tmnam20 commited on
Commit
3217983
1 Parent(s): 6c98e25

Training in progress, step 53000, checkpoint

Browse files
checkpoint-53000/added_tokens.json ADDED
@@ -0,0 +1,5 @@
 
 
 
 
 
 
1
+ {
2
+ "<bos>": 40478,
3
+ "<eos>": 40479,
4
+ "<pad>": 40480
5
+ }
checkpoint-53000/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": 2,
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-53000/merges.txt ADDED
The diff for this file is too large to render. See raw diff
 
checkpoint-53000/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d5768c90c1301387646af86a2c7b26a121187225b18ae6237d4da863e0e4046f
3
+ size 365301235
checkpoint-53000/pytorch_model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8e14830791388ea07c9928ee5cb07c005add72378a586b4c4362f85f11f18821
3
+ size 182651202
checkpoint-53000/rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:52fb288030dafa0ca6a0f98d98e6c16ae9824045cbe6fe6b41d31d1cd7e6ef46
3
+ size 14575
checkpoint-53000/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d6d04945dbc1e13d56d76c69f88e16a6169241ee902ce2b25d594920da79a9f6
3
+ size 627
checkpoint-53000/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-53000/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-53000/trainer_state.json ADDED
@@ -0,0 +1,820 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": 1.010857343673706,
3
+ "best_model_checkpoint": "/kaggle/output/checkpoint-53000",
4
+ "epoch": 2.159387222946545,
5
+ "eval_steps": 1000,
6
+ "global_step": 53000,
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": 1.1078,
15
+ "step": 1
16
+ },
17
+ {
18
+ "epoch": 0.04,
19
+ "learning_rate": 2.7750000000000004e-08,
20
+ "loss": 1.128,
21
+ "step": 1000
22
+ },
23
+ {
24
+ "epoch": 0.04,
25
+ "eval_accuracy": 0.330938123752495,
26
+ "eval_loss": 1.1083118915557861,
27
+ "eval_runtime": 12.3979,
28
+ "eval_samples_per_second": 404.1,
29
+ "eval_steps_per_second": 50.573,
30
+ "step": 1000
31
+ },
32
+ {
33
+ "epoch": 0.08,
34
+ "learning_rate": 5.5527777777777784e-08,
35
+ "loss": 1.1164,
36
+ "step": 2000
37
+ },
38
+ {
39
+ "epoch": 0.08,
40
+ "eval_accuracy": 0.3323353293413174,
41
+ "eval_loss": 1.102672815322876,
42
+ "eval_runtime": 12.2861,
43
+ "eval_samples_per_second": 407.778,
44
+ "eval_steps_per_second": 51.033,
45
+ "step": 2000
46
+ },
47
+ {
48
+ "epoch": 0.12,
49
+ "learning_rate": 8.327777777777778e-08,
50
+ "loss": 1.1158,
51
+ "step": 3000
52
+ },
53
+ {
54
+ "epoch": 0.12,
55
+ "eval_accuracy": 0.34211576846307384,
56
+ "eval_loss": 1.1024774312973022,
57
+ "eval_runtime": 12.3746,
58
+ "eval_samples_per_second": 404.863,
59
+ "eval_steps_per_second": 50.668,
60
+ "step": 3000
61
+ },
62
+ {
63
+ "epoch": 0.16,
64
+ "learning_rate": 1.1105555555555557e-07,
65
+ "loss": 1.1115,
66
+ "step": 4000
67
+ },
68
+ {
69
+ "epoch": 0.16,
70
+ "eval_accuracy": 0.34890219560878244,
71
+ "eval_loss": 1.1012382507324219,
72
+ "eval_runtime": 12.5405,
73
+ "eval_samples_per_second": 399.505,
74
+ "eval_steps_per_second": 49.998,
75
+ "step": 4000
76
+ },
77
+ {
78
+ "epoch": 0.2,
79
+ "learning_rate": 1.3880555555555558e-07,
80
+ "loss": 1.1141,
81
+ "step": 5000
82
+ },
83
+ {
84
+ "epoch": 0.2,
85
+ "eval_accuracy": 0.3534930139720559,
86
+ "eval_loss": 1.1002192497253418,
87
+ "eval_runtime": 12.3001,
88
+ "eval_samples_per_second": 407.314,
89
+ "eval_steps_per_second": 50.975,
90
+ "step": 5000
91
+ },
92
+ {
93
+ "epoch": 0.24,
94
+ "learning_rate": 1.6658333333333335e-07,
95
+ "loss": 1.112,
96
+ "step": 6000
97
+ },
98
+ {
99
+ "epoch": 0.24,
100
+ "eval_accuracy": 0.3524950099800399,
101
+ "eval_loss": 1.100056767463684,
102
+ "eval_runtime": 12.4062,
103
+ "eval_samples_per_second": 403.831,
104
+ "eval_steps_per_second": 50.539,
105
+ "step": 6000
106
+ },
107
+ {
108
+ "epoch": 0.29,
109
+ "learning_rate": 1.9433333333333334e-07,
110
+ "loss": 1.1115,
111
+ "step": 7000
112
+ },
113
+ {
114
+ "epoch": 0.29,
115
+ "eval_accuracy": 0.34331337325349304,
116
+ "eval_loss": 1.098986029624939,
117
+ "eval_runtime": 12.4456,
118
+ "eval_samples_per_second": 402.551,
119
+ "eval_steps_per_second": 50.379,
120
+ "step": 7000
121
+ },
122
+ {
123
+ "epoch": 0.33,
124
+ "learning_rate": 2.2211111111111114e-07,
125
+ "loss": 1.109,
126
+ "step": 8000
127
+ },
128
+ {
129
+ "epoch": 0.33,
130
+ "eval_accuracy": 0.3347305389221557,
131
+ "eval_loss": 1.098678469657898,
132
+ "eval_runtime": 12.5978,
133
+ "eval_samples_per_second": 397.688,
134
+ "eval_steps_per_second": 49.771,
135
+ "step": 8000
136
+ },
137
+ {
138
+ "epoch": 0.37,
139
+ "learning_rate": 2.4986111111111113e-07,
140
+ "loss": 1.1117,
141
+ "step": 9000
142
+ },
143
+ {
144
+ "epoch": 0.37,
145
+ "eval_accuracy": 0.34311377245508984,
146
+ "eval_loss": 1.0984419584274292,
147
+ "eval_runtime": 12.313,
148
+ "eval_samples_per_second": 406.886,
149
+ "eval_steps_per_second": 50.922,
150
+ "step": 9000
151
+ },
152
+ {
153
+ "epoch": 0.41,
154
+ "learning_rate": 2.776388888888889e-07,
155
+ "loss": 1.1081,
156
+ "step": 10000
157
+ },
158
+ {
159
+ "epoch": 0.41,
160
+ "eval_accuracy": 0.35429141716566864,
161
+ "eval_loss": 1.0968743562698364,
162
+ "eval_runtime": 12.4088,
163
+ "eval_samples_per_second": 403.744,
164
+ "eval_steps_per_second": 50.528,
165
+ "step": 10000
166
+ },
167
+ {
168
+ "epoch": 0.45,
169
+ "learning_rate": 3.053888888888889e-07,
170
+ "loss": 1.107,
171
+ "step": 11000
172
+ },
173
+ {
174
+ "epoch": 0.45,
175
+ "eval_accuracy": 0.3722554890219561,
176
+ "eval_loss": 1.0950665473937988,
177
+ "eval_runtime": 12.5775,
178
+ "eval_samples_per_second": 398.33,
179
+ "eval_steps_per_second": 49.851,
180
+ "step": 11000
181
+ },
182
+ {
183
+ "epoch": 0.49,
184
+ "learning_rate": 3.331666666666667e-07,
185
+ "loss": 1.1057,
186
+ "step": 12000
187
+ },
188
+ {
189
+ "epoch": 0.49,
190
+ "eval_accuracy": 0.37544910179640717,
191
+ "eval_loss": 1.0940738916397095,
192
+ "eval_runtime": 12.5186,
193
+ "eval_samples_per_second": 400.203,
194
+ "eval_steps_per_second": 50.085,
195
+ "step": 12000
196
+ },
197
+ {
198
+ "epoch": 0.53,
199
+ "learning_rate": 3.609166666666667e-07,
200
+ "loss": 1.1069,
201
+ "step": 13000
202
+ },
203
+ {
204
+ "epoch": 0.53,
205
+ "eval_accuracy": 0.3401197604790419,
206
+ "eval_loss": 1.0937162637710571,
207
+ "eval_runtime": 12.2594,
208
+ "eval_samples_per_second": 408.666,
209
+ "eval_steps_per_second": 51.144,
210
+ "step": 13000
211
+ },
212
+ {
213
+ "epoch": 0.57,
214
+ "learning_rate": 3.886944444444445e-07,
215
+ "loss": 1.1046,
216
+ "step": 14000
217
+ },
218
+ {
219
+ "epoch": 0.57,
220
+ "eval_accuracy": 0.37604790419161677,
221
+ "eval_loss": 1.0926916599273682,
222
+ "eval_runtime": 12.256,
223
+ "eval_samples_per_second": 408.78,
224
+ "eval_steps_per_second": 51.159,
225
+ "step": 14000
226
+ },
227
+ {
228
+ "epoch": 0.61,
229
+ "learning_rate": 4.164444444444445e-07,
230
+ "loss": 1.1046,
231
+ "step": 15000
232
+ },
233
+ {
234
+ "epoch": 0.61,
235
+ "eval_accuracy": 0.39201596806387223,
236
+ "eval_loss": 1.0912976264953613,
237
+ "eval_runtime": 12.5019,
238
+ "eval_samples_per_second": 400.74,
239
+ "eval_steps_per_second": 50.152,
240
+ "step": 15000
241
+ },
242
+ {
243
+ "epoch": 0.65,
244
+ "learning_rate": 4.442222222222223e-07,
245
+ "loss": 1.1052,
246
+ "step": 16000
247
+ },
248
+ {
249
+ "epoch": 0.65,
250
+ "eval_accuracy": 0.393812375249501,
251
+ "eval_loss": 1.0901482105255127,
252
+ "eval_runtime": 12.3407,
253
+ "eval_samples_per_second": 405.973,
254
+ "eval_steps_per_second": 50.807,
255
+ "step": 16000
256
+ },
257
+ {
258
+ "epoch": 0.69,
259
+ "learning_rate": 4.7197222222222224e-07,
260
+ "loss": 1.1011,
261
+ "step": 17000
262
+ },
263
+ {
264
+ "epoch": 0.69,
265
+ "eval_accuracy": 0.3932135728542914,
266
+ "eval_loss": 1.0901678800582886,
267
+ "eval_runtime": 12.3644,
268
+ "eval_samples_per_second": 405.197,
269
+ "eval_steps_per_second": 50.71,
270
+ "step": 17000
271
+ },
272
+ {
273
+ "epoch": 0.73,
274
+ "learning_rate": 4.997500000000001e-07,
275
+ "loss": 1.1,
276
+ "step": 18000
277
+ },
278
+ {
279
+ "epoch": 0.73,
280
+ "eval_accuracy": 0.38143712574850297,
281
+ "eval_loss": 1.0897998809814453,
282
+ "eval_runtime": 12.3516,
283
+ "eval_samples_per_second": 405.615,
284
+ "eval_steps_per_second": 50.763,
285
+ "step": 18000
286
+ },
287
+ {
288
+ "epoch": 0.77,
289
+ "learning_rate": 5.275277777777778e-07,
290
+ "loss": 1.1007,
291
+ "step": 19000
292
+ },
293
+ {
294
+ "epoch": 0.77,
295
+ "eval_accuracy": 0.40678642714570856,
296
+ "eval_loss": 1.0876121520996094,
297
+ "eval_runtime": 12.6608,
298
+ "eval_samples_per_second": 395.708,
299
+ "eval_steps_per_second": 49.523,
300
+ "step": 19000
301
+ },
302
+ {
303
+ "epoch": 0.81,
304
+ "learning_rate": 5.552777777777778e-07,
305
+ "loss": 1.1013,
306
+ "step": 20000
307
+ },
308
+ {
309
+ "epoch": 0.81,
310
+ "eval_accuracy": 0.4053892215568862,
311
+ "eval_loss": 1.0876350402832031,
312
+ "eval_runtime": 12.3243,
313
+ "eval_samples_per_second": 406.513,
314
+ "eval_steps_per_second": 50.875,
315
+ "step": 20000
316
+ },
317
+ {
318
+ "epoch": 0.86,
319
+ "learning_rate": 5.830555555555556e-07,
320
+ "loss": 1.0987,
321
+ "step": 21000
322
+ },
323
+ {
324
+ "epoch": 0.86,
325
+ "eval_accuracy": 0.40858283433133735,
326
+ "eval_loss": 1.085940957069397,
327
+ "eval_runtime": 12.3217,
328
+ "eval_samples_per_second": 406.598,
329
+ "eval_steps_per_second": 50.886,
330
+ "step": 21000
331
+ },
332
+ {
333
+ "epoch": 0.9,
334
+ "learning_rate": 6.108055555555556e-07,
335
+ "loss": 1.0986,
336
+ "step": 22000
337
+ },
338
+ {
339
+ "epoch": 0.9,
340
+ "eval_accuracy": 0.3972055888223553,
341
+ "eval_loss": 1.085978388786316,
342
+ "eval_runtime": 12.4285,
343
+ "eval_samples_per_second": 403.107,
344
+ "eval_steps_per_second": 50.449,
345
+ "step": 22000
346
+ },
347
+ {
348
+ "epoch": 0.94,
349
+ "learning_rate": 6.385833333333334e-07,
350
+ "loss": 1.0944,
351
+ "step": 23000
352
+ },
353
+ {
354
+ "epoch": 0.94,
355
+ "eval_accuracy": 0.41097804391217563,
356
+ "eval_loss": 1.0837496519088745,
357
+ "eval_runtime": 12.7175,
358
+ "eval_samples_per_second": 393.946,
359
+ "eval_steps_per_second": 49.302,
360
+ "step": 23000
361
+ },
362
+ {
363
+ "epoch": 0.98,
364
+ "learning_rate": 6.663611111111112e-07,
365
+ "loss": 1.095,
366
+ "step": 24000
367
+ },
368
+ {
369
+ "epoch": 0.98,
370
+ "eval_accuracy": 0.4121756487025948,
371
+ "eval_loss": 1.0820704698562622,
372
+ "eval_runtime": 12.3323,
373
+ "eval_samples_per_second": 406.251,
374
+ "eval_steps_per_second": 50.842,
375
+ "step": 24000
376
+ },
377
+ {
378
+ "epoch": 1.02,
379
+ "learning_rate": 6.941111111111112e-07,
380
+ "loss": 1.0915,
381
+ "step": 25000
382
+ },
383
+ {
384
+ "epoch": 1.02,
385
+ "eval_accuracy": 0.4177644710578842,
386
+ "eval_loss": 1.0762931108474731,
387
+ "eval_runtime": 12.3887,
388
+ "eval_samples_per_second": 404.399,
389
+ "eval_steps_per_second": 50.61,
390
+ "step": 25000
391
+ },
392
+ {
393
+ "epoch": 1.06,
394
+ "learning_rate": 7.218888888888889e-07,
395
+ "loss": 1.0889,
396
+ "step": 26000
397
+ },
398
+ {
399
+ "epoch": 1.06,
400
+ "eval_accuracy": 0.43253493013972055,
401
+ "eval_loss": 1.0695326328277588,
402
+ "eval_runtime": 12.5611,
403
+ "eval_samples_per_second": 398.85,
404
+ "eval_steps_per_second": 49.916,
405
+ "step": 26000
406
+ },
407
+ {
408
+ "epoch": 1.1,
409
+ "learning_rate": 7.49638888888889e-07,
410
+ "loss": 1.0827,
411
+ "step": 27000
412
+ },
413
+ {
414
+ "epoch": 1.1,
415
+ "eval_accuracy": 0.4343313373253493,
416
+ "eval_loss": 1.0662548542022705,
417
+ "eval_runtime": 12.641,
418
+ "eval_samples_per_second": 396.328,
419
+ "eval_steps_per_second": 49.6,
420
+ "step": 27000
421
+ },
422
+ {
423
+ "epoch": 1.14,
424
+ "learning_rate": 7.774166666666668e-07,
425
+ "loss": 1.0765,
426
+ "step": 28000
427
+ },
428
+ {
429
+ "epoch": 1.14,
430
+ "eval_accuracy": 0.4315369261477046,
431
+ "eval_loss": 1.0609794855117798,
432
+ "eval_runtime": 12.4637,
433
+ "eval_samples_per_second": 401.967,
434
+ "eval_steps_per_second": 50.306,
435
+ "step": 28000
436
+ },
437
+ {
438
+ "epoch": 1.18,
439
+ "learning_rate": 8.051666666666667e-07,
440
+ "loss": 1.0736,
441
+ "step": 29000
442
+ },
443
+ {
444
+ "epoch": 1.18,
445
+ "eval_accuracy": 0.4377245508982036,
446
+ "eval_loss": 1.0587859153747559,
447
+ "eval_runtime": 12.3801,
448
+ "eval_samples_per_second": 404.683,
449
+ "eval_steps_per_second": 50.646,
450
+ "step": 29000
451
+ },
452
+ {
453
+ "epoch": 1.22,
454
+ "learning_rate": 8.329444444444445e-07,
455
+ "loss": 1.0726,
456
+ "step": 30000
457
+ },
458
+ {
459
+ "epoch": 1.22,
460
+ "eval_accuracy": 0.4249500998003992,
461
+ "eval_loss": 1.0559338331222534,
462
+ "eval_runtime": 12.5945,
463
+ "eval_samples_per_second": 397.792,
464
+ "eval_steps_per_second": 49.784,
465
+ "step": 30000
466
+ },
467
+ {
468
+ "epoch": 1.26,
469
+ "learning_rate": 8.606944444444445e-07,
470
+ "loss": 1.0751,
471
+ "step": 31000
472
+ },
473
+ {
474
+ "epoch": 1.26,
475
+ "eval_accuracy": 0.4427145708582834,
476
+ "eval_loss": 1.0576916933059692,
477
+ "eval_runtime": 12.6324,
478
+ "eval_samples_per_second": 396.599,
479
+ "eval_steps_per_second": 49.634,
480
+ "step": 31000
481
+ },
482
+ {
483
+ "epoch": 1.3,
484
+ "learning_rate": 8.884722222222224e-07,
485
+ "loss": 1.0708,
486
+ "step": 32000
487
+ },
488
+ {
489
+ "epoch": 1.3,
490
+ "eval_accuracy": 0.42634730538922155,
491
+ "eval_loss": 1.0571542978286743,
492
+ "eval_runtime": 12.4696,
493
+ "eval_samples_per_second": 401.776,
494
+ "eval_steps_per_second": 50.282,
495
+ "step": 32000
496
+ },
497
+ {
498
+ "epoch": 1.34,
499
+ "learning_rate": 9.162222222222223e-07,
500
+ "loss": 1.0711,
501
+ "step": 33000
502
+ },
503
+ {
504
+ "epoch": 1.34,
505
+ "eval_accuracy": 0.4419161676646707,
506
+ "eval_loss": 1.0544849634170532,
507
+ "eval_runtime": 12.415,
508
+ "eval_samples_per_second": 403.544,
509
+ "eval_steps_per_second": 50.503,
510
+ "step": 33000
511
+ },
512
+ {
513
+ "epoch": 1.39,
514
+ "learning_rate": 9.440000000000001e-07,
515
+ "loss": 1.0666,
516
+ "step": 34000
517
+ },
518
+ {
519
+ "epoch": 1.39,
520
+ "eval_accuracy": 0.43333333333333335,
521
+ "eval_loss": 1.05018150806427,
522
+ "eval_runtime": 12.6226,
523
+ "eval_samples_per_second": 396.906,
524
+ "eval_steps_per_second": 49.673,
525
+ "step": 34000
526
+ },
527
+ {
528
+ "epoch": 1.43,
529
+ "learning_rate": 9.7175e-07,
530
+ "loss": 1.0698,
531
+ "step": 35000
532
+ },
533
+ {
534
+ "epoch": 1.43,
535
+ "eval_accuracy": 0.4377245508982036,
536
+ "eval_loss": 1.0488406419754028,
537
+ "eval_runtime": 12.6246,
538
+ "eval_samples_per_second": 396.844,
539
+ "eval_steps_per_second": 49.665,
540
+ "step": 35000
541
+ },
542
+ {
543
+ "epoch": 1.47,
544
+ "learning_rate": 9.995277777777778e-07,
545
+ "loss": 1.067,
546
+ "step": 36000
547
+ },
548
+ {
549
+ "epoch": 1.47,
550
+ "eval_accuracy": 0.4407185628742515,
551
+ "eval_loss": 1.047359824180603,
552
+ "eval_runtime": 12.4039,
553
+ "eval_samples_per_second": 403.906,
554
+ "eval_steps_per_second": 50.549,
555
+ "step": 36000
556
+ },
557
+ {
558
+ "epoch": 1.51,
559
+ "learning_rate": 1.027277777777778e-06,
560
+ "loss": 1.0636,
561
+ "step": 37000
562
+ },
563
+ {
564
+ "epoch": 1.51,
565
+ "eval_accuracy": 0.44451097804391215,
566
+ "eval_loss": 1.0455204248428345,
567
+ "eval_runtime": 12.3836,
568
+ "eval_samples_per_second": 404.569,
569
+ "eval_steps_per_second": 50.632,
570
+ "step": 37000
571
+ },
572
+ {
573
+ "epoch": 1.55,
574
+ "learning_rate": 1.0550555555555557e-06,
575
+ "loss": 1.0658,
576
+ "step": 38000
577
+ },
578
+ {
579
+ "epoch": 1.55,
580
+ "eval_accuracy": 0.4407185628742515,
581
+ "eval_loss": 1.0461794137954712,
582
+ "eval_runtime": 12.5546,
583
+ "eval_samples_per_second": 399.057,
584
+ "eval_steps_per_second": 49.942,
585
+ "step": 38000
586
+ },
587
+ {
588
+ "epoch": 1.59,
589
+ "learning_rate": 1.0828333333333334e-06,
590
+ "loss": 1.0635,
591
+ "step": 39000
592
+ },
593
+ {
594
+ "epoch": 1.59,
595
+ "eval_accuracy": 0.4367265469061876,
596
+ "eval_loss": 1.0461270809173584,
597
+ "eval_runtime": 12.5509,
598
+ "eval_samples_per_second": 399.173,
599
+ "eval_steps_per_second": 49.956,
600
+ "step": 39000
601
+ },
602
+ {
603
+ "epoch": 1.63,
604
+ "learning_rate": 1.1105833333333335e-06,
605
+ "loss": 1.0596,
606
+ "step": 40000
607
+ },
608
+ {
609
+ "epoch": 1.63,
610
+ "eval_accuracy": 0.449500998003992,
611
+ "eval_loss": 1.0418727397918701,
612
+ "eval_runtime": 12.4044,
613
+ "eval_samples_per_second": 403.889,
614
+ "eval_steps_per_second": 50.547,
615
+ "step": 40000
616
+ },
617
+ {
618
+ "epoch": 1.67,
619
+ "learning_rate": 1.1383611111111113e-06,
620
+ "loss": 1.0595,
621
+ "step": 41000
622
+ },
623
+ {
624
+ "epoch": 1.67,
625
+ "eval_accuracy": 0.4409181636726547,
626
+ "eval_loss": 1.0417557954788208,
627
+ "eval_runtime": 12.3642,
628
+ "eval_samples_per_second": 405.202,
629
+ "eval_steps_per_second": 50.711,
630
+ "step": 41000
631
+ },
632
+ {
633
+ "epoch": 1.71,
634
+ "learning_rate": 1.1661111111111111e-06,
635
+ "loss": 1.054,
636
+ "step": 42000
637
+ },
638
+ {
639
+ "epoch": 1.71,
640
+ "eval_accuracy": 0.4357285429141717,
641
+ "eval_loss": 1.0383535623550415,
642
+ "eval_runtime": 12.3634,
643
+ "eval_samples_per_second": 405.23,
644
+ "eval_steps_per_second": 50.714,
645
+ "step": 42000
646
+ },
647
+ {
648
+ "epoch": 1.75,
649
+ "learning_rate": 1.193888888888889e-06,
650
+ "loss": 1.0543,
651
+ "step": 43000
652
+ },
653
+ {
654
+ "epoch": 1.75,
655
+ "eval_accuracy": 0.43812375249500995,
656
+ "eval_loss": 1.0463169813156128,
657
+ "eval_runtime": 12.6398,
658
+ "eval_samples_per_second": 396.366,
659
+ "eval_steps_per_second": 49.605,
660
+ "step": 43000
661
+ },
662
+ {
663
+ "epoch": 1.79,
664
+ "learning_rate": 1.221638888888889e-06,
665
+ "loss": 1.0579,
666
+ "step": 44000
667
+ },
668
+ {
669
+ "epoch": 1.79,
670
+ "eval_accuracy": 0.4415169660678643,
671
+ "eval_loss": 1.0388957262039185,
672
+ "eval_runtime": 12.4542,
673
+ "eval_samples_per_second": 402.274,
674
+ "eval_steps_per_second": 50.344,
675
+ "step": 44000
676
+ },
677
+ {
678
+ "epoch": 1.83,
679
+ "learning_rate": 1.2494166666666668e-06,
680
+ "loss": 1.0561,
681
+ "step": 45000
682
+ },
683
+ {
684
+ "epoch": 1.83,
685
+ "eval_accuracy": 0.46187624750499,
686
+ "eval_loss": 1.031552791595459,
687
+ "eval_runtime": 12.4082,
688
+ "eval_samples_per_second": 403.764,
689
+ "eval_steps_per_second": 50.531,
690
+ "step": 45000
691
+ },
692
+ {
693
+ "epoch": 1.87,
694
+ "learning_rate": 1.2771666666666668e-06,
695
+ "loss": 1.0516,
696
+ "step": 46000
697
+ },
698
+ {
699
+ "epoch": 1.87,
700
+ "eval_accuracy": 0.4538922155688623,
701
+ "eval_loss": 1.0312999486923218,
702
+ "eval_runtime": 12.3695,
703
+ "eval_samples_per_second": 405.029,
704
+ "eval_steps_per_second": 50.689,
705
+ "step": 46000
706
+ },
707
+ {
708
+ "epoch": 1.91,
709
+ "learning_rate": 1.3049444444444446e-06,
710
+ "loss": 1.0471,
711
+ "step": 47000
712
+ },
713
+ {
714
+ "epoch": 1.91,
715
+ "eval_accuracy": 0.46387225548902195,
716
+ "eval_loss": 1.027535319328308,
717
+ "eval_runtime": 12.5846,
718
+ "eval_samples_per_second": 398.107,
719
+ "eval_steps_per_second": 49.823,
720
+ "step": 47000
721
+ },
722
+ {
723
+ "epoch": 1.96,
724
+ "learning_rate": 1.3326944444444447e-06,
725
+ "loss": 1.054,
726
+ "step": 48000
727
+ },
728
+ {
729
+ "epoch": 1.96,
730
+ "eval_accuracy": 0.4568862275449102,
731
+ "eval_loss": 1.0308438539505005,
732
+ "eval_runtime": 12.5646,
733
+ "eval_samples_per_second": 398.74,
734
+ "eval_steps_per_second": 49.902,
735
+ "step": 48000
736
+ },
737
+ {
738
+ "epoch": 2.0,
739
+ "learning_rate": 1.3604722222222224e-06,
740
+ "loss": 1.0445,
741
+ "step": 49000
742
+ },
743
+ {
744
+ "epoch": 2.0,
745
+ "eval_accuracy": 0.47045908183632734,
746
+ "eval_loss": 1.0230755805969238,
747
+ "eval_runtime": 12.4305,
748
+ "eval_samples_per_second": 403.041,
749
+ "eval_steps_per_second": 50.44,
750
+ "step": 49000
751
+ },
752
+ {
753
+ "epoch": 2.04,
754
+ "learning_rate": 1.3882222222222223e-06,
755
+ "loss": 1.0479,
756
+ "step": 50000
757
+ },
758
+ {
759
+ "epoch": 2.04,
760
+ "eval_accuracy": 0.47305389221556887,
761
+ "eval_loss": 1.0201445817947388,
762
+ "eval_runtime": 12.3672,
763
+ "eval_samples_per_second": 405.103,
764
+ "eval_steps_per_second": 50.698,
765
+ "step": 50000
766
+ },
767
+ {
768
+ "epoch": 2.08,
769
+ "learning_rate": 1.416e-06,
770
+ "loss": 1.0482,
771
+ "step": 51000
772
+ },
773
+ {
774
+ "epoch": 2.08,
775
+ "eval_accuracy": 0.4678642714570858,
776
+ "eval_loss": 1.0167714357376099,
777
+ "eval_runtime": 12.5982,
778
+ "eval_samples_per_second": 397.677,
779
+ "eval_steps_per_second": 49.769,
780
+ "step": 51000
781
+ },
782
+ {
783
+ "epoch": 2.12,
784
+ "learning_rate": 1.4437500000000002e-06,
785
+ "loss": 1.0406,
786
+ "step": 52000
787
+ },
788
+ {
789
+ "epoch": 2.12,
790
+ "eval_accuracy": 0.4722554890219561,
791
+ "eval_loss": 1.0175116062164307,
792
+ "eval_runtime": 12.6041,
793
+ "eval_samples_per_second": 397.49,
794
+ "eval_steps_per_second": 49.746,
795
+ "step": 52000
796
+ },
797
+ {
798
+ "epoch": 2.16,
799
+ "learning_rate": 1.471527777777778e-06,
800
+ "loss": 1.0457,
801
+ "step": 53000
802
+ },
803
+ {
804
+ "epoch": 2.16,
805
+ "eval_accuracy": 0.4748502994011976,
806
+ "eval_loss": 1.010857343673706,
807
+ "eval_runtime": 12.3728,
808
+ "eval_samples_per_second": 404.921,
809
+ "eval_steps_per_second": 50.676,
810
+ "step": 53000
811
+ }
812
+ ],
813
+ "logging_steps": 1000,
814
+ "max_steps": 10000000,
815
+ "num_train_epochs": 408,
816
+ "save_steps": 1000,
817
+ "total_flos": 1.846721739541709e+16,
818
+ "trial_name": null,
819
+ "trial_params": null
820
+ }
checkpoint-53000/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f48fd322744a41332d86e79cb68054df0053a36e9279c46ff138f13ef1aab365
3
+ size 4027
checkpoint-53000/vocab.json ADDED
The diff for this file is too large to render. See raw diff