alexue4 commited on
Commit
2678be8
1 Parent(s): 4a2e4f9

End of training

Browse files
Files changed (4) hide show
  1. README.md +31 -26
  2. pytorch_model.bin +1 -1
  3. trainer_state.json +957 -973
  4. training_args.bin +2 -2
README.md CHANGED
@@ -15,7 +15,7 @@ should probably proofread and complete it, then remove this comment. -->
15
 
16
  This model is a fine-tuned version of [alexue4/text-normalization-ru-new](https://huggingface.co/alexue4/text-normalization-ru-new) on the None dataset.
17
  It achieves the following results on the evaluation set:
18
- - Loss: 0.0170
19
  - Mean Distance: 0
20
  - Max Distance: 3
21
 
@@ -37,38 +37,43 @@ More information needed
37
 
38
  The following hyperparameters were used during training:
39
  - learning_rate: 0.0001
40
- - train_batch_size: 15
41
- - eval_batch_size: 15
42
  - seed: 42
43
  - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
44
  - lr_scheduler_type: linear
45
  - lr_scheduler_warmup_ratio: 0.1
46
- - num_epochs: 20
47
 
48
  ### Training results
49
 
50
- | Training Loss | Epoch | Step | Validation Loss | Mean Distance | Max Distance |
51
- |:-------------:|:-----:|:------:|:---------------:|:-------------:|:------------:|
52
- | 0.0004 | 1.0 | 13982 | 0.0145 | 0 | 4 |
53
- | 0.0005 | 2.0 | 27964 | 0.0164 | 0 | 4 |
54
- | 0.0005 | 3.0 | 41946 | 0.0167 | 0 | 3 |
55
- | 0.0005 | 4.0 | 55928 | 0.0125 | 0 | 3 |
56
- | 0.0005 | 5.0 | 69910 | 0.0185 | 0 | 3 |
57
- | 0.0005 | 6.0 | 83892 | 0.0173 | 0 | 3 |
58
- | 0.0005 | 7.0 | 97874 | 0.0168 | 0 | 5 |
59
- | 0.0004 | 8.0 | 111856 | 0.0201 | 0 | 5 |
60
- | 0.0004 | 9.0 | 125838 | 0.0215 | 0 | 2 |
61
- | 0.0004 | 10.0 | 139820 | 0.0226 | 0 | 3 |
62
- | 0.0004 | 11.0 | 153802 | 0.0199 | 0 | 3 |
63
- | 0.0003 | 12.0 | 167784 | 0.0184 | 0 | 3 |
64
- | 0.0004 | 13.0 | 181766 | 0.0196 | 0 | 3 |
65
- | 0.0003 | 14.0 | 195748 | 0.0167 | 0 | 1 |
66
- | 0.0003 | 15.0 | 209730 | 0.0177 | 0 | 3 |
67
- | 0.0003 | 16.0 | 223712 | 0.0173 | 0 | 3 |
68
- | 0.0003 | 17.0 | 237694 | 0.0173 | 0 | 3 |
69
- | 0.0003 | 18.0 | 251676 | 0.0169 | 0 | 3 |
70
- | 0.0003 | 19.0 | 265658 | 0.0167 | 0 | 3 |
71
- | 0.0003 | 20.0 | 279640 | 0.0170 | 0 | 3 |
 
 
 
 
 
72
 
73
 
74
  ### Framework versions
 
15
 
16
  This model is a fine-tuned version of [alexue4/text-normalization-ru-new](https://huggingface.co/alexue4/text-normalization-ru-new) on the None dataset.
17
  It achieves the following results on the evaluation set:
18
+ - Loss: 0.0014
19
  - Mean Distance: 0
20
  - Max Distance: 3
21
 
 
37
 
38
  The following hyperparameters were used during training:
39
  - learning_rate: 0.0001
40
+ - train_batch_size: 30
41
+ - eval_batch_size: 30
42
  - seed: 42
43
  - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
44
  - lr_scheduler_type: linear
45
  - lr_scheduler_warmup_ratio: 0.1
46
+ - num_epochs: 25
47
 
48
  ### Training results
49
 
50
+ | Training Loss | Epoch | Step | Validation Loss | Mean Distance | Max Distance |
51
+ |:-------------:|:-----:|:----:|:---------------:|:-------------:|:------------:|
52
+ | 0.0368 | 1.0 | 68 | 0.0619 | 0 | 3 |
53
+ | 0.0057 | 2.0 | 136 | 0.0346 | 0 | 3 |
54
+ | 0.0113 | 3.0 | 204 | 0.0252 | 0 | 4 |
55
+ | 0.0244 | 4.0 | 272 | 0.0111 | 0 | 3 |
56
+ | 0.001 | 5.0 | 340 | 0.0065 | 0 | 3 |
57
+ | 0.0021 | 6.0 | 408 | 0.0051 | 0 | 5 |
58
+ | 0.0008 | 7.0 | 476 | 0.0077 | 0 | 3 |
59
+ | 0.0008 | 8.0 | 544 | 0.0040 | 0 | 3 |
60
+ | 0.0025 | 9.0 | 612 | 0.0029 | 0 | 2 |
61
+ | 0.0012 | 10.0 | 680 | 0.0020 | 0 | 3 |
62
+ | 0.0012 | 11.0 | 748 | 0.0010 | 0 | 3 |
63
+ | 0.0006 | 12.0 | 816 | 0.0008 | 0 | 0 |
64
+ | 0.0012 | 13.0 | 884 | 0.0005 | 0 | 0 |
65
+ | 0.001 | 14.0 | 952 | 0.0007 | 0 | 3 |
66
+ | 0.0009 | 15.0 | 1020 | 0.0014 | 0 | 3 |
67
+ | 0.0005 | 16.0 | 1088 | 0.0015 | 0 | 3 |
68
+ | 0.0009 | 17.0 | 1156 | 0.0014 | 0 | 3 |
69
+ | 0.0006 | 18.0 | 1224 | 0.0013 | 0 | 3 |
70
+ | 0.0012 | 19.0 | 1292 | 0.0014 | 0 | 3 |
71
+ | 0.0007 | 20.0 | 1360 | 0.0014 | 0 | 3 |
72
+ | 0.0012 | 21.0 | 1428 | 0.0013 | 0 | 3 |
73
+ | 0.0013 | 22.0 | 1496 | 0.0014 | 0 | 3 |
74
+ | 0.0008 | 23.0 | 1564 | 0.0014 | 0 | 3 |
75
+ | 0.0005 | 24.0 | 1632 | 0.0014 | 0 | 3 |
76
+ | 0.001 | 25.0 | 1700 | 0.0014 | 0 | 3 |
77
 
78
 
79
  ### Framework versions
pytorch_model.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:0faab56c1613c68c3c5f4b631273360760561866b59ee610afd96d410f0b2012
3
  size 258643461
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9c78ae41e27176e6da318232c2b38c680ba5bed653558b5af6c48e9134e6fe90
3
  size 258643461
trainer_state.json CHANGED
@@ -1,1428 +1,1412 @@
1
  {
2
- "best_metric": null,
3
- "best_model_checkpoint": null,
4
- "epoch": 20.0,
5
  "eval_steps": 500,
6
- "global_step": 279640,
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": 3.576026319553712e-09,
14
- "loss": 0.0,
15
  "step": 1
16
  },
17
  {
18
- "epoch": 0.1,
19
- "learning_rate": 5.002860821055643e-06,
20
- "loss": 0.0011,
21
- "step": 1399
22
  },
23
  {
24
- "epoch": 0.2,
25
- "learning_rate": 1.0005721642111287e-05,
26
- "loss": 0.0008,
27
- "step": 2798
28
- },
29
- {
30
- "epoch": 0.3,
31
- "learning_rate": 1.500858246316693e-05,
32
- "loss": 0.0008,
33
- "step": 4197
34
  },
35
  {
36
  "epoch": 0.4,
37
- "learning_rate": 2.0011443284222573e-05,
38
- "loss": 0.0004,
39
- "step": 5596
40
  },
41
  {
42
- "epoch": 0.5,
43
- "learning_rate": 2.5014304105278213e-05,
44
- "loss": 0.0004,
45
- "step": 6995
46
  },
47
  {
48
- "epoch": 0.6,
49
- "learning_rate": 3.001716492633386e-05,
50
- "loss": 0.0004,
51
- "step": 8394
52
- },
53
- {
54
- "epoch": 0.7,
55
- "learning_rate": 3.5020025747389504e-05,
56
- "loss": 0.0005,
57
- "step": 9793
58
  },
59
  {
60
- "epoch": 0.8,
61
- "learning_rate": 4.002288656844515e-05,
62
- "loss": 0.0003,
63
- "step": 11192
64
  },
65
  {
66
- "epoch": 0.9,
67
- "learning_rate": 4.502574738950079e-05,
68
- "loss": 0.0004,
69
- "step": 12591
70
  },
71
  {
72
  "epoch": 1.0,
73
- "eval_loss": 0.01447892002761364,
74
- "eval_max_distance": 4,
75
  "eval_mean_distance": 0,
76
- "eval_runtime": 16.1519,
77
- "eval_samples_per_second": 16.593,
78
- "eval_steps_per_second": 1.114,
79
- "step": 13982
80
- },
81
- {
82
- "epoch": 1.0,
83
- "learning_rate": 5.0028608210556425e-05,
84
- "loss": 0.0004,
85
- "step": 13990
86
  },
87
  {
88
- "epoch": 1.1,
89
- "learning_rate": 5.503146903161207e-05,
90
- "loss": 0.0004,
91
- "step": 15389
92
- },
93
- {
94
- "epoch": 1.2,
95
- "learning_rate": 6.003432985266772e-05,
96
- "loss": 0.0004,
97
- "step": 16788
98
  },
99
  {
100
- "epoch": 1.3,
101
- "learning_rate": 6.503719067372336e-05,
102
- "loss": 0.0004,
103
- "step": 18187
104
  },
105
  {
106
- "epoch": 1.4,
107
- "learning_rate": 7.004005149477901e-05,
108
- "loss": 0.0004,
109
- "step": 19586
110
  },
111
  {
112
- "epoch": 1.5,
113
- "learning_rate": 7.504291231583464e-05,
114
- "loss": 0.0004,
115
- "step": 20985
116
  },
117
  {
118
- "epoch": 1.6,
119
- "learning_rate": 8.00457731368903e-05,
120
- "loss": 0.0005,
121
- "step": 22384
122
  },
123
  {
124
- "epoch": 1.7,
125
- "learning_rate": 8.504863395794593e-05,
126
- "loss": 0.0004,
127
- "step": 23783
128
  },
129
  {
130
- "epoch": 1.8,
131
- "learning_rate": 9.005149477900158e-05,
132
- "loss": 0.0004,
133
- "step": 25182
134
  },
135
  {
136
- "epoch": 1.9,
137
- "learning_rate": 9.505435560005721e-05,
138
- "loss": 0.0005,
139
- "step": 26581
140
  },
141
  {
142
  "epoch": 2.0,
143
- "eval_loss": 0.01644507423043251,
144
- "eval_max_distance": 4,
145
  "eval_mean_distance": 0,
146
- "eval_runtime": 14.8804,
147
- "eval_samples_per_second": 18.01,
148
- "eval_steps_per_second": 1.21,
149
- "step": 27964
150
  },
151
  {
152
- "epoch": 2.0,
153
- "learning_rate": 9.999364261987636e-05,
154
- "loss": 0.0004,
155
- "step": 27980
156
  },
157
  {
158
- "epoch": 2.1,
159
- "learning_rate": 9.943776919531461e-05,
160
- "loss": 0.0005,
161
- "step": 29379
162
  },
163
  {
164
- "epoch": 2.2,
165
- "learning_rate": 9.888189577075288e-05,
166
- "loss": 0.0005,
167
- "step": 30778
168
  },
169
  {
170
- "epoch": 2.3,
171
- "learning_rate": 9.832602234619113e-05,
172
- "loss": 0.0005,
173
- "step": 32177
174
  },
175
  {
176
- "epoch": 2.4,
177
- "learning_rate": 9.77701489216294e-05,
178
- "loss": 0.0005,
179
- "step": 33576
180
- },
181
- {
182
- "epoch": 2.5,
183
- "learning_rate": 9.721427549706766e-05,
184
- "loss": 0.0005,
185
- "step": 34975
186
- },
187
- {
188
- "epoch": 2.6,
189
- "learning_rate": 9.665840207250593e-05,
190
- "loss": 0.0006,
191
- "step": 36374
192
- },
193
- {
194
- "epoch": 2.7,
195
- "learning_rate": 9.61025286479442e-05,
196
  "loss": 0.0006,
197
- "step": 37773
198
  },
199
  {
200
- "epoch": 2.8,
201
- "learning_rate": 9.554665522338245e-05,
202
- "loss": 0.0005,
203
- "step": 39172
204
  },
205
  {
206
- "epoch": 2.9,
207
- "learning_rate": 9.499078179882072e-05,
208
- "loss": 0.0005,
209
- "step": 40571
210
  },
211
  {
212
  "epoch": 3.0,
213
- "eval_loss": 0.016705745831131935,
214
- "eval_max_distance": 3,
215
  "eval_mean_distance": 0,
216
- "eval_runtime": 14.6064,
217
- "eval_samples_per_second": 18.348,
218
- "eval_steps_per_second": 1.232,
219
- "step": 41946
220
  },
221
  {
222
- "epoch": 3.0,
223
- "learning_rate": 9.443490837425897e-05,
224
- "loss": 0.0006,
225
- "step": 41970
226
  },
227
  {
228
- "epoch": 3.1,
229
- "learning_rate": 9.387903494969723e-05,
230
- "loss": 0.0005,
231
- "step": 43369
232
  },
233
  {
234
- "epoch": 3.2,
235
- "learning_rate": 9.33231615251355e-05,
236
- "loss": 0.0005,
237
- "step": 44768
238
  },
239
  {
240
- "epoch": 3.3,
241
- "learning_rate": 9.276728810057375e-05,
242
- "loss": 0.0006,
243
- "step": 46167
244
  },
245
  {
246
- "epoch": 3.4,
247
- "learning_rate": 9.221141467601202e-05,
248
- "loss": 0.0005,
249
- "step": 47566
250
  },
251
  {
252
- "epoch": 3.5,
253
- "learning_rate": 9.165554125145028e-05,
254
- "loss": 0.0006,
255
- "step": 48965
256
  },
257
  {
258
- "epoch": 3.6,
259
- "learning_rate": 9.109966782688855e-05,
260
- "loss": 0.0005,
261
- "step": 50364
262
- },
263
- {
264
- "epoch": 3.7,
265
- "learning_rate": 9.054379440232682e-05,
266
- "loss": 0.0005,
267
- "step": 51763
268
  },
269
  {
270
- "epoch": 3.8,
271
- "learning_rate": 8.998792097776507e-05,
272
- "loss": 0.0005,
273
- "step": 53162
274
- },
275
- {
276
- "epoch": 3.9,
277
- "learning_rate": 8.943204755320333e-05,
278
- "loss": 0.0005,
279
- "step": 54561
280
  },
281
  {
282
  "epoch": 4.0,
283
- "eval_loss": 0.012466676533222198,
284
  "eval_max_distance": 3,
285
  "eval_mean_distance": 0,
286
- "eval_runtime": 14.6723,
287
- "eval_samples_per_second": 18.266,
288
- "eval_steps_per_second": 1.227,
289
- "step": 55928
290
- },
291
- {
292
- "epoch": 4.0,
293
- "learning_rate": 8.887617412864158e-05,
294
- "loss": 0.0006,
295
- "step": 55960
296
  },
297
  {
298
  "epoch": 4.1,
299
- "learning_rate": 8.832030070407985e-05,
300
- "loss": 0.0005,
301
- "step": 57359
302
- },
303
- {
304
- "epoch": 4.2,
305
- "learning_rate": 8.776442727951812e-05,
306
- "loss": 0.0005,
307
- "step": 58758
308
  },
309
  {
310
- "epoch": 4.3,
311
- "learning_rate": 8.720855385495637e-05,
312
- "loss": 0.0005,
313
- "step": 60157
314
  },
315
  {
316
- "epoch": 4.4,
317
- "learning_rate": 8.665268043039464e-05,
318
- "loss": 0.0005,
319
- "step": 61556
320
  },
321
  {
322
  "epoch": 4.5,
323
- "learning_rate": 8.60968070058329e-05,
324
- "loss": 0.0005,
325
- "step": 62955
326
  },
327
  {
328
- "epoch": 4.6,
329
- "learning_rate": 8.554093358127117e-05,
330
- "loss": 0.0005,
331
- "step": 64354
332
  },
333
  {
334
- "epoch": 4.7,
335
- "learning_rate": 8.498506015670943e-05,
336
- "loss": 0.0006,
337
- "step": 65753
338
- },
339
- {
340
- "epoch": 4.8,
341
- "learning_rate": 8.442918673214769e-05,
342
- "loss": 0.0005,
343
- "step": 67152
344
  },
345
  {
346
  "epoch": 4.9,
347
- "learning_rate": 8.387331330758594e-05,
348
- "loss": 0.0005,
349
- "step": 68551
350
  },
351
  {
352
  "epoch": 5.0,
353
- "eval_loss": 0.018481934443116188,
354
  "eval_max_distance": 3,
355
  "eval_mean_distance": 0,
356
- "eval_runtime": 14.7593,
357
- "eval_samples_per_second": 18.158,
358
- "eval_steps_per_second": 1.22,
359
- "step": 69910
360
  },
361
  {
362
- "epoch": 5.0,
363
- "learning_rate": 8.33174398830242e-05,
364
- "loss": 0.0005,
365
- "step": 69950
366
  },
367
  {
368
- "epoch": 5.1,
369
- "learning_rate": 8.276156645846247e-05,
370
- "loss": 0.0005,
371
- "step": 71349
372
  },
373
  {
374
- "epoch": 5.2,
375
- "learning_rate": 8.220569303390074e-05,
376
- "loss": 0.0005,
377
- "step": 72748
378
  },
379
  {
380
- "epoch": 5.3,
381
- "learning_rate": 8.164981960933899e-05,
382
- "loss": 0.0005,
383
- "step": 74147
384
  },
385
  {
386
- "epoch": 5.4,
387
- "learning_rate": 8.109394618477726e-05,
388
- "loss": 0.0005,
389
- "step": 75546
390
- },
391
- {
392
- "epoch": 5.5,
393
- "learning_rate": 8.053807276021552e-05,
394
- "loss": 0.0005,
395
- "step": 76945
396
- },
397
- {
398
- "epoch": 5.6,
399
- "learning_rate": 7.998219933565379e-05,
400
- "loss": 0.0005,
401
- "step": 78344
402
  },
403
  {
404
- "epoch": 5.7,
405
- "learning_rate": 7.942632591109205e-05,
406
- "loss": 0.0004,
407
- "step": 79743
408
  },
409
  {
410
- "epoch": 5.8,
411
- "learning_rate": 7.88704524865303e-05,
412
- "loss": 0.0005,
413
- "step": 81142
414
  },
415
  {
416
- "epoch": 5.9,
417
- "learning_rate": 7.831457906196856e-05,
418
- "loss": 0.0005,
419
- "step": 82541
420
  },
421
  {
422
  "epoch": 6.0,
423
- "eval_loss": 0.017317254096269608,
424
- "eval_max_distance": 3,
425
  "eval_mean_distance": 0,
426
- "eval_runtime": 14.4763,
427
- "eval_samples_per_second": 18.513,
428
- "eval_steps_per_second": 1.243,
429
- "step": 83892
430
  },
431
  {
432
- "epoch": 6.0,
433
- "learning_rate": 7.775870563740682e-05,
434
- "loss": 0.0005,
435
- "step": 83940
436
  },
437
  {
438
- "epoch": 6.1,
439
- "learning_rate": 7.720283221284509e-05,
440
- "loss": 0.0004,
441
- "step": 85339
442
  },
443
  {
444
- "epoch": 6.2,
445
- "learning_rate": 7.664695878828336e-05,
446
- "loss": 0.0006,
447
- "step": 86738
 
 
 
 
 
 
448
  },
449
  {
450
- "epoch": 6.3,
451
- "learning_rate": 7.609108536372161e-05,
452
  "loss": 0.0004,
453
- "step": 88137
454
  },
455
  {
456
- "epoch": 6.4,
457
- "learning_rate": 7.553521193915988e-05,
458
- "loss": 0.0005,
459
- "step": 89536
460
  },
461
  {
462
- "epoch": 6.5,
463
- "learning_rate": 7.497933851459814e-05,
464
- "loss": 0.0004,
465
- "step": 90935
466
  },
467
  {
468
- "epoch": 6.6,
469
- "learning_rate": 7.44234650900364e-05,
470
- "loss": 0.0004,
471
- "step": 92334
 
 
 
 
472
  },
473
  {
474
- "epoch": 6.7,
475
- "learning_rate": 7.386759166547467e-05,
476
- "loss": 0.0005,
477
- "step": 93733
478
  },
479
  {
480
- "epoch": 6.8,
481
- "learning_rate": 7.331171824091292e-05,
482
- "loss": 0.0004,
483
- "step": 95132
484
  },
485
  {
486
- "epoch": 6.9,
487
- "learning_rate": 7.275584481635118e-05,
488
- "loss": 0.0005,
489
- "step": 96531
490
  },
491
  {
492
- "epoch": 7.0,
493
- "eval_loss": 0.016764231026172638,
494
- "eval_max_distance": 5,
495
- "eval_mean_distance": 0,
496
- "eval_runtime": 14.4483,
497
- "eval_samples_per_second": 18.549,
498
- "eval_steps_per_second": 1.246,
499
- "step": 97874
500
  },
501
  {
502
- "epoch": 7.0,
503
- "learning_rate": 7.219997139178944e-05,
504
- "loss": 0.0005,
505
- "step": 97930
506
  },
507
  {
508
- "epoch": 7.1,
509
- "learning_rate": 7.164409796722771e-05,
510
- "loss": 0.0005,
511
- "step": 99329
512
  },
513
  {
514
- "epoch": 7.2,
515
- "learning_rate": 7.108822454266598e-05,
516
- "loss": 0.0005,
517
- "step": 100728
518
  },
519
  {
520
- "epoch": 7.3,
521
- "learning_rate": 7.053235111810423e-05,
522
- "loss": 0.0005,
523
- "step": 102127
524
  },
525
  {
526
- "epoch": 7.4,
527
- "learning_rate": 6.99764776935425e-05,
528
- "loss": 0.0005,
529
- "step": 103526
 
 
 
 
530
  },
531
  {
532
- "epoch": 7.5,
533
- "learning_rate": 6.942060426898076e-05,
534
- "loss": 0.0005,
535
- "step": 104925
536
  },
537
  {
538
- "epoch": 7.6,
539
- "learning_rate": 6.886473084441902e-05,
540
- "loss": 0.0004,
541
- "step": 106324
542
  },
543
  {
544
- "epoch": 7.7,
545
- "learning_rate": 6.830885741985728e-05,
546
- "loss": 0.0005,
547
- "step": 107723
548
  },
549
  {
550
- "epoch": 7.8,
551
- "learning_rate": 6.775298399529553e-05,
552
- "loss": 0.0004,
553
- "step": 109122
554
  },
555
  {
556
- "epoch": 7.9,
557
- "learning_rate": 6.71971105707338e-05,
558
- "loss": 0.0004,
559
- "step": 110521
560
  },
561
  {
562
- "epoch": 8.0,
563
- "eval_loss": 0.020108750090003014,
564
- "eval_max_distance": 5,
565
- "eval_mean_distance": 0,
566
- "eval_runtime": 14.4845,
567
- "eval_samples_per_second": 18.503,
568
- "eval_steps_per_second": 1.243,
569
- "step": 111856
570
  },
571
  {
572
- "epoch": 8.0,
573
- "learning_rate": 6.664123714617206e-05,
574
- "loss": 0.0004,
575
- "step": 111920
576
  },
577
  {
578
- "epoch": 8.1,
579
- "learning_rate": 6.608536372161033e-05,
580
- "loss": 0.0005,
581
- "step": 113319
582
  },
583
  {
584
- "epoch": 8.2,
585
- "learning_rate": 6.55294902970486e-05,
586
- "loss": 0.0004,
587
- "step": 114718
 
 
 
 
588
  },
589
  {
590
- "epoch": 8.3,
591
- "learning_rate": 6.497361687248685e-05,
592
- "loss": 0.0005,
593
- "step": 116117
594
  },
595
  {
596
- "epoch": 8.4,
597
- "learning_rate": 6.441774344792512e-05,
598
- "loss": 0.0005,
599
- "step": 117516
600
  },
601
  {
602
- "epoch": 8.5,
603
- "learning_rate": 6.386187002336337e-05,
604
- "loss": 0.0005,
605
- "step": 118915
606
  },
607
  {
608
- "epoch": 8.6,
609
- "learning_rate": 6.330599659880164e-05,
610
- "loss": 0.0004,
611
- "step": 120314
612
  },
613
  {
614
- "epoch": 8.7,
615
- "learning_rate": 6.27501231742399e-05,
616
- "loss": 0.0004,
617
- "step": 121713
618
  },
619
  {
620
- "epoch": 8.81,
621
- "learning_rate": 6.219424974967815e-05,
622
- "loss": 0.0004,
623
- "step": 123112
624
  },
625
  {
626
- "epoch": 8.91,
627
- "learning_rate": 6.163837632511642e-05,
628
- "loss": 0.0004,
629
- "step": 124511
630
  },
631
  {
632
- "epoch": 9.0,
633
- "eval_loss": 0.021496200934052467,
634
- "eval_max_distance": 2,
635
  "eval_mean_distance": 0,
636
- "eval_runtime": 14.5428,
637
- "eval_samples_per_second": 18.428,
638
- "eval_steps_per_second": 1.238,
639
- "step": 125838
640
  },
641
  {
642
- "epoch": 9.01,
643
- "learning_rate": 6.108250290055468e-05,
644
- "loss": 0.0004,
645
- "step": 125910
646
  },
647
  {
648
- "epoch": 9.11,
649
- "learning_rate": 6.0526629475992947e-05,
650
- "loss": 0.0004,
651
- "step": 127309
652
  },
653
  {
654
- "epoch": 9.21,
655
- "learning_rate": 5.997075605143121e-05,
656
- "loss": 0.0004,
657
- "step": 128708
658
  },
659
  {
660
- "epoch": 9.31,
661
- "learning_rate": 5.941488262686947e-05,
662
- "loss": 0.0004,
663
- "step": 130107
664
  },
665
  {
666
- "epoch": 9.41,
667
- "learning_rate": 5.885900920230773e-05,
668
- "loss": 0.0004,
669
- "step": 131506
670
  },
671
  {
672
- "epoch": 9.51,
673
- "learning_rate": 5.8303135777746e-05,
674
- "loss": 0.0004,
675
- "step": 132905
676
  },
677
  {
678
- "epoch": 9.61,
679
- "learning_rate": 5.774726235318425e-05,
680
- "loss": 0.0004,
681
- "step": 134304
682
  },
683
  {
684
- "epoch": 9.71,
685
- "learning_rate": 5.719138892862251e-05,
686
- "loss": 0.0005,
687
- "step": 135703
688
  },
689
  {
690
- "epoch": 9.81,
691
- "learning_rate": 5.663551550406077e-05,
692
- "loss": 0.0004,
693
- "step": 137102
 
 
 
 
694
  },
695
  {
696
- "epoch": 9.91,
697
- "learning_rate": 5.607964207949904e-05,
698
- "loss": 0.0004,
699
- "step": 138501
700
  },
701
  {
702
- "epoch": 10.0,
703
- "eval_loss": 0.0225870031863451,
704
- "eval_max_distance": 3,
705
- "eval_mean_distance": 0,
706
- "eval_runtime": 15.0837,
707
- "eval_samples_per_second": 17.768,
708
- "eval_steps_per_second": 1.193,
709
- "step": 139820
710
  },
711
  {
712
- "epoch": 10.01,
713
- "learning_rate": 5.5523768654937304e-05,
714
- "loss": 0.0004,
715
- "step": 139900
716
  },
717
  {
718
- "epoch": 10.11,
719
- "learning_rate": 5.4967895230375566e-05,
720
- "loss": 0.0004,
721
- "step": 141299
722
  },
723
  {
724
- "epoch": 10.21,
725
- "learning_rate": 5.441202180581383e-05,
726
- "loss": 0.0004,
727
- "step": 142698
728
  },
729
  {
730
- "epoch": 10.31,
731
- "learning_rate": 5.385614838125209e-05,
732
- "loss": 0.0004,
733
- "step": 144097
734
  },
735
  {
736
- "epoch": 10.41,
737
- "learning_rate": 5.330027495669035e-05,
738
- "loss": 0.0004,
739
- "step": 145496
740
  },
741
  {
742
- "epoch": 10.51,
743
- "learning_rate": 5.274440153212862e-05,
744
- "loss": 0.0004,
745
- "step": 146895
 
 
 
 
746
  },
747
  {
748
- "epoch": 10.61,
749
- "learning_rate": 5.218852810756687e-05,
750
- "loss": 0.0004,
751
- "step": 148294
752
  },
753
  {
754
- "epoch": 10.71,
755
- "learning_rate": 5.163265468300513e-05,
756
- "loss": 0.0004,
757
- "step": 149693
758
  },
759
  {
760
- "epoch": 10.81,
761
- "learning_rate": 5.107678125844339e-05,
762
- "loss": 0.0004,
763
- "step": 151092
764
  },
765
  {
766
- "epoch": 10.91,
767
- "learning_rate": 5.052090783388166e-05,
768
- "loss": 0.0004,
769
- "step": 152491
770
  },
771
  {
772
- "epoch": 11.0,
773
- "eval_loss": 0.019947679713368416,
774
- "eval_max_distance": 3,
775
- "eval_mean_distance": 0,
776
- "eval_runtime": 14.3656,
777
- "eval_samples_per_second": 18.656,
778
- "eval_steps_per_second": 1.253,
779
- "step": 153802
780
  },
781
  {
782
- "epoch": 11.01,
783
- "learning_rate": 4.9965034409319923e-05,
784
  "loss": 0.0004,
785
- "step": 153890
786
  },
787
  {
788
- "epoch": 11.11,
789
- "learning_rate": 4.9409160984758185e-05,
790
- "loss": 0.0003,
791
- "step": 155289
792
  },
793
  {
794
- "epoch": 11.21,
795
- "learning_rate": 4.885328756019645e-05,
796
- "loss": 0.0004,
797
- "step": 156688
798
  },
799
  {
800
- "epoch": 11.31,
801
- "learning_rate": 4.82974141356347e-05,
802
- "loss": 0.0004,
803
- "step": 158087
 
 
 
 
804
  },
805
  {
806
- "epoch": 11.41,
807
- "learning_rate": 4.774154071107297e-05,
808
- "loss": 0.0004,
809
- "step": 159486
810
  },
811
  {
812
- "epoch": 11.51,
813
- "learning_rate": 4.718566728651123e-05,
814
  "loss": 0.0004,
815
- "step": 160885
816
  },
817
  {
818
- "epoch": 11.61,
819
- "learning_rate": 4.6629793861949495e-05,
820
- "loss": 0.0004,
821
- "step": 162284
822
  },
823
  {
824
- "epoch": 11.71,
825
- "learning_rate": 4.607392043738775e-05,
826
- "loss": 0.0004,
827
- "step": 163683
828
  },
829
  {
830
- "epoch": 11.81,
831
- "learning_rate": 4.551804701282601e-05,
832
- "loss": 0.0004,
833
- "step": 165082
834
  },
835
  {
836
- "epoch": 11.91,
837
- "learning_rate": 4.496217358826428e-05,
838
- "loss": 0.0003,
839
- "step": 166481
840
  },
841
  {
842
- "epoch": 12.0,
843
- "eval_loss": 0.018351813778281212,
 
 
 
 
 
 
844
  "eval_max_distance": 3,
845
  "eval_mean_distance": 0,
846
- "eval_runtime": 14.3946,
847
- "eval_samples_per_second": 18.618,
848
- "eval_steps_per_second": 1.25,
849
- "step": 167784
850
  },
851
  {
852
- "epoch": 12.01,
853
- "learning_rate": 4.440630016370254e-05,
854
- "loss": 0.0003,
855
- "step": 167880
856
  },
857
  {
858
- "epoch": 12.11,
859
- "learning_rate": 4.3850426739140805e-05,
860
- "loss": 0.0004,
861
- "step": 169279
862
  },
863
  {
864
- "epoch": 12.21,
865
- "learning_rate": 4.329455331457906e-05,
866
- "loss": 0.0003,
867
- "step": 170678
868
  },
869
  {
870
- "epoch": 12.31,
871
- "learning_rate": 4.273867989001732e-05,
872
- "loss": 0.0003,
873
- "step": 172077
874
  },
875
  {
876
- "epoch": 12.41,
877
- "learning_rate": 4.218280646545559e-05,
878
- "loss": 0.0003,
879
- "step": 173476
880
  },
881
  {
882
- "epoch": 12.51,
883
- "learning_rate": 4.162693304089385e-05,
884
- "loss": 0.0003,
885
- "step": 174875
886
  },
887
  {
888
- "epoch": 12.61,
889
- "learning_rate": 4.1071059616332115e-05,
890
- "loss": 0.0003,
891
- "step": 176274
892
  },
893
  {
894
- "epoch": 12.71,
895
- "learning_rate": 4.051518619177037e-05,
896
- "loss": 0.0004,
897
- "step": 177673
898
  },
899
  {
900
- "epoch": 12.81,
901
- "learning_rate": 3.995931276720863e-05,
902
- "loss": 0.0003,
903
- "step": 179072
 
 
 
 
904
  },
905
  {
906
- "epoch": 12.91,
907
- "learning_rate": 3.94034393426469e-05,
908
- "loss": 0.0004,
909
- "step": 180471
910
  },
911
  {
912
- "epoch": 13.0,
913
- "eval_loss": 0.019636347889900208,
914
- "eval_max_distance": 3,
915
- "eval_mean_distance": 0,
916
- "eval_runtime": 14.3221,
917
- "eval_samples_per_second": 18.712,
918
- "eval_steps_per_second": 1.257,
919
- "step": 181766
920
  },
921
  {
922
- "epoch": 13.01,
923
- "learning_rate": 3.884756591808516e-05,
924
- "loss": 0.0004,
925
- "step": 181870
926
  },
927
  {
928
- "epoch": 13.11,
929
- "learning_rate": 3.8291692493523424e-05,
930
- "loss": 0.0004,
931
- "step": 183269
932
  },
933
  {
934
- "epoch": 13.21,
935
- "learning_rate": 3.773581906896168e-05,
936
- "loss": 0.0004,
937
- "step": 184668
938
  },
939
  {
940
- "epoch": 13.31,
941
- "learning_rate": 3.717994564439994e-05,
942
- "loss": 0.0003,
943
- "step": 186067
944
  },
945
  {
946
- "epoch": 13.41,
947
- "learning_rate": 3.662407221983821e-05,
948
- "loss": 0.0003,
949
- "step": 187466
950
  },
951
  {
952
- "epoch": 13.51,
953
- "learning_rate": 3.606819879527647e-05,
954
- "loss": 0.0003,
955
- "step": 188865
 
 
 
 
956
  },
957
  {
958
- "epoch": 13.61,
959
- "learning_rate": 3.551232537071473e-05,
960
- "loss": 0.0003,
961
- "step": 190264
962
  },
963
  {
964
- "epoch": 13.71,
965
- "learning_rate": 3.495645194615299e-05,
966
- "loss": 0.0004,
967
- "step": 191663
968
  },
969
  {
970
- "epoch": 13.81,
971
- "learning_rate": 3.440057852159125e-05,
972
- "loss": 0.0003,
973
- "step": 193062
974
  },
975
  {
976
- "epoch": 13.91,
977
- "learning_rate": 3.384470509702952e-05,
978
- "loss": 0.0003,
979
- "step": 194461
980
  },
981
  {
982
- "epoch": 14.0,
983
- "eval_loss": 0.01673734560608864,
984
- "eval_max_distance": 1,
985
- "eval_mean_distance": 0,
986
- "eval_runtime": 14.3426,
987
- "eval_samples_per_second": 18.686,
988
- "eval_steps_per_second": 1.255,
989
- "step": 195748
990
  },
991
  {
992
- "epoch": 14.01,
993
- "learning_rate": 3.328883167246778e-05,
994
- "loss": 0.0004,
995
- "step": 195860
996
  },
997
  {
998
- "epoch": 14.11,
999
- "learning_rate": 3.273295824790604e-05,
1000
- "loss": 0.0003,
1001
- "step": 197259
1002
  },
1003
  {
1004
- "epoch": 14.21,
1005
- "learning_rate": 3.21770848233443e-05,
1006
- "loss": 0.0003,
1007
- "step": 198658
1008
  },
1009
  {
1010
- "epoch": 14.31,
1011
- "learning_rate": 3.162121139878256e-05,
1012
- "loss": 0.0003,
1013
- "step": 200057
 
 
 
 
1014
  },
1015
  {
1016
- "epoch": 14.41,
1017
- "learning_rate": 3.106533797422082e-05,
1018
- "loss": 0.0003,
1019
- "step": 201456
1020
  },
1021
  {
1022
- "epoch": 14.51,
1023
- "learning_rate": 3.0509464549659088e-05,
1024
- "loss": 0.0003,
1025
- "step": 202855
1026
  },
1027
  {
1028
- "epoch": 14.61,
1029
- "learning_rate": 2.9953591125097347e-05,
1030
- "loss": 0.0003,
1031
- "step": 204254
1032
  },
1033
  {
1034
- "epoch": 14.71,
1035
- "learning_rate": 2.939771770053561e-05,
1036
- "loss": 0.0003,
1037
- "step": 205653
1038
  },
1039
  {
1040
- "epoch": 14.81,
1041
- "learning_rate": 2.8841844275973874e-05,
1042
- "loss": 0.0004,
1043
- "step": 207052
1044
  },
1045
  {
1046
- "epoch": 14.91,
1047
- "learning_rate": 2.8285970851412136e-05,
1048
- "loss": 0.0003,
1049
- "step": 208451
1050
  },
1051
  {
1052
- "epoch": 15.0,
1053
- "eval_loss": 0.017744125798344612,
 
 
 
 
 
 
 
 
 
 
 
 
1054
  "eval_max_distance": 3,
1055
  "eval_mean_distance": 0,
1056
- "eval_runtime": 14.4379,
1057
- "eval_samples_per_second": 18.562,
1058
- "eval_steps_per_second": 1.247,
1059
- "step": 209730
1060
  },
1061
  {
1062
- "epoch": 15.01,
1063
- "learning_rate": 2.7730097426850398e-05,
1064
- "loss": 0.0003,
1065
- "step": 209850
1066
  },
1067
  {
1068
- "epoch": 15.11,
1069
- "learning_rate": 2.7174224002288656e-05,
1070
- "loss": 0.0003,
1071
- "step": 211249
1072
  },
1073
  {
1074
- "epoch": 15.21,
1075
- "learning_rate": 2.661835057772692e-05,
1076
- "loss": 0.0003,
1077
- "step": 212648
1078
  },
1079
  {
1080
- "epoch": 15.31,
1081
- "learning_rate": 2.6062477153165184e-05,
1082
- "loss": 0.0003,
1083
- "step": 214047
1084
  },
1085
  {
1086
- "epoch": 15.41,
1087
- "learning_rate": 2.5506603728603446e-05,
1088
- "loss": 0.0003,
1089
- "step": 215446
1090
  },
1091
  {
1092
- "epoch": 15.51,
1093
- "learning_rate": 2.4950730304041704e-05,
1094
- "loss": 0.0003,
1095
- "step": 216845
1096
  },
1097
  {
1098
- "epoch": 15.61,
1099
- "learning_rate": 2.439485687947997e-05,
1100
- "loss": 0.0003,
1101
- "step": 218244
1102
  },
1103
  {
1104
- "epoch": 15.71,
1105
- "learning_rate": 2.3838983454918228e-05,
1106
- "loss": 0.0003,
1107
- "step": 219643
 
 
 
 
1108
  },
1109
  {
1110
- "epoch": 15.81,
1111
- "learning_rate": 2.3283110030356493e-05,
1112
- "loss": 0.0003,
1113
- "step": 221042
1114
  },
1115
  {
1116
- "epoch": 15.91,
1117
- "learning_rate": 2.2727236605794752e-05,
1118
- "loss": 0.0003,
1119
- "step": 222441
1120
  },
1121
  {
1122
- "epoch": 16.0,
1123
- "eval_loss": 0.01732315868139267,
1124
- "eval_max_distance": 3,
1125
- "eval_mean_distance": 0,
1126
- "eval_runtime": 14.4573,
1127
- "eval_samples_per_second": 18.537,
1128
- "eval_steps_per_second": 1.245,
1129
- "step": 223712
1130
  },
1131
  {
1132
- "epoch": 16.01,
1133
- "learning_rate": 2.2171363181233014e-05,
1134
- "loss": 0.0003,
1135
- "step": 223840
1136
  },
1137
  {
1138
- "epoch": 16.11,
1139
- "learning_rate": 2.1615489756671276e-05,
1140
- "loss": 0.0003,
1141
- "step": 225239
1142
  },
1143
  {
1144
- "epoch": 16.21,
1145
- "learning_rate": 2.1059616332109538e-05,
1146
- "loss": 0.0003,
1147
- "step": 226638
1148
  },
1149
  {
1150
- "epoch": 16.31,
1151
- "learning_rate": 2.0503742907547803e-05,
1152
- "loss": 0.0003,
1153
- "step": 228037
1154
  },
1155
  {
1156
- "epoch": 16.41,
1157
- "learning_rate": 1.994786948298606e-05,
1158
- "loss": 0.0003,
1159
- "step": 229436
1160
  },
1161
  {
1162
- "epoch": 16.51,
1163
- "learning_rate": 1.9391996058424324e-05,
1164
- "loss": 0.0003,
1165
- "step": 230835
 
 
 
 
1166
  },
1167
  {
1168
- "epoch": 16.61,
1169
- "learning_rate": 1.8836122633862586e-05,
1170
- "loss": 0.0003,
1171
- "step": 232234
1172
  },
1173
  {
1174
- "epoch": 16.71,
1175
- "learning_rate": 1.8280249209300847e-05,
1176
- "loss": 0.0003,
1177
- "step": 233633
1178
  },
1179
  {
1180
- "epoch": 16.81,
1181
- "learning_rate": 1.7724375784739113e-05,
1182
- "loss": 0.0003,
1183
- "step": 235032
1184
  },
1185
  {
1186
- "epoch": 16.91,
1187
- "learning_rate": 1.716850236017737e-05,
1188
- "loss": 0.0003,
1189
- "step": 236431
1190
  },
1191
  {
1192
- "epoch": 17.0,
1193
- "eval_loss": 0.01732039824128151,
1194
- "eval_max_distance": 3,
1195
- "eval_mean_distance": 0,
1196
- "eval_runtime": 14.4615,
1197
- "eval_samples_per_second": 18.532,
1198
- "eval_steps_per_second": 1.245,
1199
- "step": 237694
1200
  },
1201
  {
1202
- "epoch": 17.01,
1203
- "learning_rate": 1.6612628935615633e-05,
1204
- "loss": 0.0003,
1205
- "step": 237830
1206
  },
1207
  {
1208
- "epoch": 17.11,
1209
- "learning_rate": 1.6056755511053895e-05,
1210
- "loss": 0.0003,
1211
- "step": 239229
1212
  },
1213
  {
1214
- "epoch": 17.21,
1215
- "learning_rate": 1.5500882086492157e-05,
1216
- "loss": 0.0003,
1217
- "step": 240628
 
 
 
 
1218
  },
1219
  {
1220
- "epoch": 17.31,
1221
- "learning_rate": 1.4945008661930417e-05,
1222
- "loss": 0.0003,
1223
- "step": 242027
1224
  },
1225
  {
1226
- "epoch": 17.41,
1227
- "learning_rate": 1.4389135237368681e-05,
1228
- "loss": 0.0003,
1229
- "step": 243426
1230
  },
1231
  {
1232
- "epoch": 17.51,
1233
- "learning_rate": 1.3833261812806945e-05,
1234
- "loss": 0.0003,
1235
- "step": 244825
1236
  },
1237
  {
1238
- "epoch": 17.61,
1239
- "learning_rate": 1.3277388388245205e-05,
1240
- "loss": 0.0003,
1241
- "step": 246224
1242
  },
1243
  {
1244
- "epoch": 17.71,
1245
- "learning_rate": 1.2721514963683467e-05,
1246
- "loss": 0.0003,
1247
- "step": 247623
1248
  },
1249
  {
1250
- "epoch": 17.81,
1251
- "learning_rate": 1.2165641539121729e-05,
1252
- "loss": 0.0003,
1253
- "step": 249022
1254
  },
1255
  {
1256
- "epoch": 17.91,
1257
- "learning_rate": 1.160976811455999e-05,
1258
- "loss": 0.0003,
1259
- "step": 250421
1260
  },
1261
  {
1262
- "epoch": 18.0,
1263
- "eval_loss": 0.016897281631827354,
 
 
 
 
 
 
1264
  "eval_max_distance": 3,
1265
  "eval_mean_distance": 0,
1266
- "eval_runtime": 14.4833,
1267
- "eval_samples_per_second": 18.504,
1268
- "eval_steps_per_second": 1.243,
1269
- "step": 251676
1270
  },
1271
  {
1272
- "epoch": 18.01,
1273
- "learning_rate": 1.1053894689998253e-05,
1274
- "loss": 0.0003,
1275
- "step": 251820
1276
  },
1277
  {
1278
- "epoch": 18.11,
1279
- "learning_rate": 1.0498021265436515e-05,
1280
- "loss": 0.0003,
1281
- "step": 253219
1282
  },
1283
  {
1284
- "epoch": 18.21,
1285
- "learning_rate": 9.942147840874775e-06,
1286
- "loss": 0.0003,
1287
- "step": 254618
1288
  },
1289
  {
1290
- "epoch": 18.31,
1291
- "learning_rate": 9.386274416313037e-06,
1292
- "loss": 0.0003,
1293
- "step": 256017
1294
  },
1295
  {
1296
- "epoch": 18.41,
1297
- "learning_rate": 8.8304009917513e-06,
1298
- "loss": 0.0002,
1299
- "step": 257416
1300
  },
1301
  {
1302
- "epoch": 18.51,
1303
- "learning_rate": 8.274527567189562e-06,
1304
- "loss": 0.0003,
1305
- "step": 258815
1306
  },
1307
  {
1308
- "epoch": 18.61,
1309
- "learning_rate": 7.718654142627824e-06,
1310
- "loss": 0.0003,
1311
- "step": 260214
1312
  },
1313
  {
1314
- "epoch": 18.71,
1315
- "learning_rate": 7.1627807180660855e-06,
1316
- "loss": 0.0003,
1317
- "step": 261613
 
 
 
 
1318
  },
1319
  {
1320
- "epoch": 18.81,
1321
- "learning_rate": 6.606907293504347e-06,
1322
- "loss": 0.0002,
1323
- "step": 263012
1324
  },
1325
  {
1326
- "epoch": 18.91,
1327
- "learning_rate": 6.051033868942609e-06,
1328
  "loss": 0.0003,
1329
- "step": 264411
1330
  },
1331
  {
1332
- "epoch": 19.0,
1333
- "eval_loss": 0.016652461141347885,
1334
- "eval_max_distance": 3,
1335
- "eval_mean_distance": 0,
1336
- "eval_runtime": 14.4508,
1337
- "eval_samples_per_second": 18.546,
1338
- "eval_steps_per_second": 1.246,
1339
- "step": 265658
1340
  },
1341
  {
1342
- "epoch": 19.01,
1343
- "learning_rate": 5.4951604443808705e-06,
1344
- "loss": 0.0002,
1345
- "step": 265810
1346
  },
1347
  {
1348
- "epoch": 19.11,
1349
- "learning_rate": 4.9392870198191324e-06,
1350
- "loss": 0.0002,
1351
- "step": 267209
1352
  },
1353
  {
1354
- "epoch": 19.21,
1355
- "learning_rate": 4.383413595257395e-06,
1356
- "loss": 0.0003,
1357
- "step": 268608
1358
  },
1359
  {
1360
- "epoch": 19.31,
1361
- "learning_rate": 3.827540170695656e-06,
1362
- "loss": 0.0003,
1363
- "step": 270007
1364
  },
1365
  {
1366
- "epoch": 19.41,
1367
- "learning_rate": 3.2716667461339183e-06,
1368
- "loss": 0.0003,
1369
- "step": 271406
1370
  },
1371
  {
1372
- "epoch": 19.51,
1373
- "learning_rate": 2.71579332157218e-06,
1374
- "loss": 0.0003,
1375
- "step": 272805
 
 
 
 
1376
  },
1377
  {
1378
- "epoch": 19.61,
1379
- "learning_rate": 2.159919897010442e-06,
1380
- "loss": 0.0002,
1381
- "step": 274204
1382
  },
1383
  {
1384
- "epoch": 19.71,
1385
- "learning_rate": 1.604046472448704e-06,
1386
- "loss": 0.0003,
1387
- "step": 275603
1388
  },
1389
  {
1390
- "epoch": 19.81,
1391
- "learning_rate": 1.0481730478869658e-06,
1392
- "loss": 0.0003,
1393
- "step": 277002
1394
  },
1395
  {
1396
- "epoch": 19.91,
1397
- "learning_rate": 4.922996233252277e-07,
1398
- "loss": 0.0003,
1399
- "step": 278401
1400
  },
1401
  {
1402
- "epoch": 20.0,
1403
- "eval_loss": 0.017029350623488426,
1404
- "eval_max_distance": 3,
1405
- "eval_mean_distance": 0,
1406
- "eval_runtime": 14.4724,
1407
- "eval_samples_per_second": 18.518,
1408
- "eval_steps_per_second": 1.244,
1409
- "step": 279640
1410
  },
1411
  {
1412
- "epoch": 20.0,
1413
- "step": 279640,
1414
- "total_flos": 7.310805087246336e+16,
1415
- "train_loss": 0.00039881850996483783,
1416
- "train_runtime": 20247.2262,
1417
- "train_samples_per_second": 207.164,
1418
- "train_steps_per_second": 13.811
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1419
  }
1420
  ],
1421
- "logging_steps": 1399,
1422
- "max_steps": 279640,
1423
- "num_train_epochs": 20,
1424
- "save_steps": 2797,
1425
- "total_flos": 7.310805087246336e+16,
1426
  "trial_name": null,
1427
  "trial_params": null
1428
  }
 
1
  {
2
+ "best_metric": 0.0005441381363198161,
3
+ "best_model_checkpoint": "text-normalization-ru-new\\checkpoint-884",
4
+ "epoch": 25.0,
5
  "eval_steps": 500,
6
+ "global_step": 1700,
7
  "is_hyper_param_search": false,
8
  "is_local_process_zero": true,
9
  "is_world_process_zero": true,
10
  "log_history": [
11
  {
12
+ "epoch": 0.01,
13
+ "learning_rate": 5.882352941176471e-07,
14
+ "loss": 0.0269,
15
  "step": 1
16
  },
17
  {
18
+ "epoch": 0.13,
19
+ "learning_rate": 5.294117647058824e-06,
20
+ "loss": 0.032,
21
+ "step": 9
22
  },
23
  {
24
+ "epoch": 0.26,
25
+ "learning_rate": 1.0588235294117648e-05,
26
+ "loss": 0.0142,
27
+ "step": 18
 
 
 
 
 
 
28
  },
29
  {
30
  "epoch": 0.4,
31
+ "learning_rate": 1.588235294117647e-05,
32
+ "loss": 0.003,
33
+ "step": 27
34
  },
35
  {
36
+ "epoch": 0.53,
37
+ "learning_rate": 2.1176470588235296e-05,
38
+ "loss": 0.0185,
39
+ "step": 36
40
  },
41
  {
42
+ "epoch": 0.66,
43
+ "learning_rate": 2.647058823529412e-05,
44
+ "loss": 0.0431,
45
+ "step": 45
 
 
 
 
 
 
46
  },
47
  {
48
+ "epoch": 0.79,
49
+ "learning_rate": 3.176470588235294e-05,
50
+ "loss": 0.0182,
51
+ "step": 54
52
  },
53
  {
54
+ "epoch": 0.93,
55
+ "learning_rate": 3.705882352941177e-05,
56
+ "loss": 0.0368,
57
+ "step": 63
58
  },
59
  {
60
  "epoch": 1.0,
61
+ "eval_loss": 0.061875343322753906,
62
+ "eval_max_distance": 3,
63
  "eval_mean_distance": 0,
64
+ "eval_runtime": 1.7967,
65
+ "eval_samples_per_second": 42.855,
66
+ "eval_steps_per_second": 1.67,
67
+ "step": 68
 
 
 
 
 
 
68
  },
69
  {
70
+ "epoch": 1.06,
71
+ "learning_rate": 4.235294117647059e-05,
72
+ "loss": 0.0151,
73
+ "step": 72
 
 
 
 
 
 
74
  },
75
  {
76
+ "epoch": 1.19,
77
+ "learning_rate": 4.7647058823529414e-05,
78
+ "loss": 0.0193,
79
+ "step": 81
80
  },
81
  {
82
+ "epoch": 1.32,
83
+ "learning_rate": 5.294117647058824e-05,
84
+ "loss": 0.0076,
85
+ "step": 90
86
  },
87
  {
88
+ "epoch": 1.46,
89
+ "learning_rate": 5.823529411764707e-05,
90
+ "loss": 0.0061,
91
+ "step": 99
92
  },
93
  {
94
+ "epoch": 1.59,
95
+ "learning_rate": 6.352941176470588e-05,
96
+ "loss": 0.0111,
97
+ "step": 108
98
  },
99
  {
100
+ "epoch": 1.72,
101
+ "learning_rate": 6.882352941176471e-05,
102
+ "loss": 0.0314,
103
+ "step": 117
104
  },
105
  {
106
+ "epoch": 1.85,
107
+ "learning_rate": 7.411764705882354e-05,
108
+ "loss": 0.0253,
109
+ "step": 126
110
  },
111
  {
112
+ "epoch": 1.99,
113
+ "learning_rate": 7.941176470588235e-05,
114
+ "loss": 0.0057,
115
+ "step": 135
116
  },
117
  {
118
  "epoch": 2.0,
119
+ "eval_loss": 0.034646954387426376,
120
+ "eval_max_distance": 3,
121
  "eval_mean_distance": 0,
122
+ "eval_runtime": 1.8378,
123
+ "eval_samples_per_second": 41.898,
124
+ "eval_steps_per_second": 1.632,
125
+ "step": 136
126
  },
127
  {
128
+ "epoch": 2.12,
129
+ "learning_rate": 8.470588235294118e-05,
130
+ "loss": 0.0184,
131
+ "step": 144
132
  },
133
  {
134
+ "epoch": 2.25,
135
+ "learning_rate": 9e-05,
136
+ "loss": 0.0062,
137
+ "step": 153
138
  },
139
  {
140
+ "epoch": 2.38,
141
+ "learning_rate": 9.529411764705883e-05,
142
+ "loss": 0.0055,
143
+ "step": 162
144
  },
145
  {
146
+ "epoch": 2.51,
147
+ "learning_rate": 9.993464052287582e-05,
148
+ "loss": 0.0144,
149
+ "step": 171
150
  },
151
  {
152
+ "epoch": 2.65,
153
+ "learning_rate": 9.934640522875818e-05,
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
154
  "loss": 0.0006,
155
+ "step": 180
156
  },
157
  {
158
+ "epoch": 2.78,
159
+ "learning_rate": 9.875816993464053e-05,
160
+ "loss": 0.0101,
161
+ "step": 189
162
  },
163
  {
164
+ "epoch": 2.91,
165
+ "learning_rate": 9.816993464052288e-05,
166
+ "loss": 0.0113,
167
+ "step": 198
168
  },
169
  {
170
  "epoch": 3.0,
171
+ "eval_loss": 0.025189895182847977,
172
+ "eval_max_distance": 4,
173
  "eval_mean_distance": 0,
174
+ "eval_runtime": 1.7969,
175
+ "eval_samples_per_second": 42.851,
176
+ "eval_steps_per_second": 1.67,
177
+ "step": 204
178
  },
179
  {
180
+ "epoch": 3.04,
181
+ "learning_rate": 9.758169934640524e-05,
182
+ "loss": 0.0007,
183
+ "step": 207
184
  },
185
  {
186
+ "epoch": 3.18,
187
+ "learning_rate": 9.699346405228759e-05,
188
+ "loss": 0.0026,
189
+ "step": 216
190
  },
191
  {
192
+ "epoch": 3.31,
193
+ "learning_rate": 9.640522875816994e-05,
194
+ "loss": 0.0023,
195
+ "step": 225
196
  },
197
  {
198
+ "epoch": 3.44,
199
+ "learning_rate": 9.58169934640523e-05,
200
+ "loss": 0.0012,
201
+ "step": 234
202
  },
203
  {
204
+ "epoch": 3.57,
205
+ "learning_rate": 9.522875816993463e-05,
206
+ "loss": 0.0029,
207
+ "step": 243
208
  },
209
  {
210
+ "epoch": 3.71,
211
+ "learning_rate": 9.4640522875817e-05,
212
+ "loss": 0.0028,
213
+ "step": 252
214
  },
215
  {
216
+ "epoch": 3.84,
217
+ "learning_rate": 9.405228758169935e-05,
218
+ "loss": 0.0049,
219
+ "step": 261
 
 
 
 
 
 
220
  },
221
  {
222
+ "epoch": 3.97,
223
+ "learning_rate": 9.34640522875817e-05,
224
+ "loss": 0.0244,
225
+ "step": 270
 
 
 
 
 
 
226
  },
227
  {
228
  "epoch": 4.0,
229
+ "eval_loss": 0.011130807921290398,
230
  "eval_max_distance": 3,
231
  "eval_mean_distance": 0,
232
+ "eval_runtime": 1.9031,
233
+ "eval_samples_per_second": 40.459,
234
+ "eval_steps_per_second": 1.576,
235
+ "step": 272
 
 
 
 
 
 
236
  },
237
  {
238
  "epoch": 4.1,
239
+ "learning_rate": 9.287581699346406e-05,
240
+ "loss": 0.0163,
241
+ "step": 279
 
 
 
 
 
 
242
  },
243
  {
244
+ "epoch": 4.24,
245
+ "learning_rate": 9.228758169934641e-05,
246
+ "loss": 0.0035,
247
+ "step": 288
248
  },
249
  {
250
+ "epoch": 4.37,
251
+ "learning_rate": 9.169934640522877e-05,
252
+ "loss": 0.0011,
253
+ "step": 297
254
  },
255
  {
256
  "epoch": 4.5,
257
+ "learning_rate": 9.111111111111112e-05,
258
+ "loss": 0.0127,
259
+ "step": 306
260
  },
261
  {
262
+ "epoch": 4.63,
263
+ "learning_rate": 9.052287581699347e-05,
264
+ "loss": 0.0017,
265
+ "step": 315
266
  },
267
  {
268
+ "epoch": 4.76,
269
+ "learning_rate": 8.993464052287582e-05,
270
+ "loss": 0.0018,
271
+ "step": 324
 
 
 
 
 
 
272
  },
273
  {
274
  "epoch": 4.9,
275
+ "learning_rate": 8.934640522875818e-05,
276
+ "loss": 0.001,
277
+ "step": 333
278
  },
279
  {
280
  "epoch": 5.0,
281
+ "eval_loss": 0.006467848084867001,
282
  "eval_max_distance": 3,
283
  "eval_mean_distance": 0,
284
+ "eval_runtime": 1.7867,
285
+ "eval_samples_per_second": 43.096,
286
+ "eval_steps_per_second": 1.679,
287
+ "step": 340
288
  },
289
  {
290
+ "epoch": 5.03,
291
+ "learning_rate": 8.875816993464052e-05,
292
+ "loss": 0.0136,
293
+ "step": 342
294
  },
295
  {
296
+ "epoch": 5.16,
297
+ "learning_rate": 8.816993464052288e-05,
298
+ "loss": 0.0054,
299
+ "step": 351
300
  },
301
  {
302
+ "epoch": 5.29,
303
+ "learning_rate": 8.758169934640524e-05,
304
+ "loss": 0.0016,
305
+ "step": 360
306
  },
307
  {
308
+ "epoch": 5.43,
309
+ "learning_rate": 8.699346405228759e-05,
310
+ "loss": 0.0093,
311
+ "step": 369
312
  },
313
  {
314
+ "epoch": 5.56,
315
+ "learning_rate": 8.640522875816994e-05,
316
+ "loss": 0.0058,
317
+ "step": 378
 
 
 
 
 
 
 
 
 
 
 
 
318
  },
319
  {
320
+ "epoch": 5.69,
321
+ "learning_rate": 8.581699346405228e-05,
322
+ "loss": 0.0016,
323
+ "step": 387
324
  },
325
  {
326
+ "epoch": 5.82,
327
+ "learning_rate": 8.522875816993465e-05,
328
+ "loss": 0.0007,
329
+ "step": 396
330
  },
331
  {
332
+ "epoch": 5.96,
333
+ "learning_rate": 8.4640522875817e-05,
334
+ "loss": 0.0021,
335
+ "step": 405
336
  },
337
  {
338
  "epoch": 6.0,
339
+ "eval_loss": 0.0050986092537641525,
340
+ "eval_max_distance": 5,
341
  "eval_mean_distance": 0,
342
+ "eval_runtime": 1.8684,
343
+ "eval_samples_per_second": 41.212,
344
+ "eval_steps_per_second": 1.606,
345
+ "step": 408
346
  },
347
  {
348
+ "epoch": 6.09,
349
+ "learning_rate": 8.405228758169935e-05,
350
+ "loss": 0.0008,
351
+ "step": 414
352
  },
353
  {
354
+ "epoch": 6.22,
355
+ "learning_rate": 8.346405228758171e-05,
356
+ "loss": 0.0022,
357
+ "step": 423
358
  },
359
  {
360
+ "epoch": 6.35,
361
+ "learning_rate": 8.287581699346406e-05,
362
+ "loss": 0.0071,
363
+ "step": 432
364
+ },
365
+ {
366
+ "epoch": 6.49,
367
+ "learning_rate": 8.22875816993464e-05,
368
+ "loss": 0.004,
369
+ "step": 441
370
  },
371
  {
372
+ "epoch": 6.62,
373
+ "learning_rate": 8.169934640522877e-05,
374
  "loss": 0.0004,
375
+ "step": 450
376
  },
377
  {
378
+ "epoch": 6.75,
379
+ "learning_rate": 8.111111111111112e-05,
380
+ "loss": 0.0051,
381
+ "step": 459
382
  },
383
  {
384
+ "epoch": 6.88,
385
+ "learning_rate": 8.052287581699347e-05,
386
+ "loss": 0.0008,
387
+ "step": 468
388
  },
389
  {
390
+ "epoch": 7.0,
391
+ "eval_loss": 0.007741495501250029,
392
+ "eval_max_distance": 3,
393
+ "eval_mean_distance": 0,
394
+ "eval_runtime": 1.8063,
395
+ "eval_samples_per_second": 42.629,
396
+ "eval_steps_per_second": 1.661,
397
+ "step": 476
398
  },
399
  {
400
+ "epoch": 7.01,
401
+ "learning_rate": 7.993464052287583e-05,
402
+ "loss": 0.0071,
403
+ "step": 477
404
  },
405
  {
406
+ "epoch": 7.15,
407
+ "learning_rate": 7.934640522875816e-05,
408
+ "loss": 0.004,
409
+ "step": 486
410
  },
411
  {
412
+ "epoch": 7.28,
413
+ "learning_rate": 7.875816993464053e-05,
414
+ "loss": 0.0014,
415
+ "step": 495
416
  },
417
  {
418
+ "epoch": 7.41,
419
+ "learning_rate": 7.816993464052288e-05,
420
+ "loss": 0.003,
421
+ "step": 504
 
 
 
 
422
  },
423
  {
424
+ "epoch": 7.54,
425
+ "learning_rate": 7.758169934640524e-05,
426
+ "loss": 0.0008,
427
+ "step": 513
428
  },
429
  {
430
+ "epoch": 7.68,
431
+ "learning_rate": 7.699346405228759e-05,
432
+ "loss": 0.001,
433
+ "step": 522
434
  },
435
  {
436
+ "epoch": 7.81,
437
+ "learning_rate": 7.640522875816993e-05,
438
+ "loss": 0.0016,
439
+ "step": 531
440
  },
441
  {
442
+ "epoch": 7.94,
443
+ "learning_rate": 7.581699346405228e-05,
444
+ "loss": 0.0008,
445
+ "step": 540
446
  },
447
  {
448
+ "epoch": 8.0,
449
+ "eval_loss": 0.004006722941994667,
450
+ "eval_max_distance": 3,
451
+ "eval_mean_distance": 0,
452
+ "eval_runtime": 1.7939,
453
+ "eval_samples_per_second": 42.923,
454
+ "eval_steps_per_second": 1.672,
455
+ "step": 544
456
  },
457
  {
458
+ "epoch": 8.07,
459
+ "learning_rate": 7.522875816993465e-05,
460
+ "loss": 0.0008,
461
+ "step": 549
462
  },
463
  {
464
+ "epoch": 8.21,
465
+ "learning_rate": 7.4640522875817e-05,
466
+ "loss": 0.0015,
467
+ "step": 558
468
  },
469
  {
470
+ "epoch": 8.34,
471
+ "learning_rate": 7.405228758169935e-05,
472
+ "loss": 0.0018,
473
+ "step": 567
474
  },
475
  {
476
+ "epoch": 8.47,
477
+ "learning_rate": 7.346405228758171e-05,
478
+ "loss": 0.0015,
479
+ "step": 576
480
  },
481
  {
482
+ "epoch": 8.6,
483
+ "learning_rate": 7.287581699346405e-05,
484
+ "loss": 0.0028,
485
+ "step": 585
486
  },
487
  {
488
+ "epoch": 8.74,
489
+ "learning_rate": 7.228758169934641e-05,
490
+ "loss": 0.0017,
491
+ "step": 594
 
 
 
 
492
  },
493
  {
494
+ "epoch": 8.87,
495
+ "learning_rate": 7.169934640522877e-05,
496
+ "loss": 0.001,
497
+ "step": 603
498
  },
499
  {
500
+ "epoch": 9.0,
501
+ "learning_rate": 7.111111111111112e-05,
502
+ "loss": 0.0025,
503
+ "step": 612
504
  },
505
  {
506
+ "epoch": 9.0,
507
+ "eval_loss": 0.002946252468973398,
508
+ "eval_max_distance": 2,
509
+ "eval_mean_distance": 0,
510
+ "eval_runtime": 1.778,
511
+ "eval_samples_per_second": 43.306,
512
+ "eval_steps_per_second": 1.687,
513
+ "step": 612
514
  },
515
  {
516
+ "epoch": 9.13,
517
+ "learning_rate": 7.052287581699347e-05,
518
+ "loss": 0.0008,
519
+ "step": 621
520
  },
521
  {
522
+ "epoch": 9.26,
523
+ "learning_rate": 6.993464052287581e-05,
524
+ "loss": 0.0014,
525
+ "step": 630
526
  },
527
  {
528
+ "epoch": 9.4,
529
+ "learning_rate": 6.934640522875817e-05,
530
+ "loss": 0.0036,
531
+ "step": 639
532
  },
533
  {
534
+ "epoch": 9.53,
535
+ "learning_rate": 6.875816993464053e-05,
536
+ "loss": 0.0022,
537
+ "step": 648
538
  },
539
  {
540
+ "epoch": 9.66,
541
+ "learning_rate": 6.816993464052288e-05,
542
+ "loss": 0.0006,
543
+ "step": 657
544
  },
545
  {
546
+ "epoch": 9.79,
547
+ "learning_rate": 6.758169934640524e-05,
548
+ "loss": 0.0034,
549
+ "step": 666
550
  },
551
  {
552
+ "epoch": 9.93,
553
+ "learning_rate": 6.699346405228758e-05,
554
+ "loss": 0.0012,
555
+ "step": 675
556
  },
557
  {
558
+ "epoch": 10.0,
559
+ "eval_loss": 0.00195131148211658,
560
+ "eval_max_distance": 3,
561
  "eval_mean_distance": 0,
562
+ "eval_runtime": 1.7911,
563
+ "eval_samples_per_second": 42.99,
564
+ "eval_steps_per_second": 1.675,
565
+ "step": 680
566
  },
567
  {
568
+ "epoch": 10.06,
569
+ "learning_rate": 6.640522875816993e-05,
570
+ "loss": 0.0003,
571
+ "step": 684
572
  },
573
  {
574
+ "epoch": 10.19,
575
+ "learning_rate": 6.581699346405228e-05,
576
+ "loss": 0.0014,
577
+ "step": 693
578
  },
579
  {
580
+ "epoch": 10.32,
581
+ "learning_rate": 6.522875816993465e-05,
582
+ "loss": 0.0013,
583
+ "step": 702
584
  },
585
  {
586
+ "epoch": 10.46,
587
+ "learning_rate": 6.4640522875817e-05,
588
+ "loss": 0.0021,
589
+ "step": 711
590
  },
591
  {
592
+ "epoch": 10.59,
593
+ "learning_rate": 6.405228758169934e-05,
594
+ "loss": 0.0035,
595
+ "step": 720
596
  },
597
  {
598
+ "epoch": 10.72,
599
+ "learning_rate": 6.34640522875817e-05,
600
+ "loss": 0.0011,
601
+ "step": 729
602
  },
603
  {
604
+ "epoch": 10.85,
605
+ "learning_rate": 6.287581699346405e-05,
606
+ "loss": 0.0005,
607
+ "step": 738
608
  },
609
  {
610
+ "epoch": 10.99,
611
+ "learning_rate": 6.228758169934641e-05,
612
+ "loss": 0.0012,
613
+ "step": 747
614
  },
615
  {
616
+ "epoch": 11.0,
617
+ "eval_loss": 0.0009844042360782623,
618
+ "eval_max_distance": 3,
619
+ "eval_mean_distance": 0,
620
+ "eval_runtime": 1.7912,
621
+ "eval_samples_per_second": 42.989,
622
+ "eval_steps_per_second": 1.675,
623
+ "step": 748
624
  },
625
  {
626
+ "epoch": 11.12,
627
+ "learning_rate": 6.169934640522877e-05,
628
+ "loss": 0.0068,
629
+ "step": 756
630
  },
631
  {
632
+ "epoch": 11.25,
633
+ "learning_rate": 6.111111111111112e-05,
634
+ "loss": 0.001,
635
+ "step": 765
 
 
 
 
636
  },
637
  {
638
+ "epoch": 11.38,
639
+ "learning_rate": 6.052287581699346e-05,
640
+ "loss": 0.0009,
641
+ "step": 774
642
  },
643
  {
644
+ "epoch": 11.51,
645
+ "learning_rate": 5.993464052287582e-05,
646
+ "loss": 0.0002,
647
+ "step": 783
648
  },
649
  {
650
+ "epoch": 11.65,
651
+ "learning_rate": 5.934640522875817e-05,
652
+ "loss": 0.0014,
653
+ "step": 792
654
  },
655
  {
656
+ "epoch": 11.78,
657
+ "learning_rate": 5.8758169934640526e-05,
658
+ "loss": 0.0028,
659
+ "step": 801
660
  },
661
  {
662
+ "epoch": 11.91,
663
+ "learning_rate": 5.8169934640522886e-05,
664
+ "loss": 0.0006,
665
+ "step": 810
666
  },
667
  {
668
+ "epoch": 12.0,
669
+ "eval_loss": 0.0007958057103678584,
670
+ "eval_max_distance": 0,
671
+ "eval_mean_distance": 0,
672
+ "eval_runtime": 1.8244,
673
+ "eval_samples_per_second": 42.205,
674
+ "eval_steps_per_second": 1.644,
675
+ "step": 816
676
  },
677
  {
678
+ "epoch": 12.04,
679
+ "learning_rate": 5.7581699346405225e-05,
680
+ "loss": 0.0028,
681
+ "step": 819
682
  },
683
  {
684
+ "epoch": 12.18,
685
+ "learning_rate": 5.6993464052287585e-05,
686
+ "loss": 0.0015,
687
+ "step": 828
688
  },
689
  {
690
+ "epoch": 12.31,
691
+ "learning_rate": 5.640522875816994e-05,
692
+ "loss": 0.0009,
693
+ "step": 837
694
  },
695
  {
696
+ "epoch": 12.44,
697
+ "learning_rate": 5.581699346405229e-05,
698
+ "loss": 0.0018,
699
+ "step": 846
700
  },
701
  {
702
+ "epoch": 12.57,
703
+ "learning_rate": 5.5228758169934644e-05,
704
+ "loss": 0.0008,
705
+ "step": 855
 
 
 
 
706
  },
707
  {
708
+ "epoch": 12.71,
709
+ "learning_rate": 5.464052287581699e-05,
710
  "loss": 0.0004,
711
+ "step": 864
712
  },
713
  {
714
+ "epoch": 12.84,
715
+ "learning_rate": 5.405228758169934e-05,
716
+ "loss": 0.002,
717
+ "step": 873
718
  },
719
  {
720
+ "epoch": 12.97,
721
+ "learning_rate": 5.34640522875817e-05,
722
+ "loss": 0.0012,
723
+ "step": 882
724
  },
725
  {
726
+ "epoch": 13.0,
727
+ "eval_loss": 0.0005441381363198161,
728
+ "eval_max_distance": 0,
729
+ "eval_mean_distance": 0,
730
+ "eval_runtime": 1.7816,
731
+ "eval_samples_per_second": 43.219,
732
+ "eval_steps_per_second": 1.684,
733
+ "step": 884
734
  },
735
  {
736
+ "epoch": 13.1,
737
+ "learning_rate": 5.2875816993464056e-05,
738
+ "loss": 0.0013,
739
+ "step": 891
740
  },
741
  {
742
+ "epoch": 13.24,
743
+ "learning_rate": 5.228758169934641e-05,
744
  "loss": 0.0004,
745
+ "step": 900
746
  },
747
  {
748
+ "epoch": 13.37,
749
+ "learning_rate": 5.169934640522877e-05,
750
+ "loss": 0.0008,
751
+ "step": 909
752
  },
753
  {
754
+ "epoch": 13.5,
755
+ "learning_rate": 5.111111111111111e-05,
756
+ "loss": 0.0018,
757
+ "step": 918
758
  },
759
  {
760
+ "epoch": 13.63,
761
+ "learning_rate": 5.052287581699347e-05,
762
+ "loss": 0.0026,
763
+ "step": 927
764
  },
765
  {
766
+ "epoch": 13.76,
767
+ "learning_rate": 4.993464052287582e-05,
768
+ "loss": 0.0025,
769
+ "step": 936
770
  },
771
  {
772
+ "epoch": 13.9,
773
+ "learning_rate": 4.9346405228758174e-05,
774
+ "loss": 0.001,
775
+ "step": 945
776
+ },
777
+ {
778
+ "epoch": 14.0,
779
+ "eval_loss": 0.0007363850018009543,
780
  "eval_max_distance": 3,
781
  "eval_mean_distance": 0,
782
+ "eval_runtime": 1.7798,
783
+ "eval_samples_per_second": 43.263,
784
+ "eval_steps_per_second": 1.686,
785
+ "step": 952
786
  },
787
  {
788
+ "epoch": 14.03,
789
+ "learning_rate": 4.875816993464053e-05,
790
+ "loss": 0.0012,
791
+ "step": 954
792
  },
793
  {
794
+ "epoch": 14.16,
795
+ "learning_rate": 4.816993464052288e-05,
796
+ "loss": 0.0006,
797
+ "step": 963
798
  },
799
  {
800
+ "epoch": 14.29,
801
+ "learning_rate": 4.7581699346405226e-05,
802
+ "loss": 0.0022,
803
+ "step": 972
804
  },
805
  {
806
+ "epoch": 14.43,
807
+ "learning_rate": 4.6993464052287586e-05,
808
+ "loss": 0.001,
809
+ "step": 981
810
  },
811
  {
812
+ "epoch": 14.56,
813
+ "learning_rate": 4.640522875816994e-05,
814
+ "loss": 0.0015,
815
+ "step": 990
816
  },
817
  {
818
+ "epoch": 14.69,
819
+ "learning_rate": 4.5816993464052285e-05,
820
+ "loss": 0.0015,
821
+ "step": 999
822
  },
823
  {
824
+ "epoch": 14.82,
825
+ "learning_rate": 4.5228758169934645e-05,
826
+ "loss": 0.0013,
827
+ "step": 1008
828
  },
829
  {
830
+ "epoch": 14.96,
831
+ "learning_rate": 4.4640522875817e-05,
832
+ "loss": 0.0009,
833
+ "step": 1017
834
  },
835
  {
836
+ "epoch": 15.0,
837
+ "eval_loss": 0.0013804116751998663,
838
+ "eval_max_distance": 3,
839
+ "eval_mean_distance": 0,
840
+ "eval_runtime": 1.7896,
841
+ "eval_samples_per_second": 43.027,
842
+ "eval_steps_per_second": 1.676,
843
+ "step": 1020
844
  },
845
  {
846
+ "epoch": 15.09,
847
+ "learning_rate": 4.405228758169935e-05,
848
+ "loss": 0.0013,
849
+ "step": 1026
850
  },
851
  {
852
+ "epoch": 15.22,
853
+ "learning_rate": 4.3464052287581704e-05,
854
+ "loss": 0.0011,
855
+ "step": 1035
 
 
 
 
856
  },
857
  {
858
+ "epoch": 15.35,
859
+ "learning_rate": 4.287581699346405e-05,
860
+ "loss": 0.0024,
861
+ "step": 1044
862
  },
863
  {
864
+ "epoch": 15.49,
865
+ "learning_rate": 4.228758169934641e-05,
866
+ "loss": 0.0006,
867
+ "step": 1053
868
  },
869
  {
870
+ "epoch": 15.62,
871
+ "learning_rate": 4.169934640522876e-05,
872
+ "loss": 0.0013,
873
+ "step": 1062
874
  },
875
  {
876
+ "epoch": 15.75,
877
+ "learning_rate": 4.111111111111111e-05,
878
+ "loss": 0.0012,
879
+ "step": 1071
880
  },
881
  {
882
+ "epoch": 15.88,
883
+ "learning_rate": 4.052287581699347e-05,
884
+ "loss": 0.0005,
885
+ "step": 1080
886
  },
887
  {
888
+ "epoch": 16.0,
889
+ "eval_loss": 0.0015163180651143193,
890
+ "eval_max_distance": 3,
891
+ "eval_mean_distance": 0,
892
+ "eval_runtime": 1.7783,
893
+ "eval_samples_per_second": 43.3,
894
+ "eval_steps_per_second": 1.687,
895
+ "step": 1088
896
  },
897
  {
898
+ "epoch": 16.01,
899
+ "learning_rate": 3.993464052287582e-05,
900
+ "loss": 0.0012,
901
+ "step": 1089
902
  },
903
  {
904
+ "epoch": 16.15,
905
+ "learning_rate": 3.934640522875817e-05,
906
+ "loss": 0.0006,
907
+ "step": 1098
908
  },
909
  {
910
+ "epoch": 16.28,
911
+ "learning_rate": 3.875816993464053e-05,
912
+ "loss": 0.0018,
913
+ "step": 1107
914
  },
915
  {
916
+ "epoch": 16.41,
917
+ "learning_rate": 3.8169934640522874e-05,
918
+ "loss": 0.0008,
919
+ "step": 1116
920
  },
921
  {
922
+ "epoch": 16.54,
923
+ "learning_rate": 3.758169934640523e-05,
924
+ "loss": 0.0006,
925
+ "step": 1125
 
 
 
 
926
  },
927
  {
928
+ "epoch": 16.68,
929
+ "learning_rate": 3.699346405228759e-05,
930
+ "loss": 0.0007,
931
+ "step": 1134
932
  },
933
  {
934
+ "epoch": 16.81,
935
+ "learning_rate": 3.640522875816993e-05,
936
+ "loss": 0.0013,
937
+ "step": 1143
938
  },
939
  {
940
+ "epoch": 16.94,
941
+ "learning_rate": 3.581699346405229e-05,
942
+ "loss": 0.0009,
943
+ "step": 1152
944
  },
945
  {
946
+ "epoch": 17.0,
947
+ "eval_loss": 0.0014126226305961609,
948
+ "eval_max_distance": 3,
949
+ "eval_mean_distance": 0,
950
+ "eval_runtime": 1.7701,
951
+ "eval_samples_per_second": 43.501,
952
+ "eval_steps_per_second": 1.695,
953
+ "step": 1156
954
  },
955
  {
956
+ "epoch": 17.07,
957
+ "learning_rate": 3.5228758169934646e-05,
958
+ "loss": 0.0016,
959
+ "step": 1161
960
  },
961
  {
962
+ "epoch": 17.21,
963
+ "learning_rate": 3.464052287581699e-05,
964
+ "loss": 0.0008,
965
+ "step": 1170
966
  },
967
  {
968
+ "epoch": 17.34,
969
+ "learning_rate": 3.405228758169935e-05,
970
+ "loss": 0.001,
971
+ "step": 1179
972
  },
973
  {
974
+ "epoch": 17.47,
975
+ "learning_rate": 3.34640522875817e-05,
976
+ "loss": 0.0006,
977
+ "step": 1188
978
  },
979
  {
980
+ "epoch": 17.6,
981
+ "learning_rate": 3.287581699346405e-05,
982
+ "loss": 0.0025,
983
+ "step": 1197
984
  },
985
  {
986
+ "epoch": 17.74,
987
+ "learning_rate": 3.228758169934641e-05,
988
+ "loss": 0.0011,
989
+ "step": 1206
990
  },
991
  {
992
+ "epoch": 17.87,
993
+ "learning_rate": 3.169934640522876e-05,
994
+ "loss": 0.0008,
995
+ "step": 1215
996
+ },
997
+ {
998
+ "epoch": 18.0,
999
+ "learning_rate": 3.111111111111111e-05,
1000
+ "loss": 0.0006,
1001
+ "step": 1224
1002
+ },
1003
+ {
1004
+ "epoch": 18.0,
1005
+ "eval_loss": 0.0012967990478500724,
1006
  "eval_max_distance": 3,
1007
  "eval_mean_distance": 0,
1008
+ "eval_runtime": 1.7808,
1009
+ "eval_samples_per_second": 43.238,
1010
+ "eval_steps_per_second": 1.685,
1011
+ "step": 1224
1012
  },
1013
  {
1014
+ "epoch": 18.13,
1015
+ "learning_rate": 3.052287581699347e-05,
1016
+ "loss": 0.0011,
1017
+ "step": 1233
1018
  },
1019
  {
1020
+ "epoch": 18.26,
1021
+ "learning_rate": 2.9934640522875816e-05,
1022
+ "loss": 0.0009,
1023
+ "step": 1242
1024
  },
1025
  {
1026
+ "epoch": 18.4,
1027
+ "learning_rate": 2.9346405228758172e-05,
1028
+ "loss": 0.0009,
1029
+ "step": 1251
1030
  },
1031
  {
1032
+ "epoch": 18.53,
1033
+ "learning_rate": 2.8758169934640522e-05,
1034
+ "loss": 0.0008,
1035
+ "step": 1260
1036
  },
1037
  {
1038
+ "epoch": 18.66,
1039
+ "learning_rate": 2.8169934640522878e-05,
1040
+ "loss": 0.0005,
1041
+ "step": 1269
1042
  },
1043
  {
1044
+ "epoch": 18.79,
1045
+ "learning_rate": 2.758169934640523e-05,
1046
+ "loss": 0.001,
1047
+ "step": 1278
1048
  },
1049
  {
1050
+ "epoch": 18.93,
1051
+ "learning_rate": 2.699346405228758e-05,
1052
+ "loss": 0.0012,
1053
+ "step": 1287
1054
  },
1055
  {
1056
+ "epoch": 19.0,
1057
+ "eval_loss": 0.001441239845007658,
1058
+ "eval_max_distance": 3,
1059
+ "eval_mean_distance": 0,
1060
+ "eval_runtime": 1.824,
1061
+ "eval_samples_per_second": 42.215,
1062
+ "eval_steps_per_second": 1.645,
1063
+ "step": 1292
1064
  },
1065
  {
1066
+ "epoch": 19.06,
1067
+ "learning_rate": 2.6405228758169937e-05,
1068
+ "loss": 0.0008,
1069
+ "step": 1296
1070
  },
1071
  {
1072
+ "epoch": 19.19,
1073
+ "learning_rate": 2.581699346405229e-05,
1074
+ "loss": 0.0009,
1075
+ "step": 1305
1076
  },
1077
  {
1078
+ "epoch": 19.32,
1079
+ "learning_rate": 2.522875816993464e-05,
1080
+ "loss": 0.0009,
1081
+ "step": 1314
 
 
 
 
1082
  },
1083
  {
1084
+ "epoch": 19.46,
1085
+ "learning_rate": 2.4640522875816996e-05,
1086
+ "loss": 0.0002,
1087
+ "step": 1323
1088
  },
1089
  {
1090
+ "epoch": 19.59,
1091
+ "learning_rate": 2.405228758169935e-05,
1092
+ "loss": 0.0006,
1093
+ "step": 1332
1094
  },
1095
  {
1096
+ "epoch": 19.72,
1097
+ "learning_rate": 2.34640522875817e-05,
1098
+ "loss": 0.0004,
1099
+ "step": 1341
1100
  },
1101
  {
1102
+ "epoch": 19.85,
1103
+ "learning_rate": 2.2875816993464052e-05,
1104
+ "loss": 0.0006,
1105
+ "step": 1350
1106
  },
1107
  {
1108
+ "epoch": 19.99,
1109
+ "learning_rate": 2.2287581699346408e-05,
1110
+ "loss": 0.0007,
1111
+ "step": 1359
1112
  },
1113
  {
1114
+ "epoch": 20.0,
1115
+ "eval_loss": 0.0013804284390062094,
1116
+ "eval_max_distance": 3,
1117
+ "eval_mean_distance": 0,
1118
+ "eval_runtime": 1.9155,
1119
+ "eval_samples_per_second": 40.198,
1120
+ "eval_steps_per_second": 1.566,
1121
+ "step": 1360
1122
  },
1123
  {
1124
+ "epoch": 20.12,
1125
+ "learning_rate": 2.169934640522876e-05,
1126
+ "loss": 0.0013,
1127
+ "step": 1368
1128
  },
1129
  {
1130
+ "epoch": 20.25,
1131
+ "learning_rate": 2.111111111111111e-05,
1132
+ "loss": 0.0014,
1133
+ "step": 1377
1134
  },
1135
  {
1136
+ "epoch": 20.38,
1137
+ "learning_rate": 2.0522875816993464e-05,
1138
+ "loss": 0.0001,
1139
+ "step": 1386
1140
  },
1141
  {
1142
+ "epoch": 20.51,
1143
+ "learning_rate": 1.993464052287582e-05,
1144
+ "loss": 0.0015,
1145
+ "step": 1395
1146
  },
1147
  {
1148
+ "epoch": 20.65,
1149
+ "learning_rate": 1.9346405228758173e-05,
1150
+ "loss": 0.0008,
1151
+ "step": 1404
 
 
 
 
1152
  },
1153
  {
1154
+ "epoch": 20.78,
1155
+ "learning_rate": 1.8758169934640523e-05,
1156
+ "loss": 0.0006,
1157
+ "step": 1413
1158
  },
1159
  {
1160
+ "epoch": 20.91,
1161
+ "learning_rate": 1.8169934640522876e-05,
1162
+ "loss": 0.0012,
1163
+ "step": 1422
1164
  },
1165
  {
1166
+ "epoch": 21.0,
1167
+ "eval_loss": 0.0013173692859709263,
1168
+ "eval_max_distance": 3,
1169
+ "eval_mean_distance": 0,
1170
+ "eval_runtime": 1.8242,
1171
+ "eval_samples_per_second": 42.21,
1172
+ "eval_steps_per_second": 1.645,
1173
+ "step": 1428
1174
  },
1175
  {
1176
+ "epoch": 21.04,
1177
+ "learning_rate": 1.7581699346405232e-05,
1178
+ "loss": 0.0027,
1179
+ "step": 1431
1180
  },
1181
  {
1182
+ "epoch": 21.18,
1183
+ "learning_rate": 1.6993464052287582e-05,
1184
+ "loss": 0.0008,
1185
+ "step": 1440
1186
  },
1187
  {
1188
+ "epoch": 21.31,
1189
+ "learning_rate": 1.6405228758169935e-05,
1190
+ "loss": 0.0006,
1191
+ "step": 1449
1192
  },
1193
  {
1194
+ "epoch": 21.44,
1195
+ "learning_rate": 1.5816993464052288e-05,
1196
+ "loss": 0.0002,
1197
+ "step": 1458
1198
  },
1199
  {
1200
+ "epoch": 21.57,
1201
+ "learning_rate": 1.522875816993464e-05,
1202
+ "loss": 0.001,
1203
+ "step": 1467
1204
  },
1205
  {
1206
+ "epoch": 21.71,
1207
+ "learning_rate": 1.4640522875816995e-05,
1208
+ "loss": 0.0012,
1209
+ "step": 1476
1210
  },
1211
  {
1212
+ "epoch": 21.84,
1213
+ "learning_rate": 1.4052287581699347e-05,
1214
+ "loss": 0.0008,
1215
+ "step": 1485
1216
  },
1217
  {
1218
+ "epoch": 21.97,
1219
+ "learning_rate": 1.34640522875817e-05,
1220
+ "loss": 0.0013,
1221
+ "step": 1494
1222
+ },
1223
+ {
1224
+ "epoch": 22.0,
1225
+ "eval_loss": 0.001370716723613441,
1226
  "eval_max_distance": 3,
1227
  "eval_mean_distance": 0,
1228
+ "eval_runtime": 1.8261,
1229
+ "eval_samples_per_second": 42.166,
1230
+ "eval_steps_per_second": 1.643,
1231
+ "step": 1496
1232
  },
1233
  {
1234
+ "epoch": 22.1,
1235
+ "learning_rate": 1.2875816993464051e-05,
1236
+ "loss": 0.0004,
1237
+ "step": 1503
1238
  },
1239
  {
1240
+ "epoch": 22.24,
1241
+ "learning_rate": 1.2287581699346406e-05,
1242
+ "loss": 0.001,
1243
+ "step": 1512
1244
  },
1245
  {
1246
+ "epoch": 22.37,
1247
+ "learning_rate": 1.1699346405228759e-05,
1248
+ "loss": 0.0008,
1249
+ "step": 1521
1250
  },
1251
  {
1252
+ "epoch": 22.5,
1253
+ "learning_rate": 1.1111111111111112e-05,
1254
+ "loss": 0.0013,
1255
+ "step": 1530
1256
  },
1257
  {
1258
+ "epoch": 22.63,
1259
+ "learning_rate": 1.0522875816993465e-05,
1260
+ "loss": 0.0007,
1261
+ "step": 1539
1262
  },
1263
  {
1264
+ "epoch": 22.76,
1265
+ "learning_rate": 9.934640522875818e-06,
1266
+ "loss": 0.0013,
1267
+ "step": 1548
1268
  },
1269
  {
1270
+ "epoch": 22.9,
1271
+ "learning_rate": 9.34640522875817e-06,
1272
+ "loss": 0.0008,
1273
+ "step": 1557
1274
  },
1275
  {
1276
+ "epoch": 23.0,
1277
+ "eval_loss": 0.001398879336193204,
1278
+ "eval_max_distance": 3,
1279
+ "eval_mean_distance": 0,
1280
+ "eval_runtime": 1.8234,
1281
+ "eval_samples_per_second": 42.23,
1282
+ "eval_steps_per_second": 1.645,
1283
+ "step": 1564
1284
  },
1285
  {
1286
+ "epoch": 23.03,
1287
+ "learning_rate": 8.758169934640522e-06,
1288
+ "loss": 0.0025,
1289
+ "step": 1566
1290
  },
1291
  {
1292
+ "epoch": 23.16,
1293
+ "learning_rate": 8.169934640522877e-06,
1294
  "loss": 0.0003,
1295
+ "step": 1575
1296
  },
1297
  {
1298
+ "epoch": 23.29,
1299
+ "learning_rate": 7.581699346405229e-06,
1300
+ "loss": 0.0009,
1301
+ "step": 1584
 
 
 
 
1302
  },
1303
  {
1304
+ "epoch": 23.43,
1305
+ "learning_rate": 6.993464052287582e-06,
1306
+ "loss": 0.0011,
1307
+ "step": 1593
1308
  },
1309
  {
1310
+ "epoch": 23.56,
1311
+ "learning_rate": 6.405228758169935e-06,
1312
+ "loss": 0.0026,
1313
+ "step": 1602
1314
  },
1315
  {
1316
+ "epoch": 23.69,
1317
+ "learning_rate": 5.816993464052288e-06,
1318
+ "loss": 0.0013,
1319
+ "step": 1611
1320
  },
1321
  {
1322
+ "epoch": 23.82,
1323
+ "learning_rate": 5.228758169934641e-06,
1324
+ "loss": 0.0012,
1325
+ "step": 1620
1326
  },
1327
  {
1328
+ "epoch": 23.96,
1329
+ "learning_rate": 4.640522875816994e-06,
1330
+ "loss": 0.0005,
1331
+ "step": 1629
1332
  },
1333
  {
1334
+ "epoch": 24.0,
1335
+ "eval_loss": 0.0014485500287264585,
1336
+ "eval_max_distance": 3,
1337
+ "eval_mean_distance": 0,
1338
+ "eval_runtime": 1.8281,
1339
+ "eval_samples_per_second": 42.12,
1340
+ "eval_steps_per_second": 1.641,
1341
+ "step": 1632
1342
  },
1343
  {
1344
+ "epoch": 24.09,
1345
+ "learning_rate": 4.052287581699347e-06,
1346
+ "loss": 0.0015,
1347
+ "step": 1638
1348
  },
1349
  {
1350
+ "epoch": 24.22,
1351
+ "learning_rate": 3.4640522875816997e-06,
1352
+ "loss": 0.0013,
1353
+ "step": 1647
1354
  },
1355
  {
1356
+ "epoch": 24.35,
1357
+ "learning_rate": 2.8758169934640523e-06,
1358
+ "loss": 0.0019,
1359
+ "step": 1656
1360
  },
1361
  {
1362
+ "epoch": 24.49,
1363
+ "learning_rate": 2.2875816993464053e-06,
1364
+ "loss": 0.0005,
1365
+ "step": 1665
1366
  },
1367
  {
1368
+ "epoch": 24.62,
1369
+ "learning_rate": 1.699346405228758e-06,
1370
+ "loss": 0.0012,
1371
+ "step": 1674
 
 
 
 
1372
  },
1373
  {
1374
+ "epoch": 24.75,
1375
+ "learning_rate": 1.1111111111111112e-06,
1376
+ "loss": 0.001,
1377
+ "step": 1683
1378
+ },
1379
+ {
1380
+ "epoch": 24.88,
1381
+ "learning_rate": 5.22875816993464e-07,
1382
+ "loss": 0.001,
1383
+ "step": 1692
1384
+ },
1385
+ {
1386
+ "epoch": 25.0,
1387
+ "eval_loss": 0.0014130847994238138,
1388
+ "eval_max_distance": 3,
1389
+ "eval_mean_distance": 0,
1390
+ "eval_runtime": 1.9218,
1391
+ "eval_samples_per_second": 40.066,
1392
+ "eval_steps_per_second": 1.561,
1393
+ "step": 1700
1394
+ },
1395
+ {
1396
+ "epoch": 25.0,
1397
+ "step": 1700,
1398
+ "total_flos": 456120480651264.0,
1399
+ "train_loss": 0.003576931923401871,
1400
+ "train_runtime": 194.6035,
1401
+ "train_samples_per_second": 259.373,
1402
+ "train_steps_per_second": 8.736
1403
  }
1404
  ],
1405
+ "logging_steps": 9,
1406
+ "max_steps": 1700,
1407
+ "num_train_epochs": 25,
1408
+ "save_steps": 17,
1409
+ "total_flos": 456120480651264.0,
1410
  "trial_name": null,
1411
  "trial_params": null
1412
  }
training_args.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:7d8a3d0207b71233d440dfc4ea0e1be195fbf0ff77bc164bd8950b614fbebbd5
3
- size 4091
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e3766c94731270598b629b2e54dbd149eec47873f81720b047d2a1d23148ad6e
3
+ size 4155