.554,
660
+ "step": 43000
661
+ },
662
+ {
663
+ "epoch": 1.79,
664
+ "learning_rate": 1.2216944444444446e-06,
665
+ "loss": 1.0401,
666
+ "step": 44000
667
+ },
668
+ {
669
+ "epoch": 1.79,
670
+ "eval_accuracy": 0.4806387225548902,
671
+ "eval_loss": 1.0062618255615234,
672
+ "eval_runtime": 54.1855,
673
+ "eval_samples_per_second": 92.46,
674
+ "eval_steps_per_second": 11.571,
675
+ "step": 44000
676
+ },
677
+ {
678
+ "epoch": 1.83,
679
+ "learning_rate": 1.2494444444444447e-06,
680
+ "loss": 1.0366,
681
+ "step": 45000
682
+ },
683
+ {
684
+ "epoch": 1.83,
685
+ "eval_accuracy": 0.46966067864271455,
686
+ "eval_loss": 1.0111383199691772,
687
+ "eval_runtime": 54.4723,
688
+ "eval_samples_per_second": 91.973,
689
+ "eval_steps_per_second": 11.51,
690
+ "step": 45000
691
+ },
692
+ {
693
+ "epoch": 1.87,
694
+ "learning_rate": 1.2772222222222224e-06,
695
+ "loss": 1.0365,
696
+ "step": 46000
697
+ },
698
+ {
699
+ "epoch": 1.87,
700
+ "eval_accuracy": 0.47584830339321355,
701
+ "eval_loss": 1.0140398740768433,
702
+ "eval_runtime": 54.4788,
703
+ "eval_samples_per_second": 91.962,
704
+ "eval_steps_per_second": 11.509,
705
+ "step": 46000
706
+ },
707
+ {
708
+ "epoch": 1.91,
709
+ "learning_rate": 1.3049722222222223e-06,
710
+ "loss": 1.0299,
711
+ "step": 47000
712
+ },
713
+ {
714
+ "epoch": 1.91,
715
+ "eval_accuracy": 0.4784431137724551,
716
+ "eval_loss": 1.0014474391937256,
717
+ "eval_runtime": 54.2653,
718
+ "eval_samples_per_second": 92.324,
719
+ "eval_steps_per_second": 11.554,
720
+ "step": 47000
721
+ },
722
+ {
723
+ "epoch": 1.96,
724
+ "learning_rate": 1.33275e-06,
725
+ "loss": 1.0342,
726
+ "step": 48000
727
+ },
728
+ {
729
+ "epoch": 1.96,
730
+ "eval_accuracy": 0.47744510978043914,
731
+ "eval_loss": 1.0174448490142822,
732
+ "eval_runtime": 54.2558,
733
+ "eval_samples_per_second": 92.34,
734
+ "eval_steps_per_second": 11.556,
735
+ "step": 48000
736
+ },
737
+ {
738
+ "epoch": 2.0,
739
+ "learning_rate": 1.3605000000000001e-06,
740
+ "loss": 1.028,
741
+ "step": 49000
742
+ },
743
+ {
744
+ "epoch": 2.0,
745
+ "eval_accuracy": 0.49860279441117766,
746
+ "eval_loss": 0.9931387305259705,
747
+ "eval_runtime": 54.9876,
748
+ "eval_samples_per_second": 91.111,
749
+ "eval_steps_per_second": 11.403,
750
+ "step": 49000
751
+ },
752
+ {
753
+ "epoch": 2.04,
754
+ "learning_rate": 1.388277777777778e-06,
755
+ "loss": 1.0289,
756
+ "step": 50000
757
+ },
758
+ {
759
+ "epoch": 2.04,
760
+ "eval_accuracy": 0.48542914171656687,
761
+ "eval_loss": 0.9991692304611206,
762
+ "eval_runtime": 54.7097,
763
+ "eval_samples_per_second": 91.574,
764
+ "eval_steps_per_second": 11.46,
765
+ "step": 50000
766
+ },
767
+ {
768
+ "epoch": 2.08,
769
+ "learning_rate": 1.416027777777778e-06,
770
+ "loss": 1.0267,
771
+ "step": 51000
772
+ },
773
+ {
774
+ "epoch": 2.08,
775
+ "eval_accuracy": 0.49481037924151694,
776
+ "eval_loss": 0.9929116368293762,
777
+ "eval_runtime": 54.2323,
778
+ "eval_samples_per_second": 92.38,
779
+ "eval_steps_per_second": 11.561,
780
+ "step": 51000
781
+ }
782
+ ],
783
+ "logging_steps": 1000,
784
+ "max_steps": 10000000,
785
+ "num_train_epochs": 408,
786
+ "save_steps": 1000,
787
+ "total_flos": 1.066075890599854e+17,
788
+ "trial_name": null,
789
+ "trial_params": null
790
+ }
checkpoint-51000/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a9a52921b15d7eeb53b158a9c0f943d88288ccd9232004774ffdf3a600a048d5
3
+ size 4091
checkpoint-51000/vocab.json ADDED
The diff for this file is too large to render. See raw diff