craa commited on
Commit
ca10982
·
verified ·
1 Parent(s): a9edf0c

Training in progress, step 10000, checkpoint

Browse files
checkpoint-10000/model.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:8305ddb7dcf6cfe5a35bb48a2f25351223e7b9f898277680d5d6fa2d3db2244f
3
  size 503128704
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:53750843de32f8db3e77747274fbf050884944ef260742015078d10740e9e435
3
  size 503128704
checkpoint-10000/optimizer.pt CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:14d6af0b69f631e4761ab098bfa9fa3774cebaea8a68bbe42d54414319db9409
3
  size 1006351290
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:2789df46a3e32c34a8c21b6bfdf22dae483b2ed113da178edd743064e0580477
3
  size 1006351290
checkpoint-10000/rng_state.pth CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:c8108d586ec292d3abb40856030c3b6a3ed2558a666cffbe249076f11dc57a94
3
  size 14244
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:1abdd4c7c045dcb1f8a6404bddc83e4806c8d3e3173eb1017f9b082b3c161e85
3
  size 14244
checkpoint-10000/scheduler.pt CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:e29c5a364050c2544493214aea5be06ca5b077a072072b7010c5f1d14e8ffa82
3
  size 1064
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:baaddc0dc177f093cb7927eae851400a36fde6f45129bcaf2df443bdde5269f4
3
  size 1064
checkpoint-10000/trainer_state.json CHANGED
@@ -1,7 +1,7 @@
1
  {
2
- "best_metric": 3.7440133094787598,
3
- "best_model_checkpoint": "/scratch/cl5625/exceptions/models/100M_495/checkpoint-10000",
4
- "epoch": 1.0763104079216446,
5
  "eval_steps": 1000,
6
  "global_step": 10000,
7
  "is_hyper_param_search": false,
@@ -9,1498 +9,1498 @@
9
  "is_world_process_zero": true,
10
  "log_history": [
11
  {
12
- "epoch": 0.005381552039608223,
13
- "grad_norm": 1.6117655038833618,
14
- "learning_rate": 0.00027,
15
- "loss": 8.7843,
16
  "step": 50
17
  },
18
  {
19
- "epoch": 0.010763104079216447,
20
- "grad_norm": 1.1765116453170776,
21
- "learning_rate": 0.00057,
22
- "loss": 6.9711,
23
  "step": 100
24
  },
25
  {
26
- "epoch": 0.01614465611882467,
27
- "grad_norm": 1.6633868217468262,
28
- "learning_rate": 0.0005997090830729447,
29
- "loss": 6.5373,
30
  "step": 150
31
  },
32
  {
33
- "epoch": 0.021526208158432893,
34
- "grad_norm": 1.8836766481399536,
35
- "learning_rate": 0.0005993858420428833,
36
- "loss": 6.2429,
37
  "step": 200
38
  },
39
  {
40
- "epoch": 0.026907760198041114,
41
- "grad_norm": 1.5037431716918945,
42
- "learning_rate": 0.0005990626010128218,
43
- "loss": 6.0966,
44
  "step": 250
45
  },
46
  {
47
- "epoch": 0.03228931223764934,
48
- "grad_norm": 2.2753920555114746,
49
- "learning_rate": 0.0005987393599827604,
50
- "loss": 5.953,
51
  "step": 300
52
  },
53
  {
54
- "epoch": 0.03767086427725756,
55
- "grad_norm": 3.0884439945220947,
56
- "learning_rate": 0.000598416118952699,
57
- "loss": 5.869,
58
  "step": 350
59
  },
60
  {
61
- "epoch": 0.04305241631686579,
62
- "grad_norm": 1.4899078607559204,
63
- "learning_rate": 0.0005980928779226376,
64
- "loss": 5.8032,
65
  "step": 400
66
  },
67
  {
68
- "epoch": 0.048433968356474004,
69
- "grad_norm": 1.125209093093872,
70
- "learning_rate": 0.0005977696368925762,
71
- "loss": 5.7355,
72
  "step": 450
73
  },
74
  {
75
- "epoch": 0.05381552039608223,
76
- "grad_norm": 1.960442066192627,
77
- "learning_rate": 0.0005974463958625148,
78
- "loss": 5.6583,
79
  "step": 500
80
  },
81
  {
82
- "epoch": 0.05919707243569045,
83
- "grad_norm": 1.044176697731018,
84
- "learning_rate": 0.0005971231548324534,
85
- "loss": 5.5711,
86
  "step": 550
87
  },
88
  {
89
- "epoch": 0.06457862447529868,
90
- "grad_norm": 1.6373082399368286,
91
- "learning_rate": 0.0005967999138023919,
92
- "loss": 5.5099,
93
  "step": 600
94
  },
95
  {
96
- "epoch": 0.0699601765149069,
97
- "grad_norm": 1.8287220001220703,
98
- "learning_rate": 0.0005964766727723304,
99
- "loss": 5.4242,
100
  "step": 650
101
  },
102
  {
103
- "epoch": 0.07534172855451512,
104
- "grad_norm": 1.3936927318572998,
105
- "learning_rate": 0.0005961534317422691,
106
- "loss": 5.3571,
107
  "step": 700
108
  },
109
  {
110
- "epoch": 0.08072328059412334,
111
- "grad_norm": 0.9875015020370483,
112
- "learning_rate": 0.0005958301907122077,
113
- "loss": 5.3186,
114
  "step": 750
115
  },
116
  {
117
- "epoch": 0.08610483263373157,
118
- "grad_norm": 1.0579822063446045,
119
- "learning_rate": 0.0005955069496821463,
120
- "loss": 5.2725,
121
  "step": 800
122
  },
123
  {
124
- "epoch": 0.09148638467333979,
125
- "grad_norm": 1.3142889738082886,
126
- "learning_rate": 0.0005951837086520848,
127
- "loss": 5.2145,
128
  "step": 850
129
  },
130
  {
131
- "epoch": 0.09686793671294801,
132
- "grad_norm": 1.207129716873169,
133
- "learning_rate": 0.0005948604676220235,
134
- "loss": 5.1809,
135
  "step": 900
136
  },
137
  {
138
- "epoch": 0.10224948875255624,
139
- "grad_norm": 1.3603378534317017,
140
- "learning_rate": 0.000594537226591962,
141
- "loss": 5.1166,
142
  "step": 950
143
  },
144
  {
145
- "epoch": 0.10763104079216446,
146
- "grad_norm": 1.0100822448730469,
147
- "learning_rate": 0.0005942139855619007,
148
- "loss": 5.0872,
149
  "step": 1000
150
  },
151
  {
152
- "epoch": 0.10763104079216446,
153
- "eval_accuracy": 0.22757434275053265,
154
- "eval_loss": 5.0165934562683105,
155
- "eval_runtime": 186.2978,
156
- "eval_samples_per_second": 96.679,
157
- "eval_steps_per_second": 6.044,
158
  "step": 1000
159
  },
160
  {
161
- "epoch": 0.11301259283177269,
162
- "grad_norm": 0.9639580249786377,
163
- "learning_rate": 0.0005938907445318392,
164
- "loss": 5.0357,
165
  "step": 1050
166
  },
167
  {
168
- "epoch": 0.1183941448713809,
169
- "grad_norm": 0.8208778500556946,
170
- "learning_rate": 0.0005935675035017777,
171
- "loss": 5.0074,
172
  "step": 1100
173
  },
174
  {
175
- "epoch": 0.12377569691098914,
176
- "grad_norm": 0.9588396549224854,
177
- "learning_rate": 0.0005932442624717164,
178
- "loss": 4.9928,
179
  "step": 1150
180
  },
181
  {
182
- "epoch": 0.12915724895059735,
183
- "grad_norm": 0.8537307977676392,
184
- "learning_rate": 0.0005929210214416549,
185
- "loss": 4.9625,
186
  "step": 1200
187
  },
188
  {
189
- "epoch": 0.13453880099020557,
190
- "grad_norm": 1.1293931007385254,
191
- "learning_rate": 0.0005925977804115936,
192
- "loss": 4.9296,
193
  "step": 1250
194
  },
195
  {
196
- "epoch": 0.1399203530298138,
197
- "grad_norm": 1.039612054824829,
198
- "learning_rate": 0.0005922745393815321,
199
- "loss": 4.904,
200
  "step": 1300
201
  },
202
  {
203
- "epoch": 0.14530190506942203,
204
- "grad_norm": 0.986505925655365,
205
- "learning_rate": 0.0005919512983514707,
206
- "loss": 4.8739,
207
  "step": 1350
208
  },
209
  {
210
- "epoch": 0.15068345710903025,
211
- "grad_norm": 0.9047093391418457,
212
- "learning_rate": 0.0005916280573214093,
213
- "loss": 4.8174,
214
  "step": 1400
215
  },
216
  {
217
- "epoch": 0.15606500914863847,
218
- "grad_norm": 0.770999014377594,
219
- "learning_rate": 0.0005913048162913478,
220
- "loss": 4.8181,
221
  "step": 1450
222
  },
223
  {
224
- "epoch": 0.16144656118824668,
225
- "grad_norm": 0.9082644581794739,
226
- "learning_rate": 0.0005909815752612864,
227
- "loss": 4.8036,
228
  "step": 1500
229
  },
230
  {
231
- "epoch": 0.1668281132278549,
232
- "grad_norm": 0.8358648419380188,
233
- "learning_rate": 0.000590658334231225,
234
- "loss": 4.7798,
235
  "step": 1550
236
  },
237
  {
238
- "epoch": 0.17220966526746315,
239
- "grad_norm": 1.0487140417099,
240
- "learning_rate": 0.0005903350932011637,
241
- "loss": 4.7364,
242
  "step": 1600
243
  },
244
  {
245
- "epoch": 0.17759121730707136,
246
- "grad_norm": 1.1815866231918335,
247
- "learning_rate": 0.0005900118521711022,
248
- "loss": 4.7392,
249
  "step": 1650
250
  },
251
  {
252
- "epoch": 0.18297276934667958,
253
- "grad_norm": 0.9304591417312622,
254
- "learning_rate": 0.0005896886111410408,
255
- "loss": 4.7118,
256
  "step": 1700
257
  },
258
  {
259
- "epoch": 0.1883543213862878,
260
- "grad_norm": 1.0020699501037598,
261
- "learning_rate": 0.0005893653701109793,
262
- "loss": 4.6978,
263
  "step": 1750
264
  },
265
  {
266
- "epoch": 0.19373587342589602,
267
- "grad_norm": 0.8822704553604126,
268
- "learning_rate": 0.000589042129080918,
269
- "loss": 4.6769,
270
  "step": 1800
271
  },
272
  {
273
- "epoch": 0.19911742546550426,
274
- "grad_norm": 0.944671630859375,
275
- "learning_rate": 0.0005887188880508566,
276
- "loss": 4.6617,
277
  "step": 1850
278
  },
279
  {
280
- "epoch": 0.20449897750511248,
281
- "grad_norm": 1.0103189945220947,
282
- "learning_rate": 0.0005883956470207951,
283
- "loss": 4.6347,
284
  "step": 1900
285
  },
286
  {
287
- "epoch": 0.2098805295447207,
288
- "grad_norm": 0.9612853527069092,
289
- "learning_rate": 0.0005880724059907337,
290
- "loss": 4.5904,
291
  "step": 1950
292
  },
293
  {
294
- "epoch": 0.2152620815843289,
295
- "grad_norm": 0.9868743419647217,
296
- "learning_rate": 0.0005877491649606723,
297
- "loss": 4.58,
298
  "step": 2000
299
  },
300
  {
301
- "epoch": 0.2152620815843289,
302
- "eval_accuracy": 0.27043877621644785,
303
- "eval_loss": 4.506998538970947,
304
- "eval_runtime": 184.7948,
305
- "eval_samples_per_second": 97.465,
306
- "eval_steps_per_second": 6.093,
307
  "step": 2000
308
  },
309
  {
310
- "epoch": 0.22064363362393713,
311
- "grad_norm": 0.8025125861167908,
312
- "learning_rate": 0.0005874259239306109,
313
- "loss": 4.5503,
314
  "step": 2050
315
  },
316
  {
317
- "epoch": 0.22602518566354537,
318
- "grad_norm": 1.0225203037261963,
319
- "learning_rate": 0.0005871026829005494,
320
- "loss": 4.5634,
321
  "step": 2100
322
  },
323
  {
324
- "epoch": 0.2314067377031536,
325
- "grad_norm": 0.9046829342842102,
326
- "learning_rate": 0.0005867794418704881,
327
- "loss": 4.5213,
328
  "step": 2150
329
  },
330
  {
331
- "epoch": 0.2367882897427618,
332
- "grad_norm": 0.7576098442077637,
333
- "learning_rate": 0.0005864562008404266,
334
- "loss": 4.5007,
335
  "step": 2200
336
  },
337
  {
338
- "epoch": 0.24216984178237003,
339
- "grad_norm": 0.9998608231544495,
340
- "learning_rate": 0.0005861329598103653,
341
- "loss": 4.4993,
342
  "step": 2250
343
  },
344
  {
345
- "epoch": 0.24755139382197827,
346
- "grad_norm": 1.168628215789795,
347
- "learning_rate": 0.0005858097187803038,
348
- "loss": 4.4896,
349
  "step": 2300
350
  },
351
  {
352
- "epoch": 0.2529329458615865,
353
- "grad_norm": 0.8030685782432556,
354
- "learning_rate": 0.0005854864777502423,
355
- "loss": 4.4601,
356
  "step": 2350
357
  },
358
  {
359
- "epoch": 0.2583144979011947,
360
- "grad_norm": 1.0477194786071777,
361
- "learning_rate": 0.000585163236720181,
362
- "loss": 4.4508,
363
  "step": 2400
364
  },
365
  {
366
- "epoch": 0.2636960499408029,
367
- "grad_norm": 0.86094731092453,
368
- "learning_rate": 0.0005848399956901195,
369
- "loss": 4.428,
370
  "step": 2450
371
  },
372
  {
373
- "epoch": 0.26907760198041114,
374
- "grad_norm": 0.9053926467895508,
375
- "learning_rate": 0.0005845167546600582,
376
- "loss": 4.4063,
377
  "step": 2500
378
  },
379
  {
380
- "epoch": 0.27445915402001936,
381
- "grad_norm": 0.7576990127563477,
382
- "learning_rate": 0.0005841935136299967,
383
- "loss": 4.4131,
384
  "step": 2550
385
  },
386
  {
387
- "epoch": 0.2798407060596276,
388
- "grad_norm": 0.7213003635406494,
389
- "learning_rate": 0.0005838702725999353,
390
- "loss": 4.3988,
391
  "step": 2600
392
  },
393
  {
394
- "epoch": 0.2852222580992358,
395
- "grad_norm": 0.9967924356460571,
396
- "learning_rate": 0.0005835470315698739,
397
- "loss": 4.3945,
398
  "step": 2650
399
  },
400
  {
401
- "epoch": 0.29060381013884407,
402
- "grad_norm": 0.7419942617416382,
403
- "learning_rate": 0.0005832237905398125,
404
- "loss": 4.3636,
405
  "step": 2700
406
  },
407
  {
408
- "epoch": 0.2959853621784523,
409
- "grad_norm": 0.7668341994285583,
410
- "learning_rate": 0.0005829005495097511,
411
- "loss": 4.3658,
412
  "step": 2750
413
  },
414
  {
415
- "epoch": 0.3013669142180605,
416
- "grad_norm": 0.7742299437522888,
417
- "learning_rate": 0.0005825773084796896,
418
- "loss": 4.334,
419
  "step": 2800
420
  },
421
  {
422
- "epoch": 0.3067484662576687,
423
- "grad_norm": 0.9014106392860413,
424
- "learning_rate": 0.0005822540674496282,
425
- "loss": 4.327,
426
  "step": 2850
427
  },
428
  {
429
- "epoch": 0.31213001829727693,
430
- "grad_norm": 0.6771566867828369,
431
- "learning_rate": 0.0005819308264195668,
432
- "loss": 4.3486,
433
  "step": 2900
434
  },
435
  {
436
- "epoch": 0.31751157033688515,
437
- "grad_norm": 0.712129533290863,
438
- "learning_rate": 0.0005816075853895054,
439
- "loss": 4.3111,
440
  "step": 2950
441
  },
442
  {
443
- "epoch": 0.32289312237649337,
444
- "grad_norm": 0.6686953902244568,
445
- "learning_rate": 0.000581284344359444,
446
- "loss": 4.3079,
447
  "step": 3000
448
  },
449
  {
450
- "epoch": 0.32289312237649337,
451
- "eval_accuracy": 0.2981868766651734,
452
- "eval_loss": 4.241795063018799,
453
- "eval_runtime": 184.6997,
454
- "eval_samples_per_second": 97.515,
455
- "eval_steps_per_second": 6.096,
456
  "step": 3000
457
  },
458
  {
459
- "epoch": 0.3282746744161016,
460
- "grad_norm": 0.6886458992958069,
461
- "learning_rate": 0.0005809611033293825,
462
- "loss": 4.3161,
463
  "step": 3050
464
  },
465
  {
466
- "epoch": 0.3336562264557098,
467
- "grad_norm": 0.7012606859207153,
468
- "learning_rate": 0.0005806378622993211,
469
- "loss": 4.2908,
470
  "step": 3100
471
  },
472
  {
473
- "epoch": 0.3390377784953181,
474
- "grad_norm": 0.6854538321495056,
475
- "learning_rate": 0.0005803146212692597,
476
- "loss": 4.2775,
477
  "step": 3150
478
  },
479
  {
480
- "epoch": 0.3444193305349263,
481
- "grad_norm": 0.8339408040046692,
482
- "learning_rate": 0.0005799913802391983,
483
- "loss": 4.27,
484
  "step": 3200
485
  },
486
  {
487
- "epoch": 0.3498008825745345,
488
- "grad_norm": 0.8958104848861694,
489
- "learning_rate": 0.0005796681392091369,
490
- "loss": 4.2535,
491
  "step": 3250
492
  },
493
  {
494
- "epoch": 0.35518243461414273,
495
- "grad_norm": 0.671660304069519,
496
- "learning_rate": 0.0005793448981790755,
497
- "loss": 4.2516,
498
  "step": 3300
499
  },
500
  {
501
- "epoch": 0.36056398665375095,
502
- "grad_norm": 0.7408727407455444,
503
- "learning_rate": 0.0005790216571490141,
504
- "loss": 4.2658,
505
  "step": 3350
506
  },
507
  {
508
- "epoch": 0.36594553869335916,
509
- "grad_norm": 0.6920570731163025,
510
- "learning_rate": 0.0005786984161189527,
511
- "loss": 4.2325,
512
  "step": 3400
513
  },
514
  {
515
- "epoch": 0.3713270907329674,
516
- "grad_norm": 0.8909915685653687,
517
- "learning_rate": 0.0005783751750888912,
518
- "loss": 4.2163,
519
  "step": 3450
520
  },
521
  {
522
- "epoch": 0.3767086427725756,
523
- "grad_norm": 0.8377681970596313,
524
- "learning_rate": 0.0005780519340588297,
525
- "loss": 4.2289,
526
  "step": 3500
527
  },
528
  {
529
- "epoch": 0.3820901948121838,
530
- "grad_norm": 0.6133842468261719,
531
- "learning_rate": 0.0005777286930287684,
532
- "loss": 4.216,
533
  "step": 3550
534
  },
535
  {
536
- "epoch": 0.38747174685179203,
537
- "grad_norm": 0.5939428806304932,
538
- "learning_rate": 0.000577405451998707,
539
- "loss": 4.2227,
540
  "step": 3600
541
  },
542
  {
543
- "epoch": 0.3928532988914003,
544
- "grad_norm": 0.6408166885375977,
545
- "learning_rate": 0.0005770822109686456,
546
- "loss": 4.2049,
547
  "step": 3650
548
  },
549
  {
550
- "epoch": 0.3982348509310085,
551
- "grad_norm": 0.819424033164978,
552
- "learning_rate": 0.0005767589699385841,
553
- "loss": 4.2005,
554
  "step": 3700
555
  },
556
  {
557
- "epoch": 0.40361640297061674,
558
- "grad_norm": 0.6583958864212036,
559
- "learning_rate": 0.0005764357289085228,
560
- "loss": 4.1926,
561
  "step": 3750
562
  },
563
  {
564
- "epoch": 0.40899795501022496,
565
- "grad_norm": 0.6472761631011963,
566
- "learning_rate": 0.0005761124878784613,
567
- "loss": 4.1791,
568
  "step": 3800
569
  },
570
  {
571
- "epoch": 0.4143795070498332,
572
- "grad_norm": 0.6922594308853149,
573
- "learning_rate": 0.0005757892468484,
574
- "loss": 4.1849,
575
  "step": 3850
576
  },
577
  {
578
- "epoch": 0.4197610590894414,
579
- "grad_norm": 0.6930834054946899,
580
- "learning_rate": 0.0005754660058183385,
581
- "loss": 4.1757,
582
  "step": 3900
583
  },
584
  {
585
- "epoch": 0.4251426111290496,
586
- "grad_norm": 0.6437042355537415,
587
- "learning_rate": 0.000575142764788277,
588
- "loss": 4.1714,
589
  "step": 3950
590
  },
591
  {
592
- "epoch": 0.4305241631686578,
593
- "grad_norm": 0.7084268927574158,
594
- "learning_rate": 0.0005748195237582157,
595
- "loss": 4.1579,
596
  "step": 4000
597
  },
598
  {
599
- "epoch": 0.4305241631686578,
600
- "eval_accuracy": 0.31272235134410686,
601
- "eval_loss": 4.0872697830200195,
602
- "eval_runtime": 184.7092,
603
- "eval_samples_per_second": 97.51,
604
- "eval_steps_per_second": 6.096,
605
  "step": 4000
606
  },
607
  {
608
- "epoch": 0.43590571520826604,
609
- "grad_norm": 0.6244553327560425,
610
- "learning_rate": 0.0005744962827281542,
611
- "loss": 4.1348,
612
  "step": 4050
613
  },
614
  {
615
- "epoch": 0.44128726724787426,
616
- "grad_norm": 0.7042115330696106,
617
- "learning_rate": 0.0005741730416980928,
618
- "loss": 4.1397,
619
  "step": 4100
620
  },
621
  {
622
- "epoch": 0.44666881928748253,
623
- "grad_norm": 0.7811130881309509,
624
- "learning_rate": 0.0005738498006680314,
625
- "loss": 4.1573,
626
  "step": 4150
627
  },
628
  {
629
- "epoch": 0.45205037132709075,
630
- "grad_norm": 0.6470036506652832,
631
- "learning_rate": 0.00057352655963797,
632
- "loss": 4.1357,
633
  "step": 4200
634
  },
635
  {
636
- "epoch": 0.45743192336669897,
637
- "grad_norm": 0.6716099977493286,
638
- "learning_rate": 0.0005732033186079086,
639
- "loss": 4.1327,
640
  "step": 4250
641
  },
642
  {
643
- "epoch": 0.4628134754063072,
644
- "grad_norm": 0.7920870184898376,
645
- "learning_rate": 0.0005728800775778471,
646
- "loss": 4.1263,
647
  "step": 4300
648
  },
649
  {
650
- "epoch": 0.4681950274459154,
651
- "grad_norm": 0.7230188846588135,
652
- "learning_rate": 0.0005725568365477857,
653
- "loss": 4.1104,
654
  "step": 4350
655
  },
656
  {
657
- "epoch": 0.4735765794855236,
658
- "grad_norm": 0.7376595735549927,
659
- "learning_rate": 0.0005722335955177243,
660
- "loss": 4.1103,
661
  "step": 4400
662
  },
663
  {
664
- "epoch": 0.47895813152513184,
665
- "grad_norm": 1.1194459199905396,
666
- "learning_rate": 0.000571910354487663,
667
- "loss": 4.1222,
668
  "step": 4450
669
  },
670
  {
671
- "epoch": 0.48433968356474005,
672
- "grad_norm": 0.6437053680419922,
673
- "learning_rate": 0.0005715871134576015,
674
- "loss": 4.1131,
675
  "step": 4500
676
  },
677
  {
678
- "epoch": 0.48972123560434827,
679
- "grad_norm": 0.7767078876495361,
680
- "learning_rate": 0.0005712638724275401,
681
- "loss": 4.1017,
682
  "step": 4550
683
  },
684
  {
685
- "epoch": 0.49510278764395654,
686
- "grad_norm": 0.6522454023361206,
687
- "learning_rate": 0.0005709406313974786,
688
- "loss": 4.0979,
689
  "step": 4600
690
  },
691
  {
692
- "epoch": 0.5004843396835648,
693
- "grad_norm": 0.6566357612609863,
694
- "learning_rate": 0.0005706173903674173,
695
- "loss": 4.0735,
696
  "step": 4650
697
  },
698
  {
699
- "epoch": 0.505865891723173,
700
- "grad_norm": 0.5966187715530396,
701
- "learning_rate": 0.0005702941493373559,
702
- "loss": 4.0895,
703
  "step": 4700
704
  },
705
  {
706
- "epoch": 0.5112474437627812,
707
- "grad_norm": 0.7359771132469177,
708
- "learning_rate": 0.0005699709083072944,
709
- "loss": 4.0868,
710
  "step": 4750
711
  },
712
  {
713
- "epoch": 0.5166289958023894,
714
- "grad_norm": 0.7156068086624146,
715
- "learning_rate": 0.000569647667277233,
716
- "loss": 4.0897,
717
  "step": 4800
718
  },
719
  {
720
- "epoch": 0.5220105478419976,
721
- "grad_norm": 0.6234465837478638,
722
- "learning_rate": 0.0005693244262471716,
723
- "loss": 4.0549,
724
  "step": 4850
725
  },
726
  {
727
- "epoch": 0.5273920998816058,
728
- "grad_norm": 0.7085710167884827,
729
- "learning_rate": 0.0005690011852171102,
730
- "loss": 4.0802,
731
  "step": 4900
732
  },
733
  {
734
- "epoch": 0.5327736519212141,
735
- "grad_norm": 0.5667790174484253,
736
- "learning_rate": 0.0005686779441870487,
737
- "loss": 4.0628,
738
  "step": 4950
739
  },
740
  {
741
- "epoch": 0.5381552039608223,
742
- "grad_norm": 0.6275297999382019,
743
- "learning_rate": 0.0005683547031569874,
744
- "loss": 4.0589,
745
  "step": 5000
746
  },
747
  {
748
- "epoch": 0.5381552039608223,
749
- "eval_accuracy": 0.3213402637940002,
750
- "eval_loss": 3.990799903869629,
751
- "eval_runtime": 184.6417,
752
- "eval_samples_per_second": 97.546,
753
- "eval_steps_per_second": 6.098,
754
  "step": 5000
755
  },
756
  {
757
- "epoch": 0.5435367560004305,
758
- "grad_norm": 0.6828199625015259,
759
- "learning_rate": 0.0005680314621269259,
760
- "loss": 4.056,
761
  "step": 5050
762
  },
763
  {
764
- "epoch": 0.5489183080400387,
765
- "grad_norm": 0.6843883991241455,
766
- "learning_rate": 0.0005677082210968646,
767
- "loss": 4.0418,
768
  "step": 5100
769
  },
770
  {
771
- "epoch": 0.5542998600796469,
772
- "grad_norm": 0.5970234274864197,
773
- "learning_rate": 0.0005673849800668031,
774
- "loss": 4.0339,
775
  "step": 5150
776
  },
777
  {
778
- "epoch": 0.5596814121192552,
779
- "grad_norm": 0.6066529154777527,
780
- "learning_rate": 0.0005670617390367416,
781
- "loss": 4.0554,
782
  "step": 5200
783
  },
784
  {
785
- "epoch": 0.5650629641588634,
786
- "grad_norm": 0.7181238532066345,
787
- "learning_rate": 0.0005667384980066803,
788
- "loss": 4.0289,
789
  "step": 5250
790
  },
791
  {
792
- "epoch": 0.5704445161984716,
793
- "grad_norm": 0.6789528131484985,
794
- "learning_rate": 0.0005664152569766188,
795
- "loss": 4.0306,
796
  "step": 5300
797
  },
798
  {
799
- "epoch": 0.5758260682380799,
800
- "grad_norm": 0.6828908920288086,
801
- "learning_rate": 0.0005660920159465575,
802
- "loss": 4.0383,
803
  "step": 5350
804
  },
805
  {
806
- "epoch": 0.5812076202776881,
807
- "grad_norm": 0.5343970656394958,
808
- "learning_rate": 0.000565768774916496,
809
- "loss": 4.0472,
810
  "step": 5400
811
  },
812
  {
813
- "epoch": 0.5865891723172963,
814
- "grad_norm": 0.6703925728797913,
815
- "learning_rate": 0.0005654455338864346,
816
- "loss": 4.0139,
817
  "step": 5450
818
  },
819
  {
820
- "epoch": 0.5919707243569046,
821
- "grad_norm": 0.5589005351066589,
822
- "learning_rate": 0.0005651222928563732,
823
- "loss": 4.0171,
824
  "step": 5500
825
  },
826
  {
827
- "epoch": 0.5973522763965128,
828
- "grad_norm": 0.6419985890388489,
829
- "learning_rate": 0.0005647990518263118,
830
- "loss": 3.994,
831
  "step": 5550
832
  },
833
  {
834
- "epoch": 0.602733828436121,
835
- "grad_norm": 0.5866629481315613,
836
- "learning_rate": 0.0005644758107962504,
837
- "loss": 4.0104,
838
  "step": 5600
839
  },
840
  {
841
- "epoch": 0.6081153804757292,
842
- "grad_norm": 0.9215855002403259,
843
- "learning_rate": 0.0005641525697661889,
844
- "loss": 4.0096,
845
  "step": 5650
846
  },
847
  {
848
- "epoch": 0.6134969325153374,
849
- "grad_norm": 0.5779905319213867,
850
- "learning_rate": 0.0005638293287361275,
851
- "loss": 3.9754,
852
  "step": 5700
853
  },
854
  {
855
- "epoch": 0.6188784845549457,
856
- "grad_norm": 0.6425080299377441,
857
- "learning_rate": 0.0005635060877060661,
858
- "loss": 3.9917,
859
  "step": 5750
860
  },
861
  {
862
- "epoch": 0.6242600365945539,
863
- "grad_norm": 0.6887379884719849,
864
- "learning_rate": 0.0005631828466760047,
865
- "loss": 3.9941,
866
  "step": 5800
867
  },
868
  {
869
- "epoch": 0.6296415886341621,
870
- "grad_norm": 0.6834486722946167,
871
- "learning_rate": 0.0005628596056459433,
872
- "loss": 3.9953,
873
  "step": 5850
874
  },
875
  {
876
- "epoch": 0.6350231406737703,
877
- "grad_norm": 0.7504608631134033,
878
- "learning_rate": 0.0005625363646158818,
879
  "loss": 3.9892,
880
  "step": 5900
881
  },
882
  {
883
- "epoch": 0.6404046927133785,
884
- "grad_norm": 0.6510689854621887,
885
- "learning_rate": 0.0005622131235858205,
886
- "loss": 3.9722,
887
  "step": 5950
888
  },
889
  {
890
- "epoch": 0.6457862447529867,
891
- "grad_norm": 0.5909599661827087,
892
- "learning_rate": 0.000561889882555759,
893
- "loss": 3.9673,
894
  "step": 6000
895
  },
896
  {
897
- "epoch": 0.6457862447529867,
898
- "eval_accuracy": 0.32834229049631664,
899
- "eval_loss": 3.9165666103363037,
900
- "eval_runtime": 184.5514,
901
- "eval_samples_per_second": 97.593,
902
- "eval_steps_per_second": 6.101,
903
  "step": 6000
904
  },
905
  {
906
- "epoch": 0.651167796792595,
907
- "grad_norm": 0.6396195292472839,
908
- "learning_rate": 0.0005615666415256976,
909
- "loss": 3.997,
910
  "step": 6050
911
  },
912
  {
913
- "epoch": 0.6565493488322032,
914
- "grad_norm": 0.5541490316390991,
915
- "learning_rate": 0.0005612434004956361,
916
- "loss": 3.981,
917
  "step": 6100
918
  },
919
  {
920
- "epoch": 0.6619309008718114,
921
- "grad_norm": 0.6580690145492554,
922
- "learning_rate": 0.0005609201594655748,
923
- "loss": 3.9881,
924
  "step": 6150
925
  },
926
  {
927
- "epoch": 0.6673124529114196,
928
- "grad_norm": 0.6880823373794556,
929
- "learning_rate": 0.0005605969184355134,
930
- "loss": 3.9573,
931
  "step": 6200
932
  },
933
  {
934
- "epoch": 0.6726940049510278,
935
- "grad_norm": 0.6356266140937805,
936
- "learning_rate": 0.000560273677405452,
937
- "loss": 3.9975,
938
  "step": 6250
939
  },
940
  {
941
- "epoch": 0.6780755569906362,
942
- "grad_norm": 0.5227646827697754,
943
- "learning_rate": 0.0005599504363753905,
944
- "loss": 3.9533,
945
  "step": 6300
946
  },
947
  {
948
- "epoch": 0.6834571090302444,
949
- "grad_norm": 0.6362395882606506,
950
- "learning_rate": 0.0005596271953453291,
951
- "loss": 3.9563,
952
  "step": 6350
953
  },
954
  {
955
- "epoch": 0.6888386610698526,
956
- "grad_norm": 0.5801821947097778,
957
- "learning_rate": 0.0005593039543152677,
958
- "loss": 3.9428,
959
  "step": 6400
960
  },
961
  {
962
- "epoch": 0.6942202131094608,
963
- "grad_norm": 0.6077560782432556,
964
- "learning_rate": 0.0005589807132852063,
965
- "loss": 3.9554,
966
  "step": 6450
967
  },
968
  {
969
- "epoch": 0.699601765149069,
970
- "grad_norm": 0.5769425630569458,
971
- "learning_rate": 0.0005586574722551449,
972
- "loss": 3.9447,
973
  "step": 6500
974
  },
975
  {
976
- "epoch": 0.7049833171886772,
977
- "grad_norm": 0.6278015971183777,
978
- "learning_rate": 0.0005583342312250834,
979
- "loss": 3.9678,
980
  "step": 6550
981
  },
982
  {
983
- "epoch": 0.7103648692282855,
984
- "grad_norm": 0.5199827551841736,
985
- "learning_rate": 0.0005580109901950221,
986
- "loss": 3.9618,
987
  "step": 6600
988
  },
989
  {
990
- "epoch": 0.7157464212678937,
991
- "grad_norm": 0.659668505191803,
992
- "learning_rate": 0.0005576877491649606,
993
- "loss": 3.9206,
994
  "step": 6650
995
  },
996
  {
997
- "epoch": 0.7211279733075019,
998
- "grad_norm": 0.6791085004806519,
999
- "learning_rate": 0.0005573645081348993,
1000
  "loss": 3.9427,
1001
  "step": 6700
1002
  },
1003
  {
1004
- "epoch": 0.7265095253471101,
1005
- "grad_norm": 0.6757652759552002,
1006
- "learning_rate": 0.0005570412671048378,
1007
- "loss": 3.9381,
1008
  "step": 6750
1009
  },
1010
  {
1011
- "epoch": 0.7318910773867183,
1012
- "grad_norm": 0.5604714155197144,
1013
- "learning_rate": 0.0005567180260747763,
1014
- "loss": 3.9385,
1015
  "step": 6800
1016
  },
1017
  {
1018
- "epoch": 0.7372726294263265,
1019
- "grad_norm": 0.6604406237602234,
1020
- "learning_rate": 0.000556394785044715,
1021
- "loss": 3.9272,
1022
  "step": 6850
1023
  },
1024
  {
1025
- "epoch": 0.7426541814659348,
1026
- "grad_norm": 0.564445972442627,
1027
- "learning_rate": 0.0005560715440146535,
1028
- "loss": 3.9496,
1029
  "step": 6900
1030
  },
1031
  {
1032
- "epoch": 0.748035733505543,
1033
- "grad_norm": 0.5737416744232178,
1034
- "learning_rate": 0.0005557483029845921,
1035
- "loss": 3.9387,
1036
  "step": 6950
1037
  },
1038
  {
1039
- "epoch": 0.7534172855451512,
1040
- "grad_norm": 0.5588909983634949,
1041
- "learning_rate": 0.0005554250619545307,
1042
- "loss": 3.9074,
1043
  "step": 7000
1044
  },
1045
  {
1046
- "epoch": 0.7534172855451512,
1047
- "eval_accuracy": 0.33360815270424543,
1048
- "eval_loss": 3.859344720840454,
1049
- "eval_runtime": 184.9069,
1050
- "eval_samples_per_second": 97.406,
1051
- "eval_steps_per_second": 6.09,
1052
  "step": 7000
1053
  },
1054
  {
1055
- "epoch": 0.7587988375847594,
1056
- "grad_norm": 0.6285735964775085,
1057
- "learning_rate": 0.0005551018209244694,
1058
- "loss": 3.9333,
1059
  "step": 7050
1060
  },
1061
  {
1062
- "epoch": 0.7641803896243676,
1063
- "grad_norm": 0.5597401261329651,
1064
- "learning_rate": 0.0005547785798944079,
1065
- "loss": 3.9209,
1066
  "step": 7100
1067
  },
1068
  {
1069
- "epoch": 0.7695619416639758,
1070
- "grad_norm": 0.578063428401947,
1071
- "learning_rate": 0.0005544553388643464,
1072
- "loss": 3.9312,
1073
  "step": 7150
1074
  },
1075
  {
1076
- "epoch": 0.7749434937035841,
1077
- "grad_norm": 0.5833026170730591,
1078
- "learning_rate": 0.000554132097834285,
1079
- "loss": 3.9328,
1080
  "step": 7200
1081
  },
1082
  {
1083
- "epoch": 0.7803250457431924,
1084
- "grad_norm": 0.518353283405304,
1085
- "learning_rate": 0.0005538088568042236,
1086
- "loss": 3.9029,
1087
  "step": 7250
1088
  },
1089
  {
1090
- "epoch": 0.7857065977828006,
1091
- "grad_norm": 0.6059916019439697,
1092
- "learning_rate": 0.0005534856157741623,
1093
- "loss": 3.9199,
1094
  "step": 7300
1095
  },
1096
  {
1097
- "epoch": 0.7910881498224088,
1098
- "grad_norm": 0.5463325381278992,
1099
- "learning_rate": 0.0005531623747441008,
1100
- "loss": 3.9054,
1101
  "step": 7350
1102
  },
1103
  {
1104
- "epoch": 0.796469701862017,
1105
- "grad_norm": 0.6463024616241455,
1106
- "learning_rate": 0.0005528391337140394,
1107
- "loss": 3.9191,
1108
  "step": 7400
1109
  },
1110
  {
1111
- "epoch": 0.8018512539016253,
1112
- "grad_norm": 0.6024007797241211,
1113
- "learning_rate": 0.0005525158926839779,
1114
- "loss": 3.9284,
1115
  "step": 7450
1116
  },
1117
  {
1118
- "epoch": 0.8072328059412335,
1119
- "grad_norm": 0.5554441213607788,
1120
- "learning_rate": 0.0005521926516539166,
1121
- "loss": 3.9015,
1122
  "step": 7500
1123
  },
1124
  {
1125
- "epoch": 0.8126143579808417,
1126
- "grad_norm": 0.5887565612792969,
1127
- "learning_rate": 0.0005518694106238552,
1128
- "loss": 3.9024,
1129
  "step": 7550
1130
  },
1131
  {
1132
- "epoch": 0.8179959100204499,
1133
- "grad_norm": 0.575228214263916,
1134
- "learning_rate": 0.0005515461695937937,
1135
- "loss": 3.8981,
1136
  "step": 7600
1137
  },
1138
  {
1139
- "epoch": 0.8233774620600581,
1140
- "grad_norm": 0.6179103255271912,
1141
- "learning_rate": 0.0005512229285637323,
1142
- "loss": 3.9081,
1143
  "step": 7650
1144
  },
1145
  {
1146
- "epoch": 0.8287590140996663,
1147
- "grad_norm": 0.6597625017166138,
1148
- "learning_rate": 0.0005508996875336709,
1149
- "loss": 3.8958,
1150
  "step": 7700
1151
  },
1152
  {
1153
- "epoch": 0.8341405661392746,
1154
- "grad_norm": 0.5965509414672852,
1155
- "learning_rate": 0.0005505764465036095,
1156
- "loss": 3.8879,
1157
  "step": 7750
1158
  },
1159
  {
1160
- "epoch": 0.8395221181788828,
1161
- "grad_norm": 0.512361466884613,
1162
- "learning_rate": 0.000550253205473548,
1163
- "loss": 3.8809,
1164
  "step": 7800
1165
  },
1166
  {
1167
- "epoch": 0.844903670218491,
1168
- "grad_norm": 0.6104485988616943,
1169
- "learning_rate": 0.0005499299644434867,
1170
- "loss": 3.8852,
1171
  "step": 7850
1172
  },
1173
  {
1174
- "epoch": 0.8502852222580992,
1175
- "grad_norm": 0.6456380486488342,
1176
- "learning_rate": 0.0005496067234134252,
1177
- "loss": 3.8783,
1178
  "step": 7900
1179
  },
1180
  {
1181
- "epoch": 0.8556667742977074,
1182
- "grad_norm": 0.5954079627990723,
1183
- "learning_rate": 0.0005492834823833639,
1184
- "loss": 3.889,
1185
  "step": 7950
1186
  },
1187
  {
1188
- "epoch": 0.8610483263373157,
1189
- "grad_norm": 0.6212036609649658,
1190
- "learning_rate": 0.0005489602413533024,
1191
  "loss": 3.8712,
1192
  "step": 8000
1193
  },
1194
  {
1195
- "epoch": 0.8610483263373157,
1196
- "eval_accuracy": 0.33805890094779,
1197
- "eval_loss": 3.8169474601745605,
1198
- "eval_runtime": 184.4499,
1199
- "eval_samples_per_second": 97.647,
1200
- "eval_steps_per_second": 6.105,
1201
  "step": 8000
1202
  },
1203
  {
1204
- "epoch": 0.8664298783769239,
1205
- "grad_norm": 0.5757076144218445,
1206
- "learning_rate": 0.0005486370003232409,
1207
- "loss": 3.8725,
1208
  "step": 8050
1209
  },
1210
  {
1211
- "epoch": 0.8718114304165321,
1212
- "grad_norm": 0.6085401177406311,
1213
- "learning_rate": 0.0005483137592931796,
1214
- "loss": 3.8905,
1215
  "step": 8100
1216
  },
1217
  {
1218
- "epoch": 0.8771929824561403,
1219
- "grad_norm": 0.5095037817955017,
1220
- "learning_rate": 0.0005479905182631181,
1221
- "loss": 3.8799,
1222
  "step": 8150
1223
  },
1224
  {
1225
- "epoch": 0.8825745344957485,
1226
- "grad_norm": 0.5414556860923767,
1227
- "learning_rate": 0.0005476672772330568,
1228
- "loss": 3.8747,
1229
  "step": 8200
1230
  },
1231
  {
1232
- "epoch": 0.8879560865353568,
1233
- "grad_norm": 0.5936631560325623,
1234
- "learning_rate": 0.0005473440362029953,
1235
- "loss": 3.863,
1236
  "step": 8250
1237
  },
1238
  {
1239
- "epoch": 0.8933376385749651,
1240
- "grad_norm": 0.5667993426322937,
1241
- "learning_rate": 0.0005470207951729339,
1242
- "loss": 3.8607,
1243
  "step": 8300
1244
  },
1245
  {
1246
- "epoch": 0.8987191906145733,
1247
- "grad_norm": 0.5451300144195557,
1248
- "learning_rate": 0.0005466975541428725,
1249
- "loss": 3.8857,
1250
  "step": 8350
1251
  },
1252
  {
1253
- "epoch": 0.9041007426541815,
1254
- "grad_norm": 0.5745592713356018,
1255
- "learning_rate": 0.0005463743131128111,
1256
- "loss": 3.8554,
1257
  "step": 8400
1258
  },
1259
  {
1260
- "epoch": 0.9094822946937897,
1261
- "grad_norm": 0.5426753759384155,
1262
- "learning_rate": 0.0005460510720827497,
1263
- "loss": 3.8752,
1264
  "step": 8450
1265
  },
1266
  {
1267
- "epoch": 0.9148638467333979,
1268
- "grad_norm": 0.508082389831543,
1269
- "learning_rate": 0.0005457278310526882,
1270
- "loss": 3.8584,
1271
  "step": 8500
1272
  },
1273
  {
1274
- "epoch": 0.9202453987730062,
1275
- "grad_norm": 0.532146155834198,
1276
- "learning_rate": 0.0005454045900226268,
1277
- "loss": 3.8476,
1278
  "step": 8550
1279
  },
1280
  {
1281
- "epoch": 0.9256269508126144,
1282
- "grad_norm": 0.5778146386146545,
1283
- "learning_rate": 0.0005450813489925654,
1284
- "loss": 3.8612,
1285
  "step": 8600
1286
  },
1287
  {
1288
- "epoch": 0.9310085028522226,
1289
- "grad_norm": 0.5984019637107849,
1290
- "learning_rate": 0.000544758107962504,
1291
- "loss": 3.8478,
1292
  "step": 8650
1293
  },
1294
  {
1295
- "epoch": 0.9363900548918308,
1296
- "grad_norm": 0.609585165977478,
1297
- "learning_rate": 0.0005444348669324426,
1298
- "loss": 3.8541,
1299
  "step": 8700
1300
  },
1301
  {
1302
- "epoch": 0.941771606931439,
1303
- "grad_norm": 0.5877447724342346,
1304
- "learning_rate": 0.0005441116259023811,
1305
- "loss": 3.8298,
1306
  "step": 8750
1307
  },
1308
  {
1309
- "epoch": 0.9471531589710472,
1310
- "grad_norm": 0.6004450917243958,
1311
- "learning_rate": 0.0005437883848723198,
1312
- "loss": 3.8393,
1313
  "step": 8800
1314
  },
1315
  {
1316
- "epoch": 0.9525347110106555,
1317
- "grad_norm": 0.5448794960975647,
1318
- "learning_rate": 0.0005434651438422583,
1319
- "loss": 3.8515,
1320
  "step": 8850
1321
  },
1322
  {
1323
- "epoch": 0.9579162630502637,
1324
- "grad_norm": 0.5065131783485413,
1325
- "learning_rate": 0.0005431419028121969,
1326
- "loss": 3.8478,
1327
  "step": 8900
1328
  },
1329
  {
1330
- "epoch": 0.9632978150898719,
1331
- "grad_norm": 0.6335159540176392,
1332
- "learning_rate": 0.0005428186617821354,
1333
- "loss": 3.8411,
1334
  "step": 8950
1335
  },
1336
  {
1337
- "epoch": 0.9686793671294801,
1338
- "grad_norm": 0.5555748343467712,
1339
- "learning_rate": 0.0005424954207520741,
1340
- "loss": 3.8516,
1341
  "step": 9000
1342
  },
1343
  {
1344
- "epoch": 0.9686793671294801,
1345
- "eval_accuracy": 0.34156469502601206,
1346
- "eval_loss": 3.775557279586792,
1347
- "eval_runtime": 184.197,
1348
- "eval_samples_per_second": 97.781,
1349
- "eval_steps_per_second": 6.113,
1350
  "step": 9000
1351
  },
1352
  {
1353
- "epoch": 0.9740609191690883,
1354
- "grad_norm": 0.6176546812057495,
1355
- "learning_rate": 0.0005421721797220127,
1356
- "loss": 3.8444,
1357
  "step": 9050
1358
  },
1359
  {
1360
- "epoch": 0.9794424712086965,
1361
- "grad_norm": 0.5457547307014465,
1362
- "learning_rate": 0.0005418489386919513,
1363
- "loss": 3.8305,
1364
  "step": 9100
1365
  },
1366
  {
1367
- "epoch": 0.9848240232483048,
1368
- "grad_norm": 0.5831869840621948,
1369
- "learning_rate": 0.0005415256976618898,
1370
- "loss": 3.8196,
1371
  "step": 9150
1372
  },
1373
  {
1374
- "epoch": 0.9902055752879131,
1375
- "grad_norm": 0.5737512111663818,
1376
- "learning_rate": 0.0005412024566318284,
1377
- "loss": 3.8124,
1378
  "step": 9200
1379
  },
1380
  {
1381
- "epoch": 0.9955871273275213,
1382
- "grad_norm": 0.5983269810676575,
1383
- "learning_rate": 0.000540879215601767,
1384
- "loss": 3.8368,
1385
  "step": 9250
1386
  },
1387
  {
1388
- "epoch": 1.0009686793671295,
1389
- "grad_norm": 0.5368288159370422,
1390
- "learning_rate": 0.0005405559745717056,
1391
- "loss": 3.8048,
1392
  "step": 9300
1393
  },
1394
  {
1395
- "epoch": 1.0063502314067376,
1396
- "grad_norm": 0.5906519889831543,
1397
- "learning_rate": 0.0005402327335416442,
1398
- "loss": 3.7763,
1399
  "step": 9350
1400
  },
1401
  {
1402
- "epoch": 1.011731783446346,
1403
- "grad_norm": 0.6060296893119812,
1404
- "learning_rate": 0.0005399094925115827,
1405
- "loss": 3.7644,
1406
  "step": 9400
1407
  },
1408
  {
1409
- "epoch": 1.017113335485954,
1410
- "grad_norm": 0.5626549124717712,
1411
- "learning_rate": 0.0005395862514815214,
1412
- "loss": 3.7658,
1413
  "step": 9450
1414
  },
1415
  {
1416
- "epoch": 1.0224948875255624,
1417
- "grad_norm": 0.521769642829895,
1418
- "learning_rate": 0.0005392630104514599,
1419
- "loss": 3.7667,
1420
  "step": 9500
1421
  },
1422
  {
1423
- "epoch": 1.0278764395651705,
1424
- "grad_norm": 0.5677669048309326,
1425
- "learning_rate": 0.0005389397694213986,
1426
- "loss": 3.7803,
1427
  "step": 9550
1428
  },
1429
  {
1430
- "epoch": 1.0332579916047788,
1431
- "grad_norm": 0.5613917708396912,
1432
- "learning_rate": 0.0005386165283913371,
1433
- "loss": 3.7582,
1434
  "step": 9600
1435
  },
1436
  {
1437
- "epoch": 1.0386395436443872,
1438
- "grad_norm": 0.5857552289962769,
1439
- "learning_rate": 0.0005382932873612756,
1440
- "loss": 3.7806,
1441
  "step": 9650
1442
  },
1443
  {
1444
- "epoch": 1.0440210956839953,
1445
- "grad_norm": 0.5657854080200195,
1446
- "learning_rate": 0.0005379700463312143,
1447
- "loss": 3.7687,
1448
  "step": 9700
1449
  },
1450
  {
1451
- "epoch": 1.0494026477236036,
1452
- "grad_norm": 0.580058217048645,
1453
- "learning_rate": 0.0005376468053011528,
1454
- "loss": 3.7603,
1455
  "step": 9750
1456
  },
1457
  {
1458
- "epoch": 1.0547841997632117,
1459
- "grad_norm": 0.5268446207046509,
1460
- "learning_rate": 0.0005373235642710914,
1461
- "loss": 3.7591,
1462
  "step": 9800
1463
  },
1464
  {
1465
- "epoch": 1.06016575180282,
1466
- "grad_norm": 0.5513815879821777,
1467
- "learning_rate": 0.00053700032324103,
1468
- "loss": 3.7748,
1469
  "step": 9850
1470
  },
1471
  {
1472
- "epoch": 1.0655473038424281,
1473
- "grad_norm": 0.5946319699287415,
1474
- "learning_rate": 0.0005366770822109687,
1475
- "loss": 3.7816,
1476
  "step": 9900
1477
  },
1478
  {
1479
- "epoch": 1.0709288558820365,
1480
- "grad_norm": 0.5772368311882019,
1481
- "learning_rate": 0.0005363538411809072,
1482
- "loss": 3.754,
1483
  "step": 9950
1484
  },
1485
  {
1486
- "epoch": 1.0763104079216446,
1487
- "grad_norm": 0.6033820509910583,
1488
- "learning_rate": 0.0005360306001508457,
1489
- "loss": 3.7639,
1490
  "step": 10000
1491
  },
1492
  {
1493
- "epoch": 1.0763104079216446,
1494
- "eval_accuracy": 0.345006709859087,
1495
- "eval_loss": 3.7440133094787598,
1496
- "eval_runtime": 184.605,
1497
- "eval_samples_per_second": 97.565,
1498
- "eval_steps_per_second": 6.1,
1499
  "step": 10000
1500
  }
1501
  ],
1502
  "logging_steps": 50,
1503
- "max_steps": 92910,
1504
  "num_input_tokens_seen": 0,
1505
  "num_train_epochs": 10,
1506
  "save_steps": 10000,
@@ -1516,7 +1516,7 @@
1516
  "attributes": {}
1517
  }
1518
  },
1519
- "total_flos": 8.3611098611712e+16,
1520
  "train_batch_size": 32,
1521
  "trial_name": null,
1522
  "trial_params": null
 
1
  {
2
+ "best_metric": 3.746513843536377,
3
+ "best_model_checkpoint": "/scratch/cl5625/exceptions/models/100M_high_500_495/checkpoint-10000",
4
+ "epoch": 1.0781671159029649,
5
  "eval_steps": 1000,
6
  "global_step": 10000,
7
  "is_hyper_param_search": false,
 
9
  "is_world_process_zero": true,
10
  "log_history": [
11
  {
12
+ "epoch": 0.005390835579514825,
13
+ "grad_norm": 2.169931173324585,
14
+ "learning_rate": 0.0003,
15
+ "loss": 8.5202,
16
  "step": 50
17
  },
18
  {
19
+ "epoch": 0.01078167115902965,
20
+ "grad_norm": 4.110340118408203,
21
+ "learning_rate": 0.0006,
22
+ "loss": 6.9257,
23
  "step": 100
24
  },
25
  {
26
+ "epoch": 0.016172506738544475,
27
+ "grad_norm": 1.283010482788086,
28
+ "learning_rate": 0.0005996762007555315,
29
+ "loss": 6.5197,
30
  "step": 150
31
  },
32
  {
33
+ "epoch": 0.0215633423180593,
34
+ "grad_norm": 1.4407877922058105,
35
+ "learning_rate": 0.000599352401511063,
36
+ "loss": 6.2691,
37
  "step": 200
38
  },
39
  {
40
+ "epoch": 0.026954177897574125,
41
+ "grad_norm": 0.850818395614624,
42
+ "learning_rate": 0.0005990286022665946,
43
+ "loss": 6.0888,
44
  "step": 250
45
  },
46
  {
47
+ "epoch": 0.03234501347708895,
48
+ "grad_norm": 1.4988986253738403,
49
+ "learning_rate": 0.0005987048030221263,
50
+ "loss": 5.9867,
51
  "step": 300
52
  },
53
  {
54
+ "epoch": 0.03773584905660377,
55
+ "grad_norm": 1.1998897790908813,
56
+ "learning_rate": 0.0005983810037776578,
57
+ "loss": 5.8798,
58
  "step": 350
59
  },
60
  {
61
+ "epoch": 0.0431266846361186,
62
+ "grad_norm": 1.2407352924346924,
63
+ "learning_rate": 0.0005980572045331894,
64
+ "loss": 5.8211,
65
  "step": 400
66
  },
67
  {
68
+ "epoch": 0.04851752021563342,
69
+ "grad_norm": 0.9971290826797485,
70
+ "learning_rate": 0.0005977334052887209,
71
+ "loss": 5.7361,
72
  "step": 450
73
  },
74
  {
75
+ "epoch": 0.05390835579514825,
76
+ "grad_norm": 1.255806803703308,
77
+ "learning_rate": 0.0005974096060442526,
78
+ "loss": 5.683,
79
  "step": 500
80
  },
81
  {
82
+ "epoch": 0.05929919137466307,
83
+ "grad_norm": 1.75487220287323,
84
+ "learning_rate": 0.0005970858067997841,
85
+ "loss": 5.5637,
86
  "step": 550
87
  },
88
  {
89
+ "epoch": 0.0646900269541779,
90
+ "grad_norm": 1.2305444478988647,
91
+ "learning_rate": 0.0005967620075553157,
92
+ "loss": 5.5257,
93
  "step": 600
94
  },
95
  {
96
+ "epoch": 0.07008086253369272,
97
+ "grad_norm": 1.346718668937683,
98
+ "learning_rate": 0.0005964382083108472,
99
+ "loss": 5.4252,
100
  "step": 650
101
  },
102
  {
103
+ "epoch": 0.07547169811320754,
104
+ "grad_norm": 1.0916610956192017,
105
+ "learning_rate": 0.0005961144090663788,
106
+ "loss": 5.3969,
107
  "step": 700
108
  },
109
  {
110
+ "epoch": 0.08086253369272237,
111
+ "grad_norm": 1.0708554983139038,
112
+ "learning_rate": 0.0005957906098219104,
113
+ "loss": 5.3257,
114
  "step": 750
115
  },
116
  {
117
+ "epoch": 0.0862533692722372,
118
+ "grad_norm": 1.4724178314208984,
119
+ "learning_rate": 0.0005954668105774419,
120
+ "loss": 5.2774,
121
  "step": 800
122
  },
123
  {
124
+ "epoch": 0.09164420485175202,
125
+ "grad_norm": 1.0166338682174683,
126
+ "learning_rate": 0.0005951430113329735,
127
+ "loss": 5.2416,
128
  "step": 850
129
  },
130
  {
131
+ "epoch": 0.09703504043126684,
132
+ "grad_norm": 1.2486152648925781,
133
+ "learning_rate": 0.0005948192120885051,
134
+ "loss": 5.155,
135
  "step": 900
136
  },
137
  {
138
+ "epoch": 0.10242587601078167,
139
+ "grad_norm": 0.8233184814453125,
140
+ "learning_rate": 0.0005944954128440366,
141
+ "loss": 5.1494,
142
  "step": 950
143
  },
144
  {
145
+ "epoch": 0.1078167115902965,
146
+ "grad_norm": 1.1141846179962158,
147
+ "learning_rate": 0.0005941716135995682,
148
+ "loss": 5.0928,
149
  "step": 1000
150
  },
151
  {
152
+ "epoch": 0.1078167115902965,
153
+ "eval_accuracy": 0.22779479945990821,
154
+ "eval_loss": 5.019223690032959,
155
+ "eval_runtime": 184.6889,
156
+ "eval_samples_per_second": 97.521,
157
+ "eval_steps_per_second": 6.097,
158
  "step": 1000
159
  },
160
  {
161
+ "epoch": 0.11320754716981132,
162
+ "grad_norm": 1.1920228004455566,
163
+ "learning_rate": 0.0005938478143550997,
164
+ "loss": 5.0528,
165
  "step": 1050
166
  },
167
  {
168
+ "epoch": 0.11859838274932614,
169
+ "grad_norm": 1.3988215923309326,
170
+ "learning_rate": 0.0005935240151106314,
171
+ "loss": 5.0162,
172
  "step": 1100
173
  },
174
  {
175
+ "epoch": 0.12398921832884097,
176
+ "grad_norm": 1.1493585109710693,
177
+ "learning_rate": 0.0005932002158661629,
178
+ "loss": 4.9947,
179
  "step": 1150
180
  },
181
  {
182
+ "epoch": 0.1293800539083558,
183
+ "grad_norm": 0.8649298548698425,
184
+ "learning_rate": 0.0005928764166216945,
185
+ "loss": 4.9752,
186
  "step": 1200
187
  },
188
  {
189
+ "epoch": 0.1347708894878706,
190
+ "grad_norm": 1.2792754173278809,
191
+ "learning_rate": 0.000592552617377226,
192
+ "loss": 4.9485,
193
  "step": 1250
194
  },
195
  {
196
+ "epoch": 0.14016172506738545,
197
+ "grad_norm": 0.939761757850647,
198
+ "learning_rate": 0.0005922288181327577,
199
+ "loss": 4.8944,
200
  "step": 1300
201
  },
202
  {
203
+ "epoch": 0.14555256064690028,
204
+ "grad_norm": 1.2719601392745972,
205
+ "learning_rate": 0.0005919050188882893,
206
+ "loss": 4.8786,
207
  "step": 1350
208
  },
209
  {
210
+ "epoch": 0.1509433962264151,
211
+ "grad_norm": 0.7670081853866577,
212
+ "learning_rate": 0.0005915812196438207,
213
+ "loss": 4.8498,
214
  "step": 1400
215
  },
216
  {
217
+ "epoch": 0.15633423180592992,
218
+ "grad_norm": 1.2284456491470337,
219
+ "learning_rate": 0.0005912574203993524,
220
+ "loss": 4.8299,
221
  "step": 1450
222
  },
223
  {
224
+ "epoch": 0.16172506738544473,
225
+ "grad_norm": 0.8652415871620178,
226
+ "learning_rate": 0.0005909336211548839,
227
+ "loss": 4.8202,
228
  "step": 1500
229
  },
230
  {
231
+ "epoch": 0.16711590296495957,
232
+ "grad_norm": 0.8690969944000244,
233
+ "learning_rate": 0.0005906098219104155,
234
+ "loss": 4.7751,
235
  "step": 1550
236
  },
237
  {
238
+ "epoch": 0.1725067385444744,
239
+ "grad_norm": 1.03486967086792,
240
+ "learning_rate": 0.000590286022665947,
241
+ "loss": 4.7912,
242
  "step": 1600
243
  },
244
  {
245
+ "epoch": 0.1778975741239892,
246
+ "grad_norm": 0.8726552724838257,
247
+ "learning_rate": 0.0005899622234214787,
248
+ "loss": 4.7651,
249
  "step": 1650
250
  },
251
  {
252
+ "epoch": 0.18328840970350405,
253
+ "grad_norm": 0.8985825181007385,
254
+ "learning_rate": 0.0005896384241770102,
255
+ "loss": 4.6925,
256
  "step": 1700
257
  },
258
  {
259
+ "epoch": 0.18867924528301888,
260
+ "grad_norm": 1.0985790491104126,
261
+ "learning_rate": 0.0005893146249325418,
262
+ "loss": 4.7067,
263
  "step": 1750
264
  },
265
  {
266
+ "epoch": 0.1940700808625337,
267
+ "grad_norm": 1.0200293064117432,
268
+ "learning_rate": 0.0005889908256880733,
269
+ "loss": 4.6942,
270
  "step": 1800
271
  },
272
  {
273
+ "epoch": 0.19946091644204852,
274
+ "grad_norm": 1.0469223260879517,
275
+ "learning_rate": 0.0005886670264436049,
276
+ "loss": 4.6484,
277
  "step": 1850
278
  },
279
  {
280
+ "epoch": 0.20485175202156333,
281
+ "grad_norm": 0.8023126721382141,
282
+ "learning_rate": 0.0005883432271991365,
283
+ "loss": 4.6433,
284
  "step": 1900
285
  },
286
  {
287
+ "epoch": 0.21024258760107817,
288
+ "grad_norm": 1.0328996181488037,
289
+ "learning_rate": 0.0005880194279546681,
290
+ "loss": 4.6104,
291
  "step": 1950
292
  },
293
  {
294
+ "epoch": 0.215633423180593,
295
+ "grad_norm": 1.1157277822494507,
296
+ "learning_rate": 0.0005876956287101996,
297
+ "loss": 4.586,
298
  "step": 2000
299
  },
300
  {
301
+ "epoch": 0.215633423180593,
302
+ "eval_accuracy": 0.26792661279728924,
303
+ "eval_loss": 4.539734363555908,
304
+ "eval_runtime": 183.597,
305
+ "eval_samples_per_second": 98.101,
306
+ "eval_steps_per_second": 6.133,
307
  "step": 2000
308
  },
309
  {
310
+ "epoch": 0.2210242587601078,
311
+ "grad_norm": 0.802883505821228,
312
+ "learning_rate": 0.0005873718294657312,
313
+ "loss": 4.6011,
314
  "step": 2050
315
  },
316
  {
317
+ "epoch": 0.22641509433962265,
318
+ "grad_norm": 0.9819656014442444,
319
+ "learning_rate": 0.0005870480302212628,
320
+ "loss": 4.5627,
321
  "step": 2100
322
  },
323
  {
324
+ "epoch": 0.23180592991913745,
325
+ "grad_norm": 1.0422033071517944,
326
+ "learning_rate": 0.0005867242309767943,
327
+ "loss": 4.5141,
328
  "step": 2150
329
  },
330
  {
331
+ "epoch": 0.2371967654986523,
332
+ "grad_norm": 0.9134729504585266,
333
+ "learning_rate": 0.0005864004317323259,
334
+ "loss": 4.5497,
335
  "step": 2200
336
  },
337
  {
338
+ "epoch": 0.24258760107816713,
339
+ "grad_norm": 0.7439409494400024,
340
+ "learning_rate": 0.0005860766324878575,
341
+ "loss": 4.486,
342
  "step": 2250
343
  },
344
  {
345
+ "epoch": 0.24797843665768193,
346
+ "grad_norm": 1.0813612937927246,
347
+ "learning_rate": 0.000585752833243389,
348
+ "loss": 4.4643,
349
  "step": 2300
350
  },
351
  {
352
+ "epoch": 0.25336927223719674,
353
+ "grad_norm": 0.8331607580184937,
354
+ "learning_rate": 0.0005854290339989206,
355
+ "loss": 4.4584,
356
  "step": 2350
357
  },
358
  {
359
+ "epoch": 0.2587601078167116,
360
+ "grad_norm": 0.8642703294754028,
361
+ "learning_rate": 0.0005851052347544521,
362
+ "loss": 4.4498,
363
  "step": 2400
364
  },
365
  {
366
+ "epoch": 0.2641509433962264,
367
+ "grad_norm": 0.829148530960083,
368
+ "learning_rate": 0.0005847814355099838,
369
+ "loss": 4.453,
370
  "step": 2450
371
  },
372
  {
373
+ "epoch": 0.2695417789757412,
374
+ "grad_norm": 0.9349790215492249,
375
+ "learning_rate": 0.0005844576362655154,
376
+ "loss": 4.4526,
377
  "step": 2500
378
  },
379
  {
380
+ "epoch": 0.2749326145552561,
381
+ "grad_norm": 0.8753833770751953,
382
+ "learning_rate": 0.0005841338370210469,
383
+ "loss": 4.412,
384
  "step": 2550
385
  },
386
  {
387
+ "epoch": 0.2803234501347709,
388
+ "grad_norm": 0.7982025146484375,
389
+ "learning_rate": 0.0005838100377765785,
390
+ "loss": 4.4007,
391
  "step": 2600
392
  },
393
  {
394
+ "epoch": 0.2857142857142857,
395
+ "grad_norm": 1.1045323610305786,
396
+ "learning_rate": 0.0005834862385321101,
397
+ "loss": 4.3978,
398
  "step": 2650
399
  },
400
  {
401
+ "epoch": 0.29110512129380056,
402
+ "grad_norm": 0.873582124710083,
403
+ "learning_rate": 0.0005831624392876417,
404
+ "loss": 4.3911,
405
  "step": 2700
406
  },
407
  {
408
+ "epoch": 0.29649595687331537,
409
+ "grad_norm": 0.9419072866439819,
410
+ "learning_rate": 0.0005828386400431731,
411
+ "loss": 4.3527,
412
  "step": 2750
413
  },
414
  {
415
+ "epoch": 0.3018867924528302,
416
+ "grad_norm": 0.6602005958557129,
417
+ "learning_rate": 0.0005825148407987048,
418
+ "loss": 4.3873,
419
  "step": 2800
420
  },
421
  {
422
+ "epoch": 0.30727762803234504,
423
+ "grad_norm": 0.7998479008674622,
424
+ "learning_rate": 0.0005821910415542363,
425
+ "loss": 4.3582,
426
  "step": 2850
427
  },
428
  {
429
+ "epoch": 0.31266846361185985,
430
+ "grad_norm": 0.7541155219078064,
431
+ "learning_rate": 0.0005818672423097679,
432
+ "loss": 4.337,
433
  "step": 2900
434
  },
435
  {
436
+ "epoch": 0.31805929919137466,
437
+ "grad_norm": 0.8266378045082092,
438
+ "learning_rate": 0.0005815434430652994,
439
+ "loss": 4.3355,
440
  "step": 2950
441
  },
442
  {
443
+ "epoch": 0.32345013477088946,
444
+ "grad_norm": 0.7615757584571838,
445
+ "learning_rate": 0.0005812196438208311,
446
+ "loss": 4.3331,
447
  "step": 3000
448
  },
449
  {
450
+ "epoch": 0.32345013477088946,
451
+ "eval_accuracy": 0.2979562065843433,
452
+ "eval_loss": 4.236091613769531,
453
+ "eval_runtime": 183.7161,
454
+ "eval_samples_per_second": 98.037,
455
+ "eval_steps_per_second": 6.129,
456
  "step": 3000
457
  },
458
  {
459
+ "epoch": 0.3288409703504043,
460
+ "grad_norm": 0.9040746688842773,
461
+ "learning_rate": 0.0005808958445763626,
462
+ "loss": 4.3105,
463
  "step": 3050
464
  },
465
  {
466
+ "epoch": 0.33423180592991913,
467
+ "grad_norm": 0.6944011449813843,
468
+ "learning_rate": 0.0005805720453318942,
469
+ "loss": 4.3025,
470
  "step": 3100
471
  },
472
  {
473
+ "epoch": 0.33962264150943394,
474
+ "grad_norm": 0.6618968844413757,
475
+ "learning_rate": 0.0005802482460874257,
476
+ "loss": 4.2711,
477
  "step": 3150
478
  },
479
  {
480
+ "epoch": 0.3450134770889488,
481
+ "grad_norm": 0.7913966178894043,
482
+ "learning_rate": 0.0005799244468429573,
483
+ "loss": 4.279,
484
  "step": 3200
485
  },
486
  {
487
+ "epoch": 0.3504043126684636,
488
+ "grad_norm": 0.7068710923194885,
489
+ "learning_rate": 0.0005796006475984889,
490
+ "loss": 4.2522,
491
  "step": 3250
492
  },
493
  {
494
+ "epoch": 0.3557951482479784,
495
+ "grad_norm": 0.7139419317245483,
496
+ "learning_rate": 0.0005792768483540205,
497
+ "loss": 4.2757,
498
  "step": 3300
499
  },
500
  {
501
+ "epoch": 0.3611859838274933,
502
+ "grad_norm": 0.7112064957618713,
503
+ "learning_rate": 0.000578953049109552,
504
+ "loss": 4.255,
505
  "step": 3350
506
  },
507
  {
508
+ "epoch": 0.3665768194070081,
509
+ "grad_norm": 0.824454665184021,
510
+ "learning_rate": 0.0005786292498650836,
511
+ "loss": 4.2453,
512
  "step": 3400
513
  },
514
  {
515
+ "epoch": 0.3719676549865229,
516
+ "grad_norm": 0.651171863079071,
517
+ "learning_rate": 0.0005783054506206152,
518
+ "loss": 4.2315,
519
  "step": 3450
520
  },
521
  {
522
+ "epoch": 0.37735849056603776,
523
+ "grad_norm": 0.6538099646568298,
524
+ "learning_rate": 0.0005779816513761467,
525
+ "loss": 4.2093,
526
  "step": 3500
527
  },
528
  {
529
+ "epoch": 0.38274932614555257,
530
+ "grad_norm": 0.6883544325828552,
531
+ "learning_rate": 0.0005776578521316782,
532
+ "loss": 4.2153,
533
  "step": 3550
534
  },
535
  {
536
+ "epoch": 0.3881401617250674,
537
+ "grad_norm": 0.8138344287872314,
538
+ "learning_rate": 0.0005773340528872099,
539
+ "loss": 4.2188,
540
  "step": 3600
541
  },
542
  {
543
+ "epoch": 0.3935309973045822,
544
+ "grad_norm": 0.9981099963188171,
545
+ "learning_rate": 0.0005770102536427414,
546
+ "loss": 4.2288,
547
  "step": 3650
548
  },
549
  {
550
+ "epoch": 0.39892183288409705,
551
+ "grad_norm": 0.7214846014976501,
552
+ "learning_rate": 0.000576686454398273,
553
+ "loss": 4.1953,
554
  "step": 3700
555
  },
556
  {
557
+ "epoch": 0.40431266846361186,
558
+ "grad_norm": 0.7932049036026001,
559
+ "learning_rate": 0.0005763626551538045,
560
+ "loss": 4.1891,
561
  "step": 3750
562
  },
563
  {
564
+ "epoch": 0.40970350404312667,
565
+ "grad_norm": 0.7664895057678223,
566
+ "learning_rate": 0.0005760388559093362,
567
+ "loss": 4.214,
568
  "step": 3800
569
  },
570
  {
571
+ "epoch": 0.41509433962264153,
572
+ "grad_norm": 0.8450409770011902,
573
+ "learning_rate": 0.0005757150566648678,
574
+ "loss": 4.1912,
575
  "step": 3850
576
  },
577
  {
578
+ "epoch": 0.42048517520215634,
579
+ "grad_norm": 0.9285402297973633,
580
+ "learning_rate": 0.0005753912574203993,
581
+ "loss": 4.1599,
582
  "step": 3900
583
  },
584
  {
585
+ "epoch": 0.42587601078167114,
586
+ "grad_norm": 0.6313654184341431,
587
+ "learning_rate": 0.0005750674581759309,
588
+ "loss": 4.1601,
589
  "step": 3950
590
  },
591
  {
592
+ "epoch": 0.431266846361186,
593
+ "grad_norm": 0.6990293264389038,
594
+ "learning_rate": 0.0005747436589314624,
595
+ "loss": 4.1641,
596
  "step": 4000
597
  },
598
  {
599
+ "epoch": 0.431266846361186,
600
+ "eval_accuracy": 0.31198546745894906,
601
+ "eval_loss": 4.094001293182373,
602
+ "eval_runtime": 183.5145,
603
+ "eval_samples_per_second": 98.145,
604
+ "eval_steps_per_second": 6.136,
605
  "step": 4000
606
  },
607
  {
608
+ "epoch": 0.4366576819407008,
609
+ "grad_norm": 0.6404671669006348,
610
+ "learning_rate": 0.0005744198596869941,
611
+ "loss": 4.1621,
612
  "step": 4050
613
  },
614
  {
615
+ "epoch": 0.4420485175202156,
616
+ "grad_norm": 0.6013748049736023,
617
+ "learning_rate": 0.0005740960604425255,
618
+ "loss": 4.1719,
619
  "step": 4100
620
  },
621
  {
622
+ "epoch": 0.4474393530997305,
623
+ "grad_norm": 0.7267091274261475,
624
+ "learning_rate": 0.0005737722611980572,
625
+ "loss": 4.1522,
626
  "step": 4150
627
  },
628
  {
629
+ "epoch": 0.4528301886792453,
630
+ "grad_norm": 0.6014766097068787,
631
+ "learning_rate": 0.0005734484619535887,
632
+ "loss": 4.1425,
633
  "step": 4200
634
  },
635
  {
636
+ "epoch": 0.4582210242587601,
637
+ "grad_norm": 0.5657269358634949,
638
+ "learning_rate": 0.0005731246627091203,
639
+ "loss": 4.1337,
640
  "step": 4250
641
  },
642
  {
643
+ "epoch": 0.4636118598382749,
644
+ "grad_norm": 0.6378490328788757,
645
+ "learning_rate": 0.0005728008634646518,
646
+ "loss": 4.126,
647
  "step": 4300
648
  },
649
  {
650
+ "epoch": 0.46900269541778977,
651
+ "grad_norm": 0.739479660987854,
652
+ "learning_rate": 0.0005724770642201835,
653
+ "loss": 4.1349,
654
  "step": 4350
655
  },
656
  {
657
+ "epoch": 0.4743935309973046,
658
+ "grad_norm": 0.7157710790634155,
659
+ "learning_rate": 0.000572153264975715,
660
+ "loss": 4.1215,
661
  "step": 4400
662
  },
663
  {
664
+ "epoch": 0.4797843665768194,
665
+ "grad_norm": 0.7867320775985718,
666
+ "learning_rate": 0.0005718294657312466,
667
+ "loss": 4.1252,
668
  "step": 4450
669
  },
670
  {
671
+ "epoch": 0.48517520215633425,
672
+ "grad_norm": 0.622015118598938,
673
+ "learning_rate": 0.0005715056664867781,
674
+ "loss": 4.108,
675
  "step": 4500
676
  },
677
  {
678
+ "epoch": 0.49056603773584906,
679
+ "grad_norm": 0.6655036807060242,
680
+ "learning_rate": 0.0005711818672423097,
681
+ "loss": 4.0987,
682
  "step": 4550
683
  },
684
  {
685
+ "epoch": 0.49595687331536387,
686
+ "grad_norm": 0.6496259570121765,
687
+ "learning_rate": 0.0005708580679978413,
688
+ "loss": 4.0999,
689
  "step": 4600
690
  },
691
  {
692
+ "epoch": 0.5013477088948787,
693
+ "grad_norm": 0.6310088038444519,
694
+ "learning_rate": 0.0005705342687533729,
695
+ "loss": 4.0796,
696
  "step": 4650
697
  },
698
  {
699
+ "epoch": 0.5067385444743935,
700
+ "grad_norm": 0.8397471308708191,
701
+ "learning_rate": 0.0005702104695089044,
702
+ "loss": 4.0885,
703
  "step": 4700
704
  },
705
  {
706
+ "epoch": 0.5121293800539084,
707
+ "grad_norm": 0.6343526840209961,
708
+ "learning_rate": 0.000569886670264436,
709
+ "loss": 4.0699,
710
  "step": 4750
711
  },
712
  {
713
+ "epoch": 0.5175202156334232,
714
+ "grad_norm": 0.6864327192306519,
715
+ "learning_rate": 0.0005695628710199675,
716
+ "loss": 4.0858,
717
  "step": 4800
718
  },
719
  {
720
+ "epoch": 0.522911051212938,
721
+ "grad_norm": 0.6211134195327759,
722
+ "learning_rate": 0.0005692390717754991,
723
+ "loss": 4.0786,
724
  "step": 4850
725
  },
726
  {
727
+ "epoch": 0.5283018867924528,
728
+ "grad_norm": 0.6251089572906494,
729
+ "learning_rate": 0.0005689152725310306,
730
+ "loss": 4.0887,
731
  "step": 4900
732
  },
733
  {
734
+ "epoch": 0.5336927223719676,
735
+ "grad_norm": 0.691362738609314,
736
+ "learning_rate": 0.0005685914732865623,
737
+ "loss": 4.061,
738
  "step": 4950
739
  },
740
  {
741
+ "epoch": 0.5390835579514824,
742
+ "grad_norm": 0.6830545663833618,
743
+ "learning_rate": 0.0005682676740420939,
744
+ "loss": 4.0484,
745
  "step": 5000
746
  },
747
  {
748
+ "epoch": 0.5390835579514824,
749
+ "eval_accuracy": 0.3213867672299848,
750
+ "eval_loss": 3.994671106338501,
751
+ "eval_runtime": 183.5628,
752
+ "eval_samples_per_second": 98.119,
753
+ "eval_steps_per_second": 6.134,
754
  "step": 5000
755
  },
756
  {
757
+ "epoch": 0.5444743935309974,
758
+ "grad_norm": 0.5907337069511414,
759
+ "learning_rate": 0.0005679438747976254,
760
+ "loss": 4.045,
761
  "step": 5050
762
  },
763
  {
764
+ "epoch": 0.5498652291105122,
765
+ "grad_norm": 0.8197788000106812,
766
+ "learning_rate": 0.000567620075553157,
767
+ "loss": 4.0488,
768
  "step": 5100
769
  },
770
  {
771
+ "epoch": 0.555256064690027,
772
+ "grad_norm": 0.6527996063232422,
773
+ "learning_rate": 0.0005672962763086886,
774
+ "loss": 4.0536,
775
  "step": 5150
776
  },
777
  {
778
+ "epoch": 0.5606469002695418,
779
+ "grad_norm": 0.6831157803535461,
780
+ "learning_rate": 0.0005669724770642202,
781
+ "loss": 4.0314,
782
  "step": 5200
783
  },
784
  {
785
+ "epoch": 0.5660377358490566,
786
+ "grad_norm": 0.6056663990020752,
787
+ "learning_rate": 0.0005666486778197517,
788
+ "loss": 4.0345,
789
  "step": 5250
790
  },
791
  {
792
+ "epoch": 0.5714285714285714,
793
+ "grad_norm": 0.70479416847229,
794
+ "learning_rate": 0.0005663248785752833,
795
+ "loss": 4.0151,
796
  "step": 5300
797
  },
798
  {
799
+ "epoch": 0.5768194070080862,
800
+ "grad_norm": 0.6906822919845581,
801
+ "learning_rate": 0.0005660010793308148,
802
+ "loss": 4.0303,
803
  "step": 5350
804
  },
805
  {
806
+ "epoch": 0.5822102425876011,
807
+ "grad_norm": 0.6968085765838623,
808
+ "learning_rate": 0.0005656772800863465,
809
+ "loss": 4.0215,
810
  "step": 5400
811
  },
812
  {
813
+ "epoch": 0.5876010781671159,
814
+ "grad_norm": 0.8021159768104553,
815
+ "learning_rate": 0.0005653534808418779,
816
+ "loss": 4.0326,
817
  "step": 5450
818
  },
819
  {
820
+ "epoch": 0.5929919137466307,
821
+ "grad_norm": 0.6045252084732056,
822
+ "learning_rate": 0.0005650296815974096,
823
+ "loss": 4.033,
824
  "step": 5500
825
  },
826
  {
827
+ "epoch": 0.5983827493261455,
828
+ "grad_norm": 0.6504480838775635,
829
+ "learning_rate": 0.0005647058823529411,
830
+ "loss": 4.0368,
831
  "step": 5550
832
  },
833
  {
834
+ "epoch": 0.6037735849056604,
835
+ "grad_norm": 0.6711751222610474,
836
+ "learning_rate": 0.0005643820831084727,
837
+ "loss": 4.0154,
838
  "step": 5600
839
  },
840
  {
841
+ "epoch": 0.6091644204851752,
842
+ "grad_norm": 0.5429975390434265,
843
+ "learning_rate": 0.0005640582838640042,
844
+ "loss": 4.004,
845
  "step": 5650
846
  },
847
  {
848
+ "epoch": 0.6145552560646901,
849
+ "grad_norm": 0.6812418103218079,
850
+ "learning_rate": 0.0005637344846195358,
851
+ "loss": 4.0092,
852
  "step": 5700
853
  },
854
  {
855
+ "epoch": 0.6199460916442049,
856
+ "grad_norm": 0.67169588804245,
857
+ "learning_rate": 0.0005634106853750674,
858
+ "loss": 4.0189,
859
  "step": 5750
860
  },
861
  {
862
+ "epoch": 0.6253369272237197,
863
+ "grad_norm": 0.67845219373703,
864
+ "learning_rate": 0.000563086886130599,
865
+ "loss": 4.0127,
866
  "step": 5800
867
  },
868
  {
869
+ "epoch": 0.6307277628032345,
870
+ "grad_norm": 0.6689065098762512,
871
+ "learning_rate": 0.0005627630868861305,
872
+ "loss": 4.0195,
873
  "step": 5850
874
  },
875
  {
876
+ "epoch": 0.6361185983827493,
877
+ "grad_norm": 0.6098775267601013,
878
+ "learning_rate": 0.0005624392876416621,
879
  "loss": 3.9892,
880
  "step": 5900
881
  },
882
  {
883
+ "epoch": 0.6415094339622641,
884
+ "grad_norm": 0.5924386382102966,
885
+ "learning_rate": 0.0005621154883971937,
886
+ "loss": 3.9839,
887
  "step": 5950
888
  },
889
  {
890
+ "epoch": 0.6469002695417789,
891
+ "grad_norm": 0.5603781342506409,
892
+ "learning_rate": 0.0005617916891527253,
893
+ "loss": 3.9955,
894
  "step": 6000
895
  },
896
  {
897
+ "epoch": 0.6469002695417789,
898
+ "eval_accuracy": 0.32812498472068763,
899
+ "eval_loss": 3.916694164276123,
900
+ "eval_runtime": 183.5467,
901
+ "eval_samples_per_second": 98.128,
902
+ "eval_steps_per_second": 6.135,
903
  "step": 6000
904
  },
905
  {
906
+ "epoch": 0.6522911051212938,
907
+ "grad_norm": 0.7233008742332458,
908
+ "learning_rate": 0.0005614678899082568,
909
+ "loss": 3.9679,
910
  "step": 6050
911
  },
912
  {
913
+ "epoch": 0.6576819407008087,
914
+ "grad_norm": 0.7258248329162598,
915
+ "learning_rate": 0.0005611440906637884,
916
+ "loss": 3.9993,
917
  "step": 6100
918
  },
919
  {
920
+ "epoch": 0.6630727762803235,
921
+ "grad_norm": 0.6090084910392761,
922
+ "learning_rate": 0.00056082029141932,
923
+ "loss": 3.9792,
924
  "step": 6150
925
  },
926
  {
927
+ "epoch": 0.6684636118598383,
928
+ "grad_norm": 0.667877197265625,
929
+ "learning_rate": 0.0005605029681597409,
930
+ "loss": 3.9585,
931
  "step": 6200
932
  },
933
  {
934
+ "epoch": 0.6738544474393531,
935
+ "grad_norm": 0.6215923428535461,
936
+ "learning_rate": 0.0005601791689152725,
937
+ "loss": 3.9753,
938
  "step": 6250
939
  },
940
  {
941
+ "epoch": 0.6792452830188679,
942
+ "grad_norm": 0.6659998893737793,
943
+ "learning_rate": 0.000559855369670804,
944
+ "loss": 3.9925,
945
  "step": 6300
946
  },
947
  {
948
+ "epoch": 0.6846361185983828,
949
+ "grad_norm": 0.6712725758552551,
950
+ "learning_rate": 0.0005595315704263357,
951
+ "loss": 3.9557,
952
  "step": 6350
953
  },
954
  {
955
+ "epoch": 0.6900269541778976,
956
+ "grad_norm": 0.5836709141731262,
957
+ "learning_rate": 0.0005592077711818672,
958
+ "loss": 3.9608,
959
  "step": 6400
960
  },
961
  {
962
+ "epoch": 0.6954177897574124,
963
+ "grad_norm": 0.6350057721138,
964
+ "learning_rate": 0.0005588839719373988,
965
+ "loss": 3.9449,
966
  "step": 6450
967
  },
968
  {
969
+ "epoch": 0.7008086253369272,
970
+ "grad_norm": 0.6150210499763489,
971
+ "learning_rate": 0.0005585601726929303,
972
+ "loss": 3.9582,
973
  "step": 6500
974
  },
975
  {
976
+ "epoch": 0.706199460916442,
977
+ "grad_norm": 0.598513126373291,
978
+ "learning_rate": 0.0005582363734484619,
979
+ "loss": 3.9816,
980
  "step": 6550
981
  },
982
  {
983
+ "epoch": 0.7115902964959568,
984
+ "grad_norm": 0.6591358184814453,
985
+ "learning_rate": 0.0005579125742039935,
986
+ "loss": 3.937,
987
  "step": 6600
988
  },
989
  {
990
+ "epoch": 0.7169811320754716,
991
+ "grad_norm": 0.5885124206542969,
992
+ "learning_rate": 0.0005575887749595251,
993
+ "loss": 3.9678,
994
  "step": 6650
995
  },
996
  {
997
+ "epoch": 0.7223719676549866,
998
+ "grad_norm": 0.6395810842514038,
999
+ "learning_rate": 0.0005572649757150566,
1000
  "loss": 3.9427,
1001
  "step": 6700
1002
  },
1003
  {
1004
+ "epoch": 0.7277628032345014,
1005
+ "grad_norm": 0.5958187580108643,
1006
+ "learning_rate": 0.0005569411764705882,
1007
+ "loss": 3.9486,
1008
  "step": 6750
1009
  },
1010
  {
1011
+ "epoch": 0.7331536388140162,
1012
+ "grad_norm": 0.5831095576286316,
1013
+ "learning_rate": 0.0005566173772261198,
1014
+ "loss": 3.9383,
1015
  "step": 6800
1016
  },
1017
  {
1018
+ "epoch": 0.738544474393531,
1019
+ "grad_norm": 0.6585524082183838,
1020
+ "learning_rate": 0.0005562935779816513,
1021
+ "loss": 3.9173,
1022
  "step": 6850
1023
  },
1024
  {
1025
+ "epoch": 0.7439353099730458,
1026
+ "grad_norm": 0.705803394317627,
1027
+ "learning_rate": 0.0005559697787371828,
1028
+ "loss": 3.945,
1029
  "step": 6900
1030
  },
1031
  {
1032
+ "epoch": 0.7493261455525606,
1033
+ "grad_norm": 0.634233295917511,
1034
+ "learning_rate": 0.0005556459794927145,
1035
+ "loss": 3.9407,
1036
  "step": 6950
1037
  },
1038
  {
1039
+ "epoch": 0.7547169811320755,
1040
+ "grad_norm": 0.6854012608528137,
1041
+ "learning_rate": 0.000555322180248246,
1042
+ "loss": 3.9273,
1043
  "step": 7000
1044
  },
1045
  {
1046
+ "epoch": 0.7547169811320755,
1047
+ "eval_accuracy": 0.3336400966532629,
1048
+ "eval_loss": 3.8616080284118652,
1049
+ "eval_runtime": 183.6309,
1050
+ "eval_samples_per_second": 98.083,
1051
+ "eval_steps_per_second": 6.132,
1052
  "step": 7000
1053
  },
1054
  {
1055
+ "epoch": 0.7601078167115903,
1056
+ "grad_norm": 0.6441482901573181,
1057
+ "learning_rate": 0.0005549983810037776,
1058
+ "loss": 3.9186,
1059
  "step": 7050
1060
  },
1061
  {
1062
+ "epoch": 0.7654986522911051,
1063
+ "grad_norm": 0.6202889084815979,
1064
+ "learning_rate": 0.0005546745817593091,
1065
+ "loss": 3.9302,
1066
  "step": 7100
1067
  },
1068
  {
1069
+ "epoch": 0.77088948787062,
1070
+ "grad_norm": 0.518191933631897,
1071
+ "learning_rate": 0.0005543507825148408,
1072
+ "loss": 3.9288,
1073
  "step": 7150
1074
  },
1075
  {
1076
+ "epoch": 0.7762803234501348,
1077
+ "grad_norm": 0.6395387053489685,
1078
+ "learning_rate": 0.0005540269832703723,
1079
+ "loss": 3.9317,
1080
  "step": 7200
1081
  },
1082
  {
1083
+ "epoch": 0.7816711590296496,
1084
+ "grad_norm": 0.578109860420227,
1085
+ "learning_rate": 0.0005537031840259039,
1086
+ "loss": 3.887,
1087
  "step": 7250
1088
  },
1089
  {
1090
+ "epoch": 0.7870619946091644,
1091
+ "grad_norm": 0.7136623859405518,
1092
+ "learning_rate": 0.0005533793847814354,
1093
+ "loss": 3.9282,
1094
  "step": 7300
1095
  },
1096
  {
1097
+ "epoch": 0.7924528301886793,
1098
+ "grad_norm": 0.615642786026001,
1099
+ "learning_rate": 0.000553055585536967,
1100
+ "loss": 3.9137,
1101
  "step": 7350
1102
  },
1103
  {
1104
+ "epoch": 0.7978436657681941,
1105
+ "grad_norm": 0.6420246362686157,
1106
+ "learning_rate": 0.0005527317862924987,
1107
+ "loss": 3.9119,
1108
  "step": 7400
1109
  },
1110
  {
1111
+ "epoch": 0.8032345013477089,
1112
+ "grad_norm": 0.5536246299743652,
1113
+ "learning_rate": 0.0005524079870480301,
1114
+ "loss": 3.9007,
1115
  "step": 7450
1116
  },
1117
  {
1118
+ "epoch": 0.8086253369272237,
1119
+ "grad_norm": 0.6325304508209229,
1120
+ "learning_rate": 0.0005520841878035618,
1121
+ "loss": 3.8882,
1122
  "step": 7500
1123
  },
1124
  {
1125
+ "epoch": 0.8140161725067385,
1126
+ "grad_norm": 0.6307028532028198,
1127
+ "learning_rate": 0.0005517603885590933,
1128
+ "loss": 3.9159,
1129
  "step": 7550
1130
  },
1131
  {
1132
+ "epoch": 0.8194070080862533,
1133
+ "grad_norm": 0.5926609039306641,
1134
+ "learning_rate": 0.0005514365893146249,
1135
+ "loss": 3.9175,
1136
  "step": 7600
1137
  },
1138
  {
1139
+ "epoch": 0.8247978436657682,
1140
+ "grad_norm": 0.5906722545623779,
1141
+ "learning_rate": 0.0005511127900701564,
1142
+ "loss": 3.8928,
1143
  "step": 7650
1144
  },
1145
  {
1146
+ "epoch": 0.8301886792452831,
1147
+ "grad_norm": 0.5734378099441528,
1148
+ "learning_rate": 0.000550788990825688,
1149
+ "loss": 3.8949,
1150
  "step": 7700
1151
  },
1152
  {
1153
+ "epoch": 0.8355795148247979,
1154
+ "grad_norm": 0.56657475233078,
1155
+ "learning_rate": 0.0005504651915812196,
1156
+ "loss": 3.878,
1157
  "step": 7750
1158
  },
1159
  {
1160
+ "epoch": 0.8409703504043127,
1161
+ "grad_norm": 0.6782810091972351,
1162
+ "learning_rate": 0.0005501413923367512,
1163
+ "loss": 3.9073,
1164
  "step": 7800
1165
  },
1166
  {
1167
+ "epoch": 0.8463611859838275,
1168
+ "grad_norm": 0.6757487654685974,
1169
+ "learning_rate": 0.0005498175930922827,
1170
+ "loss": 3.8846,
1171
  "step": 7850
1172
  },
1173
  {
1174
+ "epoch": 0.8517520215633423,
1175
+ "grad_norm": 0.5899352431297302,
1176
+ "learning_rate": 0.0005494937938478143,
1177
+ "loss": 3.8869,
1178
  "step": 7900
1179
  },
1180
  {
1181
+ "epoch": 0.8571428571428571,
1182
+ "grad_norm": 0.5763821005821228,
1183
+ "learning_rate": 0.0005491699946033459,
1184
+ "loss": 3.8827,
1185
  "step": 7950
1186
  },
1187
  {
1188
+ "epoch": 0.862533692722372,
1189
+ "grad_norm": 0.6570541858673096,
1190
+ "learning_rate": 0.0005488461953588775,
1191
  "loss": 3.8712,
1192
  "step": 8000
1193
  },
1194
  {
1195
+ "epoch": 0.862533692722372,
1196
+ "eval_accuracy": 0.3377064309797198,
1197
+ "eval_loss": 3.816032886505127,
1198
+ "eval_runtime": 183.5243,
1199
+ "eval_samples_per_second": 98.14,
1200
+ "eval_steps_per_second": 6.135,
1201
  "step": 8000
1202
  },
1203
  {
1204
+ "epoch": 0.8679245283018868,
1205
+ "grad_norm": 0.5602306723594666,
1206
+ "learning_rate": 0.000548522396114409,
1207
+ "loss": 3.8805,
1208
  "step": 8050
1209
  },
1210
  {
1211
+ "epoch": 0.8733153638814016,
1212
+ "grad_norm": 0.6562096476554871,
1213
+ "learning_rate": 0.0005481985968699406,
1214
+ "loss": 3.8669,
1215
  "step": 8100
1216
  },
1217
  {
1218
+ "epoch": 0.8787061994609164,
1219
+ "grad_norm": 0.5427864789962769,
1220
+ "learning_rate": 0.0005478747976254721,
1221
+ "loss": 3.8766,
1222
  "step": 8150
1223
  },
1224
  {
1225
+ "epoch": 0.8840970350404312,
1226
+ "grad_norm": 0.5421739816665649,
1227
+ "learning_rate": 0.0005475574743658931,
1228
+ "loss": 3.8668,
1229
  "step": 8200
1230
  },
1231
  {
1232
+ "epoch": 0.889487870619946,
1233
+ "grad_norm": 0.6107007265090942,
1234
+ "learning_rate": 0.0005472336751214246,
1235
+ "loss": 3.8544,
1236
  "step": 8250
1237
  },
1238
  {
1239
+ "epoch": 0.894878706199461,
1240
+ "grad_norm": 0.6804000735282898,
1241
+ "learning_rate": 0.0005469098758769562,
1242
+ "loss": 3.8797,
1243
  "step": 8300
1244
  },
1245
  {
1246
+ "epoch": 0.9002695417789758,
1247
+ "grad_norm": 0.6543049216270447,
1248
+ "learning_rate": 0.0005465860766324878,
1249
+ "loss": 3.861,
1250
  "step": 8350
1251
  },
1252
  {
1253
+ "epoch": 0.9056603773584906,
1254
+ "grad_norm": 0.7077866196632385,
1255
+ "learning_rate": 0.0005462622773880194,
1256
+ "loss": 3.874,
1257
  "step": 8400
1258
  },
1259
  {
1260
+ "epoch": 0.9110512129380054,
1261
+ "grad_norm": 0.5970845818519592,
1262
+ "learning_rate": 0.000545938478143551,
1263
+ "loss": 3.8836,
1264
  "step": 8450
1265
  },
1266
  {
1267
+ "epoch": 0.9164420485175202,
1268
+ "grad_norm": 0.5863073468208313,
1269
+ "learning_rate": 0.0005456146788990825,
1270
+ "loss": 3.8752,
1271
  "step": 8500
1272
  },
1273
  {
1274
+ "epoch": 0.921832884097035,
1275
+ "grad_norm": 0.5704134702682495,
1276
+ "learning_rate": 0.000545290879654614,
1277
+ "loss": 3.8553,
1278
  "step": 8550
1279
  },
1280
  {
1281
+ "epoch": 0.9272237196765498,
1282
+ "grad_norm": 0.5353178381919861,
1283
+ "learning_rate": 0.0005449670804101457,
1284
+ "loss": 3.8502,
1285
  "step": 8600
1286
  },
1287
  {
1288
+ "epoch": 0.9326145552560647,
1289
+ "grad_norm": 0.5742438435554504,
1290
+ "learning_rate": 0.0005446432811656773,
1291
+ "loss": 3.8445,
1292
  "step": 8650
1293
  },
1294
  {
1295
+ "epoch": 0.9380053908355795,
1296
+ "grad_norm": 0.6159672141075134,
1297
+ "learning_rate": 0.0005443194819212088,
1298
+ "loss": 3.85,
1299
  "step": 8700
1300
  },
1301
  {
1302
+ "epoch": 0.9433962264150944,
1303
+ "grad_norm": 0.5223457217216492,
1304
+ "learning_rate": 0.0005439956826767404,
1305
+ "loss": 3.8475,
1306
  "step": 8750
1307
  },
1308
  {
1309
+ "epoch": 0.9487870619946092,
1310
+ "grad_norm": 0.5645538568496704,
1311
+ "learning_rate": 0.000543671883432272,
1312
+ "loss": 3.8401,
1313
  "step": 8800
1314
  },
1315
  {
1316
+ "epoch": 0.954177897574124,
1317
+ "grad_norm": 0.5175207853317261,
1318
+ "learning_rate": 0.0005433480841878035,
1319
+ "loss": 3.8302,
1320
  "step": 8850
1321
  },
1322
  {
1323
+ "epoch": 0.9595687331536388,
1324
+ "grad_norm": 0.5572501420974731,
1325
+ "learning_rate": 0.000543024284943335,
1326
+ "loss": 3.8494,
1327
  "step": 8900
1328
  },
1329
  {
1330
+ "epoch": 0.9649595687331537,
1331
+ "grad_norm": 0.6185976266860962,
1332
+ "learning_rate": 0.0005427004856988667,
1333
+ "loss": 3.8528,
1334
  "step": 8950
1335
  },
1336
  {
1337
+ "epoch": 0.9703504043126685,
1338
+ "grad_norm": 0.5853222608566284,
1339
+ "learning_rate": 0.0005423766864543982,
1340
+ "loss": 3.8428,
1341
  "step": 9000
1342
  },
1343
  {
1344
+ "epoch": 0.9703504043126685,
1345
+ "eval_accuracy": 0.34173299834923665,
1346
+ "eval_loss": 3.77936053276062,
1347
+ "eval_runtime": 183.7681,
1348
+ "eval_samples_per_second": 98.009,
1349
+ "eval_steps_per_second": 6.127,
1350
  "step": 9000
1351
  },
1352
  {
1353
+ "epoch": 0.9757412398921833,
1354
+ "grad_norm": 0.5816961526870728,
1355
+ "learning_rate": 0.0005420528872099298,
1356
+ "loss": 3.8353,
1357
  "step": 9050
1358
  },
1359
  {
1360
+ "epoch": 0.9811320754716981,
1361
+ "grad_norm": 0.5886768102645874,
1362
+ "learning_rate": 0.0005417290879654613,
1363
+ "loss": 3.8505,
1364
  "step": 9100
1365
  },
1366
  {
1367
+ "epoch": 0.9865229110512129,
1368
+ "grad_norm": 0.5620021224021912,
1369
+ "learning_rate": 0.000541405288720993,
1370
+ "loss": 3.8231,
1371
  "step": 9150
1372
  },
1373
  {
1374
+ "epoch": 0.9919137466307277,
1375
+ "grad_norm": 0.5953831672668457,
1376
+ "learning_rate": 0.0005410814894765245,
1377
+ "loss": 3.8198,
1378
  "step": 9200
1379
  },
1380
  {
1381
+ "epoch": 0.9973045822102425,
1382
+ "grad_norm": 0.6049351096153259,
1383
+ "learning_rate": 0.0005407576902320561,
1384
+ "loss": 3.8327,
1385
  "step": 9250
1386
  },
1387
  {
1388
+ "epoch": 1.0026954177897573,
1389
+ "grad_norm": 0.5849935412406921,
1390
+ "learning_rate": 0.0005404338909875876,
1391
+ "loss": 3.8049,
1392
  "step": 9300
1393
  },
1394
  {
1395
+ "epoch": 1.0080862533692723,
1396
+ "grad_norm": 0.6319761276245117,
1397
+ "learning_rate": 0.0005401100917431192,
1398
+ "loss": 3.772,
1399
  "step": 9350
1400
  },
1401
  {
1402
+ "epoch": 1.013477088948787,
1403
+ "grad_norm": 0.599040687084198,
1404
+ "learning_rate": 0.0005397862924986508,
1405
+ "loss": 3.7662,
1406
  "step": 9400
1407
  },
1408
  {
1409
+ "epoch": 1.0188679245283019,
1410
+ "grad_norm": 0.6134844422340393,
1411
+ "learning_rate": 0.0005394624932541824,
1412
+ "loss": 3.7678,
1413
  "step": 9450
1414
  },
1415
  {
1416
+ "epoch": 1.0242587601078168,
1417
+ "grad_norm": 0.6510705351829529,
1418
+ "learning_rate": 0.0005391386940097139,
1419
+ "loss": 3.7646,
1420
  "step": 9500
1421
  },
1422
  {
1423
+ "epoch": 1.0296495956873315,
1424
+ "grad_norm": 0.5537397861480713,
1425
+ "learning_rate": 0.0005388148947652455,
1426
+ "loss": 3.7639,
1427
  "step": 9550
1428
  },
1429
  {
1430
+ "epoch": 1.0350404312668464,
1431
+ "grad_norm": 0.6508437395095825,
1432
+ "learning_rate": 0.000538491095520777,
1433
+ "loss": 3.753,
1434
  "step": 9600
1435
  },
1436
  {
1437
+ "epoch": 1.0404312668463611,
1438
+ "grad_norm": 0.6333539485931396,
1439
+ "learning_rate": 0.0005381672962763086,
1440
+ "loss": 3.7944,
1441
  "step": 9650
1442
  },
1443
  {
1444
+ "epoch": 1.045822102425876,
1445
+ "grad_norm": 0.6207062005996704,
1446
+ "learning_rate": 0.0005378434970318403,
1447
+ "loss": 3.7775,
1448
  "step": 9700
1449
  },
1450
  {
1451
+ "epoch": 1.0512129380053907,
1452
+ "grad_norm": 0.5319339632987976,
1453
+ "learning_rate": 0.0005375196977873718,
1454
+ "loss": 3.779,
1455
  "step": 9750
1456
  },
1457
  {
1458
+ "epoch": 1.0566037735849056,
1459
+ "grad_norm": 0.586555540561676,
1460
+ "learning_rate": 0.0005371958985429034,
1461
+ "loss": 3.7858,
1462
  "step": 9800
1463
  },
1464
  {
1465
+ "epoch": 1.0619946091644206,
1466
+ "grad_norm": 0.5690563321113586,
1467
+ "learning_rate": 0.0005368720992984349,
1468
+ "loss": 3.7811,
1469
  "step": 9850
1470
  },
1471
  {
1472
+ "epoch": 1.0673854447439353,
1473
+ "grad_norm": 0.5293869972229004,
1474
+ "learning_rate": 0.0005365483000539665,
1475
+ "loss": 3.7734,
1476
  "step": 9900
1477
  },
1478
  {
1479
+ "epoch": 1.0727762803234502,
1480
+ "grad_norm": 0.5877575278282166,
1481
+ "learning_rate": 0.0005362245008094981,
1482
+ "loss": 3.7691,
1483
  "step": 9950
1484
  },
1485
  {
1486
+ "epoch": 1.0781671159029649,
1487
+ "grad_norm": 0.6360712051391602,
1488
+ "learning_rate": 0.0005359007015650297,
1489
+ "loss": 3.759,
1490
  "step": 10000
1491
  },
1492
  {
1493
+ "epoch": 1.0781671159029649,
1494
+ "eval_accuracy": 0.34497118036477165,
1495
+ "eval_loss": 3.746513843536377,
1496
+ "eval_runtime": 183.499,
1497
+ "eval_samples_per_second": 98.153,
1498
+ "eval_steps_per_second": 6.136,
1499
  "step": 10000
1500
  }
1501
  ],
1502
  "logging_steps": 50,
1503
+ "max_steps": 92750,
1504
  "num_input_tokens_seen": 0,
1505
  "num_train_epochs": 10,
1506
  "save_steps": 10000,
 
1516
  "attributes": {}
1517
  }
1518
  },
1519
+ "total_flos": 8.36069179392e+16,
1520
  "train_batch_size": 32,
1521
  "trial_name": null,
1522
  "trial_params": null
checkpoint-10000/training_args.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:a7e9ad18c4133cca362e8ca729cda30cbeb2c0db779f81e5371625847996c992
3
  size 5304
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c424126a3e03dc0994f6fc899ff4f1ff7015775f3423397f5c1c16e9ebf50480
3
  size 5304