craa commited on
Commit
f40ab24
·
verified ·
1 Parent(s): f5a4044

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:f49a59670c717e56d6a4311ed243bbd975926e8990d9e46a3f5234b3c409750c
3
  size 503128704
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:18bcd1c1737518486f8d3c4c8949b8d131644b4eca28f7962a7aeca7a70338a7
3
  size 503128704
checkpoint-10000/optimizer.pt CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:317ce9079bf77ed5e4d3e2644b763fed8157cb653f508847c9507ff162ebe752
3
  size 1006351290
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:abadb84a828ae1cde4669f147e6c74709632b3d334ead5406ea61de144dab15f
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:d4f8db97fd2d786dd18e1660a629749563cc5b59d400bd62a512ac18ce773e2a
3
  size 14244
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:95d415f5a2495efee5f0bd162340bac975693e203cab718136e9d616a832b275
3
  size 14244
checkpoint-10000/scheduler.pt CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:e5a954fd0891b9a63bb5018718359621b31d6887975b43dbbefda70f75ac15a5
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": 3.7542877197265625,
3
- "best_model_checkpoint": "/scratch/cl5625/exceptions/models/100M_634/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": 0.947643518447876,
14
- "learning_rate": 0.00028199999999999997,
15
- "loss": 9.02,
16
  "step": 50
17
  },
18
  {
19
- "epoch": 0.010763104079216447,
20
- "grad_norm": 2.6564507484436035,
21
- "learning_rate": 0.0005819999999999999,
22
- "loss": 6.9671,
23
  "step": 100
24
  },
25
  {
26
- "epoch": 0.01614465611882467,
27
- "grad_norm": 2.533383369445801,
28
- "learning_rate": 0.0005996961534317422,
29
- "loss": 6.5213,
30
  "step": 150
31
  },
32
  {
33
- "epoch": 0.021526208158432893,
34
- "grad_norm": 3.2595882415771484,
35
- "learning_rate": 0.0005993729124016807,
36
- "loss": 6.2948,
37
  "step": 200
38
  },
39
  {
40
- "epoch": 0.026907760198041114,
41
- "grad_norm": 1.9747371673583984,
42
- "learning_rate": 0.0005990496713716194,
43
- "loss": 6.1227,
44
  "step": 250
45
  },
46
  {
47
- "epoch": 0.03228931223764934,
48
- "grad_norm": 1.3486372232437134,
49
- "learning_rate": 0.000598726430341558,
50
- "loss": 5.989,
51
  "step": 300
52
  },
53
  {
54
- "epoch": 0.03767086427725756,
55
- "grad_norm": 2.030965805053711,
56
- "learning_rate": 0.0005984031893114966,
57
- "loss": 5.8853,
58
  "step": 350
59
  },
60
  {
61
- "epoch": 0.04305241631686579,
62
- "grad_norm": 1.8007402420043945,
63
- "learning_rate": 0.0005980799482814351,
64
- "loss": 5.8246,
65
  "step": 400
66
  },
67
  {
68
- "epoch": 0.048433968356474004,
69
- "grad_norm": 1.0434612035751343,
70
- "learning_rate": 0.0005977567072513737,
71
- "loss": 5.7759,
72
  "step": 450
73
  },
74
  {
75
- "epoch": 0.05381552039608223,
76
- "grad_norm": 1.30581796169281,
77
- "learning_rate": 0.0005974334662213123,
78
- "loss": 5.6783,
79
  "step": 500
80
  },
81
  {
82
- "epoch": 0.05919707243569045,
83
- "grad_norm": 1.3575013875961304,
84
- "learning_rate": 0.0005971102251912509,
85
- "loss": 5.6197,
86
  "step": 550
87
  },
88
  {
89
- "epoch": 0.06457862447529868,
90
- "grad_norm": 1.027186632156372,
91
- "learning_rate": 0.0005967869841611895,
92
- "loss": 5.5417,
93
  "step": 600
94
  },
95
  {
96
- "epoch": 0.0699601765149069,
97
- "grad_norm": 1.0279064178466797,
98
- "learning_rate": 0.000596463743131128,
99
- "loss": 5.4727,
100
  "step": 650
101
  },
102
  {
103
- "epoch": 0.07534172855451512,
104
- "grad_norm": 1.1722652912139893,
105
- "learning_rate": 0.0005961405021010667,
106
- "loss": 5.418,
107
  "step": 700
108
  },
109
  {
110
- "epoch": 0.08072328059412334,
111
- "grad_norm": 0.9741182327270508,
112
- "learning_rate": 0.0005958172610710052,
113
- "loss": 5.3362,
114
  "step": 750
115
  },
116
  {
117
- "epoch": 0.08610483263373157,
118
- "grad_norm": 1.3713257312774658,
119
- "learning_rate": 0.0005954940200409439,
120
- "loss": 5.2961,
121
  "step": 800
122
  },
123
  {
124
- "epoch": 0.09148638467333979,
125
- "grad_norm": 1.626237392425537,
126
- "learning_rate": 0.0005951707790108824,
127
- "loss": 5.2518,
128
  "step": 850
129
  },
130
  {
131
- "epoch": 0.09686793671294801,
132
- "grad_norm": 1.4366179704666138,
133
- "learning_rate": 0.0005948475379808209,
134
- "loss": 5.2072,
135
  "step": 900
136
  },
137
  {
138
- "epoch": 0.10224948875255624,
139
- "grad_norm": 1.0901954174041748,
140
- "learning_rate": 0.0005945242969507596,
141
- "loss": 5.1598,
142
  "step": 950
143
  },
144
  {
145
- "epoch": 0.10763104079216446,
146
- "grad_norm": 1.056020975112915,
147
- "learning_rate": 0.0005942010559206981,
148
- "loss": 5.1238,
149
  "step": 1000
150
  },
151
  {
152
- "epoch": 0.10763104079216446,
153
- "eval_accuracy": 0.22448784016638668,
154
- "eval_loss": 5.047372341156006,
155
- "eval_runtime": 186.0066,
156
- "eval_samples_per_second": 96.83,
157
- "eval_steps_per_second": 6.054,
158
  "step": 1000
159
  },
160
  {
161
- "epoch": 0.11301259283177269,
162
- "grad_norm": 1.2061820030212402,
163
- "learning_rate": 0.0005938778148906367,
164
- "loss": 5.0644,
165
  "step": 1050
166
  },
167
  {
168
- "epoch": 0.1183941448713809,
169
- "grad_norm": 0.8459726572036743,
170
- "learning_rate": 0.0005935545738605753,
171
- "loss": 5.0343,
172
  "step": 1100
173
  },
174
  {
175
- "epoch": 0.12377569691098914,
176
- "grad_norm": 1.117684006690979,
177
- "learning_rate": 0.000593231332830514,
178
- "loss": 4.9795,
179
  "step": 1150
180
  },
181
  {
182
- "epoch": 0.12915724895059735,
183
- "grad_norm": 1.4578666687011719,
184
- "learning_rate": 0.0005929080918004525,
185
- "loss": 4.9842,
186
  "step": 1200
187
  },
188
  {
189
- "epoch": 0.13453880099020557,
190
- "grad_norm": 0.9818866848945618,
191
- "learning_rate": 0.000592584850770391,
192
- "loss": 4.9541,
193
  "step": 1250
194
  },
195
  {
196
- "epoch": 0.1399203530298138,
197
- "grad_norm": 0.8982384204864502,
198
- "learning_rate": 0.0005922616097403296,
199
- "loss": 4.9085,
200
  "step": 1300
201
  },
202
  {
203
- "epoch": 0.14530190506942203,
204
- "grad_norm": 1.4174669981002808,
205
- "learning_rate": 0.0005919383687102682,
206
- "loss": 4.8986,
207
  "step": 1350
208
  },
209
  {
210
- "epoch": 0.15068345710903025,
211
- "grad_norm": 0.9739001393318176,
212
- "learning_rate": 0.0005916151276802069,
213
- "loss": 4.873,
214
  "step": 1400
215
  },
216
  {
217
- "epoch": 0.15606500914863847,
218
- "grad_norm": 1.062203049659729,
219
- "learning_rate": 0.0005912918866501454,
220
- "loss": 4.8505,
221
  "step": 1450
222
  },
223
  {
224
- "epoch": 0.16144656118824668,
225
- "grad_norm": 1.0389671325683594,
226
- "learning_rate": 0.000590968645620084,
227
- "loss": 4.8168,
228
  "step": 1500
229
  },
230
  {
231
- "epoch": 0.1668281132278549,
232
- "grad_norm": 1.0054130554199219,
233
- "learning_rate": 0.0005906454045900226,
234
- "loss": 4.8127,
235
  "step": 1550
236
  },
237
  {
238
- "epoch": 0.17220966526746315,
239
- "grad_norm": 1.2191665172576904,
240
- "learning_rate": 0.0005903221635599612,
241
- "loss": 4.7919,
242
  "step": 1600
243
  },
244
  {
245
- "epoch": 0.17759121730707136,
246
- "grad_norm": 0.8363758325576782,
247
- "learning_rate": 0.0005899989225298998,
248
- "loss": 4.7589,
249
  "step": 1650
250
  },
251
  {
252
- "epoch": 0.18297276934667958,
253
- "grad_norm": 1.000256061553955,
254
- "learning_rate": 0.0005896756814998383,
255
- "loss": 4.7222,
256
  "step": 1700
257
  },
258
  {
259
- "epoch": 0.1883543213862878,
260
- "grad_norm": 1.0218900442123413,
261
- "learning_rate": 0.0005893524404697769,
262
- "loss": 4.7129,
263
  "step": 1750
264
  },
265
  {
266
- "epoch": 0.19373587342589602,
267
- "grad_norm": 0.8401052951812744,
268
- "learning_rate": 0.0005890291994397155,
269
- "loss": 4.6975,
270
  "step": 1800
271
  },
272
  {
273
- "epoch": 0.19911742546550426,
274
- "grad_norm": 1.0717366933822632,
275
- "learning_rate": 0.0005887059584096541,
276
- "loss": 4.6681,
277
  "step": 1850
278
  },
279
  {
280
- "epoch": 0.20449897750511248,
281
- "grad_norm": 0.8167716264724731,
282
- "learning_rate": 0.0005883827173795926,
283
- "loss": 4.6492,
284
  "step": 1900
285
  },
286
  {
287
- "epoch": 0.2098805295447207,
288
- "grad_norm": 0.7771570682525635,
289
- "learning_rate": 0.0005880594763495313,
290
- "loss": 4.6232,
291
  "step": 1950
292
  },
293
  {
294
- "epoch": 0.2152620815843289,
295
- "grad_norm": 1.1614071130752563,
296
- "learning_rate": 0.0005877362353194698,
297
- "loss": 4.6101,
298
  "step": 2000
299
  },
300
  {
301
- "epoch": 0.2152620815843289,
302
- "eval_accuracy": 0.2659886798902302,
303
- "eval_loss": 4.5428266525268555,
304
- "eval_runtime": 184.7289,
305
- "eval_samples_per_second": 97.5,
306
- "eval_steps_per_second": 6.095,
307
  "step": 2000
308
  },
309
  {
310
- "epoch": 0.22064363362393713,
311
- "grad_norm": 0.8067775368690491,
312
- "learning_rate": 0.0005874129942894084,
313
- "loss": 4.5945,
314
  "step": 2050
315
  },
316
  {
317
- "epoch": 0.22602518566354537,
318
- "grad_norm": 1.3052033185958862,
319
- "learning_rate": 0.000587089753259347,
320
- "loss": 4.576,
321
  "step": 2100
322
  },
323
  {
324
- "epoch": 0.2314067377031536,
325
- "grad_norm": 1.0830671787261963,
326
- "learning_rate": 0.0005867665122292855,
327
- "loss": 4.573,
328
  "step": 2150
329
  },
330
  {
331
- "epoch": 0.2367882897427618,
332
- "grad_norm": 1.00840425491333,
333
- "learning_rate": 0.0005864432711992242,
334
- "loss": 4.5518,
335
  "step": 2200
336
  },
337
  {
338
- "epoch": 0.24216984178237003,
339
- "grad_norm": 1.4997429847717285,
340
- "learning_rate": 0.0005861200301691628,
341
- "loss": 4.5326,
342
  "step": 2250
343
  },
344
  {
345
- "epoch": 0.24755139382197827,
346
- "grad_norm": 0.8558619618415833,
347
- "learning_rate": 0.0005857967891391014,
348
- "loss": 4.4986,
349
  "step": 2300
350
  },
351
  {
352
- "epoch": 0.2529329458615865,
353
- "grad_norm": 0.8113582134246826,
354
- "learning_rate": 0.0005854735481090399,
355
- "loss": 4.5091,
356
  "step": 2350
357
  },
358
  {
359
- "epoch": 0.2583144979011947,
360
- "grad_norm": 0.98548823595047,
361
- "learning_rate": 0.0005851503070789784,
362
- "loss": 4.4801,
363
  "step": 2400
364
  },
365
  {
366
- "epoch": 0.2636960499408029,
367
- "grad_norm": 0.8301020860671997,
368
- "learning_rate": 0.0005848270660489171,
369
- "loss": 4.4595,
370
  "step": 2450
371
  },
372
  {
373
- "epoch": 0.26907760198041114,
374
- "grad_norm": 1.0569989681243896,
375
- "learning_rate": 0.0005845038250188556,
376
- "loss": 4.4474,
377
  "step": 2500
378
  },
379
  {
380
- "epoch": 0.27445915402001936,
381
- "grad_norm": 0.9343509674072266,
382
- "learning_rate": 0.0005841805839887943,
383
- "loss": 4.4274,
384
  "step": 2550
385
  },
386
  {
387
- "epoch": 0.2798407060596276,
388
- "grad_norm": 0.8506145477294922,
389
- "learning_rate": 0.0005838573429587328,
390
- "loss": 4.4131,
391
  "step": 2600
392
  },
393
  {
394
- "epoch": 0.2852222580992358,
395
- "grad_norm": 0.9148919582366943,
396
- "learning_rate": 0.0005835341019286715,
397
- "loss": 4.3894,
398
  "step": 2650
399
  },
400
  {
401
- "epoch": 0.29060381013884407,
402
- "grad_norm": 0.8359649777412415,
403
- "learning_rate": 0.00058321086089861,
404
- "loss": 4.4011,
405
  "step": 2700
406
  },
407
  {
408
- "epoch": 0.2959853621784523,
409
- "grad_norm": 0.9812710881233215,
410
- "learning_rate": 0.0005828876198685486,
411
- "loss": 4.3753,
412
  "step": 2750
413
  },
414
  {
415
- "epoch": 0.3013669142180605,
416
- "grad_norm": 1.0767602920532227,
417
- "learning_rate": 0.0005825643788384872,
418
- "loss": 4.374,
419
  "step": 2800
420
  },
421
  {
422
- "epoch": 0.3067484662576687,
423
- "grad_norm": 0.8539724349975586,
424
- "learning_rate": 0.0005822411378084257,
425
- "loss": 4.3591,
426
  "step": 2850
427
  },
428
  {
429
- "epoch": 0.31213001829727693,
430
- "grad_norm": 0.8400486707687378,
431
- "learning_rate": 0.0005819178967783644,
432
- "loss": 4.3441,
433
  "step": 2900
434
  },
435
  {
436
- "epoch": 0.31751157033688515,
437
- "grad_norm": 0.9034419059753418,
438
- "learning_rate": 0.0005815946557483029,
439
- "loss": 4.3381,
440
  "step": 2950
441
  },
442
  {
443
- "epoch": 0.32289312237649337,
444
- "grad_norm": 0.8509873151779175,
445
- "learning_rate": 0.0005812714147182415,
446
- "loss": 4.327,
447
  "step": 3000
448
  },
449
  {
450
- "epoch": 0.32289312237649337,
451
- "eval_accuracy": 0.2964122490484995,
452
- "eval_loss": 4.251293659210205,
453
- "eval_runtime": 184.751,
454
- "eval_samples_per_second": 97.488,
455
- "eval_steps_per_second": 6.095,
456
  "step": 3000
457
  },
458
  {
459
- "epoch": 0.3282746744161016,
460
- "grad_norm": 0.8212876915931702,
461
- "learning_rate": 0.0005809481736881801,
462
- "loss": 4.3028,
463
  "step": 3050
464
  },
465
  {
466
- "epoch": 0.3336562264557098,
467
- "grad_norm": 0.7386678457260132,
468
- "learning_rate": 0.0005806249326581187,
469
- "loss": 4.3224,
470
  "step": 3100
471
  },
472
  {
473
- "epoch": 0.3390377784953181,
474
- "grad_norm": 0.8408326506614685,
475
- "learning_rate": 0.0005803016916280573,
476
- "loss": 4.3054,
477
  "step": 3150
478
  },
479
  {
480
- "epoch": 0.3444193305349263,
481
- "grad_norm": 0.7891979217529297,
482
- "learning_rate": 0.0005799784505979959,
483
- "loss": 4.2832,
484
  "step": 3200
485
  },
486
  {
487
- "epoch": 0.3498008825745345,
488
- "grad_norm": 1.1157724857330322,
489
- "learning_rate": 0.0005796552095679344,
490
- "loss": 4.2875,
491
  "step": 3250
492
  },
493
  {
494
- "epoch": 0.35518243461414273,
495
- "grad_norm": 0.7911354899406433,
496
- "learning_rate": 0.000579331968537873,
497
- "loss": 4.2468,
498
  "step": 3300
499
  },
500
  {
501
- "epoch": 0.36056398665375095,
502
- "grad_norm": 0.670813798904419,
503
- "learning_rate": 0.0005790087275078116,
504
- "loss": 4.2655,
505
  "step": 3350
506
  },
507
  {
508
- "epoch": 0.36594553869335916,
509
- "grad_norm": 0.7761114835739136,
510
- "learning_rate": 0.0005786854864777502,
511
- "loss": 4.2688,
512
  "step": 3400
513
  },
514
  {
515
- "epoch": 0.3713270907329674,
516
- "grad_norm": 0.8588844537734985,
517
- "learning_rate": 0.0005783622454476888,
518
- "loss": 4.2547,
519
  "step": 3450
520
  },
521
  {
522
- "epoch": 0.3767086427725756,
523
- "grad_norm": 0.8100800514221191,
524
- "learning_rate": 0.0005780390044176273,
525
- "loss": 4.2286,
526
  "step": 3500
527
  },
528
  {
529
- "epoch": 0.3820901948121838,
530
- "grad_norm": 0.8621886372566223,
531
- "learning_rate": 0.000577715763387566,
532
- "loss": 4.2537,
533
  "step": 3550
534
  },
535
  {
536
- "epoch": 0.38747174685179203,
537
- "grad_norm": 0.681401252746582,
538
- "learning_rate": 0.0005773925223575045,
539
- "loss": 4.2342,
540
  "step": 3600
541
  },
542
  {
543
- "epoch": 0.3928532988914003,
544
- "grad_norm": 0.8861917853355408,
545
- "learning_rate": 0.0005770692813274432,
546
- "loss": 4.2298,
547
  "step": 3650
548
  },
549
  {
550
- "epoch": 0.3982348509310085,
551
- "grad_norm": 0.7586415410041809,
552
- "learning_rate": 0.0005767460402973817,
553
- "loss": 4.1935,
554
  "step": 3700
555
  },
556
  {
557
- "epoch": 0.40361640297061674,
558
- "grad_norm": 0.7726384401321411,
559
- "learning_rate": 0.0005764227992673203,
560
- "loss": 4.1981,
561
  "step": 3750
562
  },
563
  {
564
- "epoch": 0.40899795501022496,
565
- "grad_norm": 0.8231189846992493,
566
- "learning_rate": 0.0005760995582372589,
567
- "loss": 4.1916,
568
  "step": 3800
569
  },
570
  {
571
- "epoch": 0.4143795070498332,
572
- "grad_norm": 0.6635996103286743,
573
- "learning_rate": 0.0005757763172071974,
574
- "loss": 4.194,
575
  "step": 3850
576
  },
577
  {
578
- "epoch": 0.4197610590894414,
579
- "grad_norm": 0.6655986905097961,
580
- "learning_rate": 0.000575453076177136,
581
- "loss": 4.2108,
582
  "step": 3900
583
  },
584
  {
585
- "epoch": 0.4251426111290496,
586
- "grad_norm": 0.6464502215385437,
587
- "learning_rate": 0.0005751298351470746,
588
- "loss": 4.1773,
589
  "step": 3950
590
  },
591
  {
592
- "epoch": 0.4305241631686578,
593
- "grad_norm": 1.0790879726409912,
594
- "learning_rate": 0.0005748065941170133,
595
- "loss": 4.18,
596
  "step": 4000
597
  },
598
  {
599
- "epoch": 0.4305241631686578,
600
- "eval_accuracy": 0.3111644862386228,
601
- "eval_loss": 4.104041576385498,
602
- "eval_runtime": 184.532,
603
- "eval_samples_per_second": 97.604,
604
- "eval_steps_per_second": 6.102,
605
  "step": 4000
606
  },
607
  {
608
- "epoch": 0.43590571520826604,
609
- "grad_norm": 0.7633262872695923,
610
- "learning_rate": 0.0005744833530869518,
611
- "loss": 4.1895,
612
  "step": 4050
613
  },
614
  {
615
- "epoch": 0.44128726724787426,
616
- "grad_norm": 0.7926012277603149,
617
- "learning_rate": 0.0005741601120568903,
618
- "loss": 4.1728,
619
  "step": 4100
620
  },
621
  {
622
- "epoch": 0.44666881928748253,
623
- "grad_norm": 0.6977135539054871,
624
- "learning_rate": 0.0005738368710268289,
625
- "loss": 4.1502,
626
  "step": 4150
627
  },
628
  {
629
- "epoch": 0.45205037132709075,
630
- "grad_norm": 0.7949364185333252,
631
- "learning_rate": 0.0005735136299967675,
632
- "loss": 4.181,
633
  "step": 4200
634
  },
635
  {
636
- "epoch": 0.45743192336669897,
637
- "grad_norm": 0.6541943550109863,
638
- "learning_rate": 0.0005731903889667062,
639
- "loss": 4.1484,
640
  "step": 4250
641
  },
642
  {
643
- "epoch": 0.4628134754063072,
644
- "grad_norm": 0.7301810383796692,
645
- "learning_rate": 0.0005728671479366447,
646
- "loss": 4.1502,
647
  "step": 4300
648
  },
649
  {
650
- "epoch": 0.4681950274459154,
651
- "grad_norm": 0.7047900557518005,
652
- "learning_rate": 0.0005725439069065833,
653
- "loss": 4.1516,
654
  "step": 4350
655
  },
656
  {
657
- "epoch": 0.4735765794855236,
658
- "grad_norm": 0.6980259418487549,
659
- "learning_rate": 0.0005722206658765219,
660
- "loss": 4.1214,
661
  "step": 4400
662
  },
663
  {
664
- "epoch": 0.47895813152513184,
665
- "grad_norm": 0.7646435499191284,
666
- "learning_rate": 0.0005718974248464605,
667
- "loss": 4.1348,
668
  "step": 4450
669
  },
670
  {
671
- "epoch": 0.48433968356474005,
672
- "grad_norm": 0.7655755877494812,
673
- "learning_rate": 0.000571574183816399,
674
- "loss": 4.1344,
675
  "step": 4500
676
  },
677
  {
678
- "epoch": 0.48972123560434827,
679
- "grad_norm": 0.7866830825805664,
680
- "learning_rate": 0.0005712509427863376,
681
- "loss": 4.1116,
682
  "step": 4550
683
  },
684
  {
685
- "epoch": 0.49510278764395654,
686
- "grad_norm": 0.7541496753692627,
687
- "learning_rate": 0.0005709277017562762,
688
- "loss": 4.0941,
689
  "step": 4600
690
  },
691
  {
692
- "epoch": 0.5004843396835648,
693
- "grad_norm": 0.7112383842468262,
694
- "learning_rate": 0.0005706044607262148,
695
- "loss": 4.0955,
696
  "step": 4650
697
  },
698
  {
699
- "epoch": 0.505865891723173,
700
- "grad_norm": 0.7648486495018005,
701
- "learning_rate": 0.0005702812196961534,
702
- "loss": 4.1013,
703
  "step": 4700
704
  },
705
  {
706
- "epoch": 0.5112474437627812,
707
- "grad_norm": 0.6658865213394165,
708
- "learning_rate": 0.0005699579786660919,
709
- "loss": 4.0833,
710
  "step": 4750
711
  },
712
  {
713
- "epoch": 0.5166289958023894,
714
- "grad_norm": 0.6460030674934387,
715
- "learning_rate": 0.0005696347376360306,
716
- "loss": 4.0976,
717
  "step": 4800
718
  },
719
  {
720
- "epoch": 0.5220105478419976,
721
- "grad_norm": 0.7306205630302429,
722
- "learning_rate": 0.0005693114966059691,
723
- "loss": 4.0811,
724
  "step": 4850
725
  },
726
  {
727
- "epoch": 0.5273920998816058,
728
- "grad_norm": 0.6842286586761475,
729
- "learning_rate": 0.0005689882555759077,
730
- "loss": 4.0634,
731
  "step": 4900
732
  },
733
  {
734
- "epoch": 0.5327736519212141,
735
- "grad_norm": 0.8138746619224548,
736
- "learning_rate": 0.0005686650145458463,
737
- "loss": 4.0802,
738
  "step": 4950
739
  },
740
  {
741
- "epoch": 0.5381552039608223,
742
- "grad_norm": 0.6573743224143982,
743
- "learning_rate": 0.0005683417735157848,
744
- "loss": 4.0788,
745
  "step": 5000
746
  },
747
  {
748
- "epoch": 0.5381552039608223,
749
- "eval_accuracy": 0.3209131492920015,
750
- "eval_loss": 4.001318454742432,
751
- "eval_runtime": 184.7779,
752
- "eval_samples_per_second": 97.474,
753
- "eval_steps_per_second": 6.094,
754
  "step": 5000
755
  },
756
  {
757
- "epoch": 0.5435367560004305,
758
- "grad_norm": 0.621402382850647,
759
- "learning_rate": 0.0005680185324857235,
760
- "loss": 4.0548,
761
  "step": 5050
762
  },
763
  {
764
- "epoch": 0.5489183080400387,
765
- "grad_norm": 0.8887001872062683,
766
- "learning_rate": 0.0005676952914556621,
767
- "loss": 4.0738,
768
  "step": 5100
769
  },
770
  {
771
- "epoch": 0.5542998600796469,
772
- "grad_norm": 0.7014524936676025,
773
- "learning_rate": 0.0005673720504256007,
774
- "loss": 4.0823,
775
  "step": 5150
776
  },
777
  {
778
- "epoch": 0.5596814121192552,
779
- "grad_norm": 0.7665441036224365,
780
- "learning_rate": 0.0005670488093955392,
781
- "loss": 4.0523,
782
  "step": 5200
783
  },
784
  {
785
- "epoch": 0.5650629641588634,
786
- "grad_norm": 0.6104534864425659,
787
- "learning_rate": 0.0005667255683654777,
788
- "loss": 4.0362,
789
  "step": 5250
790
  },
791
  {
792
- "epoch": 0.5704445161984716,
793
- "grad_norm": 0.6866149306297302,
794
- "learning_rate": 0.0005664023273354164,
795
- "loss": 4.0581,
796
  "step": 5300
797
  },
798
  {
799
- "epoch": 0.5758260682380799,
800
- "grad_norm": 0.6914085745811462,
801
- "learning_rate": 0.0005660790863053549,
802
- "loss": 4.0439,
803
  "step": 5350
804
  },
805
  {
806
- "epoch": 0.5812076202776881,
807
- "grad_norm": 0.604280948638916,
808
- "learning_rate": 0.0005657558452752936,
809
- "loss": 4.0293,
810
  "step": 5400
811
  },
812
  {
813
- "epoch": 0.5865891723172963,
814
- "grad_norm": 0.8009333610534668,
815
- "learning_rate": 0.0005654326042452321,
816
- "loss": 4.0299,
817
  "step": 5450
818
  },
819
  {
820
- "epoch": 0.5919707243569046,
821
- "grad_norm": 0.7749080061912537,
822
- "learning_rate": 0.0005651093632151708,
823
- "loss": 4.0229,
824
  "step": 5500
825
  },
826
  {
827
- "epoch": 0.5973522763965128,
828
- "grad_norm": 0.7085071802139282,
829
- "learning_rate": 0.0005647861221851093,
830
- "loss": 4.0196,
831
  "step": 5550
832
  },
833
  {
834
- "epoch": 0.602733828436121,
835
- "grad_norm": 0.6664130687713623,
836
- "learning_rate": 0.0005644628811550479,
837
- "loss": 4.0094,
838
  "step": 5600
839
  },
840
  {
841
- "epoch": 0.6081153804757292,
842
- "grad_norm": 0.7595982551574707,
843
- "learning_rate": 0.0005641396401249865,
844
- "loss": 4.0446,
845
  "step": 5650
846
  },
847
  {
848
- "epoch": 0.6134969325153374,
849
- "grad_norm": 0.6277487874031067,
850
- "learning_rate": 0.000563816399094925,
851
- "loss": 4.0226,
852
  "step": 5700
853
  },
854
  {
855
- "epoch": 0.6188784845549457,
856
- "grad_norm": 0.5675387978553772,
857
- "learning_rate": 0.0005634931580648637,
858
- "loss": 4.0005,
859
  "step": 5750
860
  },
861
  {
862
- "epoch": 0.6242600365945539,
863
- "grad_norm": 0.6105149388313293,
864
- "learning_rate": 0.0005631699170348022,
865
- "loss": 4.0048,
866
  "step": 5800
867
  },
868
  {
869
- "epoch": 0.6296415886341621,
870
- "grad_norm": 0.822024941444397,
871
- "learning_rate": 0.0005628466760047408,
872
- "loss": 4.0139,
873
  "step": 5850
874
  },
875
  {
876
- "epoch": 0.6350231406737703,
877
- "grad_norm": 0.5809662342071533,
878
- "learning_rate": 0.0005625234349746794,
879
- "loss": 3.9919,
880
  "step": 5900
881
  },
882
  {
883
- "epoch": 0.6404046927133785,
884
- "grad_norm": 0.6869716048240662,
885
- "learning_rate": 0.000562200193944618,
886
- "loss": 4.0053,
887
  "step": 5950
888
  },
889
  {
890
- "epoch": 0.6457862447529867,
891
- "grad_norm": 0.7093315720558167,
892
- "learning_rate": 0.0005618769529145566,
893
- "loss": 3.9809,
894
  "step": 6000
895
  },
896
  {
897
- "epoch": 0.6457862447529867,
898
- "eval_accuracy": 0.3273809297449341,
899
- "eval_loss": 3.9289309978485107,
900
- "eval_runtime": 184.1772,
901
- "eval_samples_per_second": 97.792,
902
- "eval_steps_per_second": 6.114,
903
  "step": 6000
904
  },
905
  {
906
- "epoch": 0.651167796792595,
907
- "grad_norm": 0.5664177536964417,
908
- "learning_rate": 0.0005615537118844952,
909
- "loss": 4.0017,
910
  "step": 6050
911
  },
912
  {
913
- "epoch": 0.6565493488322032,
914
- "grad_norm": 0.6920751929283142,
915
- "learning_rate": 0.0005612304708544337,
916
- "loss": 3.9825,
917
  "step": 6100
918
  },
919
  {
920
- "epoch": 0.6619309008718114,
921
- "grad_norm": 0.5844790935516357,
922
- "learning_rate": 0.0005609072298243723,
923
- "loss": 4.0016,
924
  "step": 6150
925
  },
926
  {
927
- "epoch": 0.6673124529114196,
928
- "grad_norm": 0.6675333976745605,
929
- "learning_rate": 0.0005605839887943109,
930
- "loss": 3.9938,
931
  "step": 6200
932
  },
933
  {
934
- "epoch": 0.6726940049510278,
935
- "grad_norm": 0.6434072256088257,
936
- "learning_rate": 0.0005602607477642495,
937
- "loss": 3.9887,
938
  "step": 6250
939
  },
940
  {
941
- "epoch": 0.6780755569906362,
942
- "grad_norm": 0.660064160823822,
943
- "learning_rate": 0.0005599375067341881,
944
- "loss": 3.9893,
945
  "step": 6300
946
  },
947
  {
948
- "epoch": 0.6834571090302444,
949
- "grad_norm": 0.7599339485168457,
950
- "learning_rate": 0.0005596142657041266,
951
- "loss": 3.969,
952
  "step": 6350
953
  },
954
  {
955
- "epoch": 0.6888386610698526,
956
- "grad_norm": 0.675517201423645,
957
- "learning_rate": 0.0005592910246740653,
958
- "loss": 3.9902,
959
  "step": 6400
960
  },
961
  {
962
- "epoch": 0.6942202131094608,
963
- "grad_norm": 0.7125230431556702,
964
- "learning_rate": 0.0005589677836440038,
965
- "loss": 3.9791,
966
  "step": 6450
967
  },
968
  {
969
- "epoch": 0.699601765149069,
970
- "grad_norm": 0.6018815636634827,
971
- "learning_rate": 0.0005586445426139425,
972
- "loss": 3.9474,
973
  "step": 6500
974
  },
975
  {
976
- "epoch": 0.7049833171886772,
977
- "grad_norm": 0.6196553111076355,
978
- "learning_rate": 0.000558321301583881,
979
- "loss": 3.9771,
980
  "step": 6550
981
  },
982
  {
983
- "epoch": 0.7103648692282855,
984
- "grad_norm": 0.5549377799034119,
985
- "learning_rate": 0.0005579980605538196,
986
- "loss": 3.9651,
987
  "step": 6600
988
  },
989
  {
990
- "epoch": 0.7157464212678937,
991
- "grad_norm": 0.6142178177833557,
992
- "learning_rate": 0.0005576748195237582,
993
- "loss": 3.9533,
994
  "step": 6650
995
  },
996
  {
997
- "epoch": 0.7211279733075019,
998
- "grad_norm": 0.6508194208145142,
999
- "learning_rate": 0.0005573515784936967,
1000
- "loss": 3.972,
1001
  "step": 6700
1002
  },
1003
  {
1004
- "epoch": 0.7265095253471101,
1005
- "grad_norm": 0.6435700058937073,
1006
- "learning_rate": 0.0005570283374636353,
1007
- "loss": 3.9371,
1008
  "step": 6750
1009
  },
1010
  {
1011
- "epoch": 0.7318910773867183,
1012
- "grad_norm": 0.6877986192703247,
1013
- "learning_rate": 0.0005567050964335739,
1014
- "loss": 3.9625,
1015
  "step": 6800
1016
  },
1017
  {
1018
- "epoch": 0.7372726294263265,
1019
- "grad_norm": 0.6768865585327148,
1020
- "learning_rate": 0.0005563818554035126,
1021
- "loss": 3.9651,
1022
  "step": 6850
1023
  },
1024
  {
1025
- "epoch": 0.7426541814659348,
1026
- "grad_norm": 0.5656898021697998,
1027
- "learning_rate": 0.0005560586143734511,
1028
- "loss": 3.9244,
1029
  "step": 6900
1030
  },
1031
  {
1032
- "epoch": 0.748035733505543,
1033
- "grad_norm": 0.7026826739311218,
1034
- "learning_rate": 0.0005557353733433896,
1035
- "loss": 3.9553,
1036
  "step": 6950
1037
  },
1038
  {
1039
- "epoch": 0.7534172855451512,
1040
- "grad_norm": 0.6179747581481934,
1041
- "learning_rate": 0.0005554121323133283,
1042
- "loss": 3.9432,
1043
  "step": 7000
1044
  },
1045
  {
1046
- "epoch": 0.7534172855451512,
1047
- "eval_accuracy": 0.332839433522958,
1048
- "eval_loss": 3.8685381412506104,
1049
- "eval_runtime": 184.8874,
1050
- "eval_samples_per_second": 97.416,
1051
- "eval_steps_per_second": 6.09,
1052
  "step": 7000
1053
  },
1054
  {
1055
- "epoch": 0.7587988375847594,
1056
- "grad_norm": 0.5971046090126038,
1057
- "learning_rate": 0.0005550888912832668,
1058
- "loss": 3.9327,
1059
  "step": 7050
1060
  },
1061
  {
1062
- "epoch": 0.7641803896243676,
1063
- "grad_norm": 0.6499930620193481,
1064
- "learning_rate": 0.0005547656502532055,
1065
- "loss": 3.9357,
1066
  "step": 7100
1067
  },
1068
  {
1069
- "epoch": 0.7695619416639758,
1070
- "grad_norm": 0.652708113193512,
1071
- "learning_rate": 0.000554442409223144,
1072
- "loss": 3.9197,
1073
  "step": 7150
1074
  },
1075
  {
1076
- "epoch": 0.7749434937035841,
1077
- "grad_norm": 0.6613510847091675,
1078
- "learning_rate": 0.0005541191681930826,
1079
- "loss": 3.9122,
1080
  "step": 7200
1081
  },
1082
  {
1083
- "epoch": 0.7803250457431924,
1084
- "grad_norm": 0.6252796053886414,
1085
- "learning_rate": 0.0005537959271630212,
1086
- "loss": 3.9254,
1087
  "step": 7250
1088
  },
1089
  {
1090
- "epoch": 0.7857065977828006,
1091
- "grad_norm": 0.6944841742515564,
1092
- "learning_rate": 0.0005534726861329598,
1093
- "loss": 3.9347,
1094
  "step": 7300
1095
  },
1096
  {
1097
- "epoch": 0.7910881498224088,
1098
- "grad_norm": 0.6067506074905396,
1099
- "learning_rate": 0.0005531494451028983,
1100
- "loss": 3.9181,
1101
  "step": 7350
1102
  },
1103
  {
1104
- "epoch": 0.796469701862017,
1105
- "grad_norm": 0.671901524066925,
1106
- "learning_rate": 0.0005528262040728369,
1107
- "loss": 3.9295,
1108
  "step": 7400
1109
  },
1110
  {
1111
- "epoch": 0.8018512539016253,
1112
- "grad_norm": 0.5454155802726746,
1113
- "learning_rate": 0.0005525029630427755,
1114
- "loss": 3.8984,
1115
  "step": 7450
1116
  },
1117
  {
1118
- "epoch": 0.8072328059412335,
1119
- "grad_norm": 0.6592690348625183,
1120
- "learning_rate": 0.0005521797220127141,
1121
- "loss": 3.9373,
1122
  "step": 7500
1123
  },
1124
  {
1125
- "epoch": 0.8126143579808417,
1126
- "grad_norm": 0.5588394999504089,
1127
- "learning_rate": 0.0005518564809826527,
1128
- "loss": 3.9046,
1129
  "step": 7550
1130
  },
1131
  {
1132
- "epoch": 0.8179959100204499,
1133
- "grad_norm": 0.7025091648101807,
1134
- "learning_rate": 0.0005515332399525912,
1135
- "loss": 3.9059,
1136
  "step": 7600
1137
  },
1138
  {
1139
- "epoch": 0.8233774620600581,
1140
- "grad_norm": 0.6414318084716797,
1141
- "learning_rate": 0.0005512099989225299,
1142
- "loss": 3.9109,
1143
  "step": 7650
1144
  },
1145
  {
1146
- "epoch": 0.8287590140996663,
1147
- "grad_norm": 0.6233903169631958,
1148
- "learning_rate": 0.0005508867578924685,
1149
- "loss": 3.9003,
1150
  "step": 7700
1151
  },
1152
  {
1153
- "epoch": 0.8341405661392746,
1154
- "grad_norm": 0.690694272518158,
1155
- "learning_rate": 0.000550563516862407,
1156
- "loss": 3.9203,
1157
  "step": 7750
1158
  },
1159
  {
1160
- "epoch": 0.8395221181788828,
1161
- "grad_norm": 0.5982156991958618,
1162
- "learning_rate": 0.0005502402758323456,
1163
- "loss": 3.9269,
1164
  "step": 7800
1165
  },
1166
  {
1167
- "epoch": 0.844903670218491,
1168
- "grad_norm": 0.5323824882507324,
1169
- "learning_rate": 0.0005499170348022841,
1170
- "loss": 3.9035,
1171
  "step": 7850
1172
  },
1173
  {
1174
- "epoch": 0.8502852222580992,
1175
- "grad_norm": 0.6599534749984741,
1176
- "learning_rate": 0.0005495937937722228,
1177
- "loss": 3.8998,
1178
  "step": 7900
1179
  },
1180
  {
1181
- "epoch": 0.8556667742977074,
1182
- "grad_norm": 0.5557947158813477,
1183
- "learning_rate": 0.0005492705527421614,
1184
- "loss": 3.9154,
1185
  "step": 7950
1186
  },
1187
  {
1188
- "epoch": 0.8610483263373157,
1189
- "grad_norm": 0.6077902317047119,
1190
- "learning_rate": 0.0005489473117121,
1191
- "loss": 3.8951,
1192
  "step": 8000
1193
  },
1194
  {
1195
- "epoch": 0.8610483263373157,
1196
- "eval_accuracy": 0.3374400140987987,
1197
- "eval_loss": 3.822192907333374,
1198
- "eval_runtime": 184.6794,
1199
- "eval_samples_per_second": 97.526,
1200
- "eval_steps_per_second": 6.097,
1201
  "step": 8000
1202
  },
1203
  {
1204
- "epoch": 0.8664298783769239,
1205
- "grad_norm": 0.571945309638977,
1206
- "learning_rate": 0.0005486240706820385,
1207
- "loss": 3.8844,
1208
  "step": 8050
1209
  },
1210
  {
1211
- "epoch": 0.8718114304165321,
1212
- "grad_norm": 0.5621715784072876,
1213
- "learning_rate": 0.0005483008296519772,
1214
- "loss": 3.8981,
1215
  "step": 8100
1216
  },
1217
  {
1218
- "epoch": 0.8771929824561403,
1219
- "grad_norm": 0.6691741943359375,
1220
- "learning_rate": 0.0005479775886219157,
1221
- "loss": 3.9025,
1222
  "step": 8150
1223
  },
1224
  {
1225
- "epoch": 0.8825745344957485,
1226
- "grad_norm": 0.6594086289405823,
1227
- "learning_rate": 0.0005476543475918542,
1228
- "loss": 3.8895,
1229
  "step": 8200
1230
  },
1231
  {
1232
- "epoch": 0.8879560865353568,
1233
- "grad_norm": 0.56702721118927,
1234
- "learning_rate": 0.0005473311065617929,
1235
- "loss": 3.8608,
1236
  "step": 8250
1237
  },
1238
  {
1239
- "epoch": 0.8933376385749651,
1240
- "grad_norm": 0.5397810935974121,
1241
- "learning_rate": 0.0005470078655317314,
1242
- "loss": 3.8865,
1243
  "step": 8300
1244
  },
1245
  {
1246
- "epoch": 0.8987191906145733,
1247
- "grad_norm": 0.5843083262443542,
1248
- "learning_rate": 0.0005466846245016701,
1249
- "loss": 3.8676,
1250
  "step": 8350
1251
  },
1252
  {
1253
- "epoch": 0.9041007426541815,
1254
- "grad_norm": 0.7101298570632935,
1255
- "learning_rate": 0.0005463613834716086,
1256
- "loss": 3.8729,
1257
  "step": 8400
1258
  },
1259
  {
1260
- "epoch": 0.9094822946937897,
1261
- "grad_norm": 0.6375505924224854,
1262
- "learning_rate": 0.0005460381424415472,
1263
- "loss": 3.8818,
1264
  "step": 8450
1265
  },
1266
  {
1267
- "epoch": 0.9148638467333979,
1268
- "grad_norm": 0.6644571423530579,
1269
- "learning_rate": 0.0005457149014114858,
1270
- "loss": 3.8883,
1271
  "step": 8500
1272
  },
1273
  {
1274
- "epoch": 0.9202453987730062,
1275
- "grad_norm": 0.654179036617279,
1276
- "learning_rate": 0.0005453916603814243,
1277
- "loss": 3.8855,
1278
  "step": 8550
1279
  },
1280
  {
1281
- "epoch": 0.9256269508126144,
1282
- "grad_norm": 0.6557681560516357,
1283
- "learning_rate": 0.000545068419351363,
1284
- "loss": 3.8691,
1285
  "step": 8600
1286
  },
1287
  {
1288
- "epoch": 0.9310085028522226,
1289
- "grad_norm": 0.5975877642631531,
1290
- "learning_rate": 0.0005447451783213015,
1291
- "loss": 3.8716,
1292
  "step": 8650
1293
  },
1294
  {
1295
- "epoch": 0.9363900548918308,
1296
- "grad_norm": 0.5875534415245056,
1297
- "learning_rate": 0.0005444219372912401,
1298
- "loss": 3.8659,
1299
  "step": 8700
1300
  },
1301
  {
1302
- "epoch": 0.941771606931439,
1303
- "grad_norm": 0.5575305223464966,
1304
- "learning_rate": 0.0005440986962611787,
1305
- "loss": 3.8586,
1306
  "step": 8750
1307
  },
1308
  {
1309
- "epoch": 0.9471531589710472,
1310
- "grad_norm": 0.5902034640312195,
1311
- "learning_rate": 0.0005437754552311173,
1312
- "loss": 3.859,
1313
  "step": 8800
1314
  },
1315
  {
1316
- "epoch": 0.9525347110106555,
1317
- "grad_norm": 0.6572847366333008,
1318
- "learning_rate": 0.0005434522142010559,
1319
- "loss": 3.8656,
1320
  "step": 8850
1321
  },
1322
  {
1323
- "epoch": 0.9579162630502637,
1324
- "grad_norm": 0.5779463648796082,
1325
- "learning_rate": 0.0005431289731709945,
1326
- "loss": 3.8557,
1327
  "step": 8900
1328
  },
1329
  {
1330
- "epoch": 0.9632978150898719,
1331
- "grad_norm": 0.6097962260246277,
1332
- "learning_rate": 0.000542805732140933,
1333
- "loss": 3.8618,
1334
  "step": 8950
1335
  },
1336
  {
1337
- "epoch": 0.9686793671294801,
1338
- "grad_norm": 0.6319593191146851,
1339
- "learning_rate": 0.0005424824911108716,
1340
- "loss": 3.8465,
1341
  "step": 9000
1342
  },
1343
  {
1344
- "epoch": 0.9686793671294801,
1345
- "eval_accuracy": 0.34105967640345036,
1346
- "eval_loss": 3.7847001552581787,
1347
- "eval_runtime": 184.8533,
1348
- "eval_samples_per_second": 97.434,
1349
- "eval_steps_per_second": 6.091,
1350
  "step": 9000
1351
  },
1352
  {
1353
- "epoch": 0.9740609191690883,
1354
- "grad_norm": 0.6114616394042969,
1355
- "learning_rate": 0.0005421592500808102,
1356
- "loss": 3.859,
1357
  "step": 9050
1358
  },
1359
  {
1360
- "epoch": 0.9794424712086965,
1361
- "grad_norm": 0.5971494317054749,
1362
- "learning_rate": 0.0005418360090507488,
1363
- "loss": 3.8401,
1364
  "step": 9100
1365
  },
1366
  {
1367
- "epoch": 0.9848240232483048,
1368
- "grad_norm": 0.6119955778121948,
1369
- "learning_rate": 0.0005415127680206874,
1370
- "loss": 3.8575,
1371
  "step": 9150
1372
  },
1373
  {
1374
- "epoch": 0.9902055752879131,
1375
- "grad_norm": 0.5487962365150452,
1376
- "learning_rate": 0.0005411895269906259,
1377
- "loss": 3.8552,
1378
  "step": 9200
1379
  },
1380
  {
1381
- "epoch": 0.9955871273275213,
1382
- "grad_norm": 0.631289005279541,
1383
- "learning_rate": 0.0005408662859605646,
1384
- "loss": 3.8509,
1385
  "step": 9250
1386
  },
1387
  {
1388
- "epoch": 1.0009686793671295,
1389
- "grad_norm": 0.615967333316803,
1390
- "learning_rate": 0.0005405430449305031,
1391
- "loss": 3.8556,
1392
  "step": 9300
1393
  },
1394
  {
1395
- "epoch": 1.0063502314067376,
1396
- "grad_norm": 0.5872229337692261,
1397
- "learning_rate": 0.0005402198039004416,
1398
- "loss": 3.7921,
1399
  "step": 9350
1400
  },
1401
  {
1402
- "epoch": 1.011731783446346,
1403
- "grad_norm": 0.5617460012435913,
1404
- "learning_rate": 0.0005398965628703803,
1405
- "loss": 3.7792,
1406
  "step": 9400
1407
  },
1408
  {
1409
- "epoch": 1.017113335485954,
1410
- "grad_norm": 0.6194376349449158,
1411
- "learning_rate": 0.0005395733218403189,
1412
- "loss": 3.7967,
1413
  "step": 9450
1414
  },
1415
  {
1416
- "epoch": 1.0224948875255624,
1417
- "grad_norm": 0.5620608329772949,
1418
- "learning_rate": 0.0005392500808102575,
1419
- "loss": 3.7958,
1420
  "step": 9500
1421
  },
1422
  {
1423
- "epoch": 1.0278764395651705,
1424
- "grad_norm": 0.6231437921524048,
1425
- "learning_rate": 0.000538926839780196,
1426
- "loss": 3.8046,
1427
  "step": 9550
1428
  },
1429
  {
1430
- "epoch": 1.0332579916047788,
1431
- "grad_norm": 0.6144508719444275,
1432
- "learning_rate": 0.0005386035987501346,
1433
- "loss": 3.7608,
1434
  "step": 9600
1435
  },
1436
  {
1437
- "epoch": 1.0386395436443872,
1438
- "grad_norm": 0.8581711649894714,
1439
- "learning_rate": 0.0005382803577200732,
1440
- "loss": 3.7768,
1441
  "step": 9650
1442
  },
1443
  {
1444
- "epoch": 1.0440210956839953,
1445
- "grad_norm": 0.5918218493461609,
1446
- "learning_rate": 0.0005379571166900119,
1447
- "loss": 3.7755,
1448
  "step": 9700
1449
  },
1450
  {
1451
- "epoch": 1.0494026477236036,
1452
- "grad_norm": 0.6462705135345459,
1453
- "learning_rate": 0.0005376338756599504,
1454
- "loss": 3.7889,
1455
  "step": 9750
1456
  },
1457
  {
1458
- "epoch": 1.0547841997632117,
1459
- "grad_norm": 0.7305418848991394,
1460
- "learning_rate": 0.0005373106346298889,
1461
- "loss": 3.7776,
1462
  "step": 9800
1463
  },
1464
  {
1465
- "epoch": 1.06016575180282,
1466
- "grad_norm": 0.6626386642456055,
1467
- "learning_rate": 0.0005369873935998276,
1468
- "loss": 3.7625,
1469
  "step": 9850
1470
  },
1471
  {
1472
- "epoch": 1.0655473038424281,
1473
- "grad_norm": 0.5469115376472473,
1474
- "learning_rate": 0.0005366641525697661,
1475
- "loss": 3.7704,
1476
  "step": 9900
1477
  },
1478
  {
1479
- "epoch": 1.0709288558820365,
1480
- "grad_norm": 0.607982337474823,
1481
- "learning_rate": 0.0005363409115397048,
1482
- "loss": 3.7885,
1483
  "step": 9950
1484
  },
1485
  {
1486
- "epoch": 1.0763104079216446,
1487
- "grad_norm": 0.5802627205848694,
1488
- "learning_rate": 0.0005360176705096433,
1489
- "loss": 3.7776,
1490
  "step": 10000
1491
  },
1492
  {
1493
- "epoch": 1.0763104079216446,
1494
- "eval_accuracy": 0.34371732603939253,
1495
- "eval_loss": 3.7542877197265625,
1496
- "eval_runtime": 184.5847,
1497
- "eval_samples_per_second": 97.576,
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": 8.3610314735616e+16,
1520
  "train_batch_size": 32,
1521
  "trial_name": null,
1522
  "trial_params": null
 
1
  {
2
+ "best_metric": 3.7473487854003906,
3
+ "best_model_checkpoint": "/scratch/cl5625/exceptions/models/100M_high_2000_634/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": 1.4153761863708496,
14
+ "learning_rate": 0.0003,
15
+ "loss": 8.6258,
16
  "step": 50
17
  },
18
  {
19
+ "epoch": 0.01078167115902965,
20
+ "grad_norm": 4.798243045806885,
21
+ "learning_rate": 0.0006,
22
+ "loss": 6.9282,
23
  "step": 100
24
  },
25
  {
26
+ "epoch": 0.016172506738544475,
27
+ "grad_norm": 2.44549560546875,
28
+ "learning_rate": 0.0005996762007555315,
29
+ "loss": 6.5183,
30
  "step": 150
31
  },
32
  {
33
+ "epoch": 0.0215633423180593,
34
+ "grad_norm": 1.1002956628799438,
35
+ "learning_rate": 0.000599352401511063,
36
+ "loss": 6.263,
37
  "step": 200
38
  },
39
  {
40
+ "epoch": 0.026954177897574125,
41
+ "grad_norm": 0.991017758846283,
42
+ "learning_rate": 0.0005990286022665946,
43
+ "loss": 6.1066,
44
  "step": 250
45
  },
46
  {
47
+ "epoch": 0.03234501347708895,
48
+ "grad_norm": 1.7810102701187134,
49
+ "learning_rate": 0.0005987048030221263,
50
+ "loss": 5.9974,
51
  "step": 300
52
  },
53
  {
54
+ "epoch": 0.03773584905660377,
55
+ "grad_norm": 1.4339019060134888,
56
+ "learning_rate": 0.0005983810037776578,
57
+ "loss": 5.8832,
58
  "step": 350
59
  },
60
  {
61
+ "epoch": 0.0431266846361186,
62
+ "grad_norm": 1.7068945169448853,
63
+ "learning_rate": 0.0005980572045331894,
64
+ "loss": 5.7989,
65
  "step": 400
66
  },
67
  {
68
+ "epoch": 0.04851752021563342,
69
+ "grad_norm": 1.2276661396026611,
70
+ "learning_rate": 0.0005977334052887209,
71
+ "loss": 5.7294,
72
  "step": 450
73
  },
74
  {
75
+ "epoch": 0.05390835579514825,
76
+ "grad_norm": 1.4345930814743042,
77
+ "learning_rate": 0.0005974096060442526,
78
+ "loss": 5.6735,
79
  "step": 500
80
  },
81
  {
82
+ "epoch": 0.05929919137466307,
83
+ "grad_norm": 1.651483178138733,
84
+ "learning_rate": 0.0005970858067997841,
85
+ "loss": 5.5941,
86
  "step": 550
87
  },
88
  {
89
+ "epoch": 0.0646900269541779,
90
+ "grad_norm": 1.4423189163208008,
91
+ "learning_rate": 0.0005967620075553157,
92
+ "loss": 5.528,
93
  "step": 600
94
  },
95
  {
96
+ "epoch": 0.07008086253369272,
97
+ "grad_norm": 0.9989488124847412,
98
+ "learning_rate": 0.0005964382083108472,
99
+ "loss": 5.4406,
100
  "step": 650
101
  },
102
  {
103
+ "epoch": 0.07547169811320754,
104
+ "grad_norm": 1.5697565078735352,
105
+ "learning_rate": 0.0005961144090663788,
106
+ "loss": 5.3854,
107
  "step": 700
108
  },
109
  {
110
+ "epoch": 0.08086253369272237,
111
+ "grad_norm": 1.3698347806930542,
112
+ "learning_rate": 0.0005957906098219104,
113
+ "loss": 5.316,
114
  "step": 750
115
  },
116
  {
117
+ "epoch": 0.0862533692722372,
118
+ "grad_norm": 0.9174061417579651,
119
+ "learning_rate": 0.0005954668105774419,
120
+ "loss": 5.2632,
121
  "step": 800
122
  },
123
  {
124
+ "epoch": 0.09164420485175202,
125
+ "grad_norm": 0.8781394958496094,
126
+ "learning_rate": 0.0005951430113329735,
127
+ "loss": 5.2044,
128
  "step": 850
129
  },
130
  {
131
+ "epoch": 0.09703504043126684,
132
+ "grad_norm": 0.9764671325683594,
133
+ "learning_rate": 0.0005948192120885051,
134
+ "loss": 5.1792,
135
  "step": 900
136
  },
137
  {
138
+ "epoch": 0.10242587601078167,
139
+ "grad_norm": 1.7501579523086548,
140
+ "learning_rate": 0.0005944954128440366,
141
+ "loss": 5.1443,
142
  "step": 950
143
  },
144
  {
145
+ "epoch": 0.1078167115902965,
146
+ "grad_norm": 1.8602075576782227,
147
+ "learning_rate": 0.0005941716135995682,
148
+ "loss": 5.0898,
149
  "step": 1000
150
  },
151
  {
152
+ "epoch": 0.1078167115902965,
153
+ "eval_accuracy": 0.22601321805841418,
154
+ "eval_loss": 5.037795066833496,
155
+ "eval_runtime": 187.186,
156
+ "eval_samples_per_second": 96.22,
157
+ "eval_steps_per_second": 6.015,
158
  "step": 1000
159
  },
160
  {
161
+ "epoch": 0.11320754716981132,
162
+ "grad_norm": 0.9920917749404907,
163
+ "learning_rate": 0.0005938478143550997,
164
+ "loss": 5.0498,
165
  "step": 1050
166
  },
167
  {
168
+ "epoch": 0.11859838274932614,
169
+ "grad_norm": 1.1556609869003296,
170
+ "learning_rate": 0.0005935240151106314,
171
+ "loss": 5.0218,
172
  "step": 1100
173
  },
174
  {
175
+ "epoch": 0.12398921832884097,
176
+ "grad_norm": 1.180494785308838,
177
+ "learning_rate": 0.0005932002158661629,
178
+ "loss": 4.9949,
179
  "step": 1150
180
  },
181
  {
182
+ "epoch": 0.1293800539083558,
183
+ "grad_norm": 1.130318522453308,
184
+ "learning_rate": 0.0005928764166216945,
185
+ "loss": 4.9329,
186
  "step": 1200
187
  },
188
  {
189
+ "epoch": 0.1347708894878706,
190
+ "grad_norm": 1.003656029701233,
191
+ "learning_rate": 0.000592552617377226,
192
+ "loss": 4.9114,
193
  "step": 1250
194
  },
195
  {
196
+ "epoch": 0.14016172506738545,
197
+ "grad_norm": 1.564238429069519,
198
+ "learning_rate": 0.0005922288181327577,
199
+ "loss": 4.8916,
200
  "step": 1300
201
  },
202
  {
203
+ "epoch": 0.14555256064690028,
204
+ "grad_norm": 0.8959649801254272,
205
+ "learning_rate": 0.0005919050188882893,
206
+ "loss": 4.8536,
207
  "step": 1350
208
  },
209
  {
210
+ "epoch": 0.1509433962264151,
211
+ "grad_norm": 0.8476941585540771,
212
+ "learning_rate": 0.0005915812196438207,
213
+ "loss": 4.8471,
214
  "step": 1400
215
  },
216
  {
217
+ "epoch": 0.15633423180592992,
218
+ "grad_norm": 1.2242474555969238,
219
+ "learning_rate": 0.0005912574203993524,
220
+ "loss": 4.8488,
221
  "step": 1450
222
  },
223
  {
224
+ "epoch": 0.16172506738544473,
225
+ "grad_norm": 0.8395281434059143,
226
+ "learning_rate": 0.0005909336211548839,
227
+ "loss": 4.8122,
228
  "step": 1500
229
  },
230
  {
231
+ "epoch": 0.16711590296495957,
232
+ "grad_norm": 1.100954294204712,
233
+ "learning_rate": 0.0005906098219104155,
234
+ "loss": 4.7627,
235
  "step": 1550
236
  },
237
  {
238
+ "epoch": 0.1725067385444744,
239
+ "grad_norm": 0.8952234387397766,
240
+ "learning_rate": 0.000590286022665947,
241
+ "loss": 4.7551,
242
  "step": 1600
243
  },
244
  {
245
+ "epoch": 0.1778975741239892,
246
+ "grad_norm": 0.9563343524932861,
247
+ "learning_rate": 0.0005899622234214787,
248
+ "loss": 4.7286,
249
  "step": 1650
250
  },
251
  {
252
+ "epoch": 0.18328840970350405,
253
+ "grad_norm": 0.9019905924797058,
254
+ "learning_rate": 0.0005896384241770102,
255
+ "loss": 4.7041,
256
  "step": 1700
257
  },
258
  {
259
+ "epoch": 0.18867924528301888,
260
+ "grad_norm": 1.0329313278198242,
261
+ "learning_rate": 0.0005893146249325418,
262
+ "loss": 4.714,
263
  "step": 1750
264
  },
265
  {
266
+ "epoch": 0.1940700808625337,
267
+ "grad_norm": 1.022636890411377,
268
+ "learning_rate": 0.0005889908256880733,
269
+ "loss": 4.679,
270
  "step": 1800
271
  },
272
  {
273
+ "epoch": 0.19946091644204852,
274
+ "grad_norm": 1.012805700302124,
275
+ "learning_rate": 0.0005886670264436049,
276
+ "loss": 4.6485,
277
  "step": 1850
278
  },
279
  {
280
+ "epoch": 0.20485175202156333,
281
+ "grad_norm": 1.017619252204895,
282
+ "learning_rate": 0.0005883432271991365,
283
+ "loss": 4.6367,
284
  "step": 1900
285
  },
286
  {
287
+ "epoch": 0.21024258760107817,
288
+ "grad_norm": 1.0787638425827026,
289
+ "learning_rate": 0.0005880194279546681,
290
+ "loss": 4.6371,
291
  "step": 1950
292
  },
293
  {
294
+ "epoch": 0.215633423180593,
295
+ "grad_norm": 1.0098811388015747,
296
+ "learning_rate": 0.0005876956287101996,
297
+ "loss": 4.5856,
298
  "step": 2000
299
  },
300
  {
301
+ "epoch": 0.215633423180593,
302
+ "eval_accuracy": 0.26980304816984535,
303
+ "eval_loss": 4.516435146331787,
304
+ "eval_runtime": 185.8323,
305
+ "eval_samples_per_second": 96.921,
306
+ "eval_steps_per_second": 6.059,
307
  "step": 2000
308
  },
309
  {
310
+ "epoch": 0.2210242587601078,
311
+ "grad_norm": 0.8449752330780029,
312
+ "learning_rate": 0.0005873718294657312,
313
+ "loss": 4.5674,
314
  "step": 2050
315
  },
316
  {
317
+ "epoch": 0.22641509433962265,
318
+ "grad_norm": 0.6898910999298096,
319
+ "learning_rate": 0.0005870480302212628,
320
+ "loss": 4.5584,
321
  "step": 2100
322
  },
323
  {
324
+ "epoch": 0.23180592991913745,
325
+ "grad_norm": 0.9888122081756592,
326
+ "learning_rate": 0.0005867242309767943,
327
+ "loss": 4.5375,
328
  "step": 2150
329
  },
330
  {
331
+ "epoch": 0.2371967654986523,
332
+ "grad_norm": 0.8149335384368896,
333
+ "learning_rate": 0.0005864004317323259,
334
+ "loss": 4.4984,
335
  "step": 2200
336
  },
337
  {
338
+ "epoch": 0.24258760107816713,
339
+ "grad_norm": 1.0397859811782837,
340
+ "learning_rate": 0.0005860766324878575,
341
+ "loss": 4.5126,
342
  "step": 2250
343
  },
344
  {
345
+ "epoch": 0.24797843665768193,
346
+ "grad_norm": 0.8346261978149414,
347
+ "learning_rate": 0.000585752833243389,
348
+ "loss": 4.4873,
349
  "step": 2300
350
  },
351
  {
352
+ "epoch": 0.25336927223719674,
353
+ "grad_norm": 0.9166735410690308,
354
+ "learning_rate": 0.0005854290339989206,
355
+ "loss": 4.4653,
356
  "step": 2350
357
  },
358
  {
359
+ "epoch": 0.2587601078167116,
360
+ "grad_norm": 0.9180364012718201,
361
+ "learning_rate": 0.0005851052347544521,
362
+ "loss": 4.4583,
363
  "step": 2400
364
  },
365
  {
366
+ "epoch": 0.2641509433962264,
367
+ "grad_norm": 0.8483642339706421,
368
+ "learning_rate": 0.0005847814355099838,
369
+ "loss": 4.4384,
370
  "step": 2450
371
  },
372
  {
373
+ "epoch": 0.2695417789757412,
374
+ "grad_norm": 0.7761411070823669,
375
+ "learning_rate": 0.0005844576362655154,
376
+ "loss": 4.4318,
377
  "step": 2500
378
  },
379
  {
380
+ "epoch": 0.2749326145552561,
381
+ "grad_norm": 0.8839765787124634,
382
+ "learning_rate": 0.0005841338370210469,
383
+ "loss": 4.4097,
384
  "step": 2550
385
  },
386
  {
387
+ "epoch": 0.2803234501347709,
388
+ "grad_norm": 1.1479021310806274,
389
+ "learning_rate": 0.0005838100377765785,
390
+ "loss": 4.3848,
391
  "step": 2600
392
  },
393
  {
394
+ "epoch": 0.2857142857142857,
395
+ "grad_norm": 0.8199343085289001,
396
+ "learning_rate": 0.0005834862385321101,
397
+ "loss": 4.3774,
398
  "step": 2650
399
  },
400
  {
401
+ "epoch": 0.29110512129380056,
402
+ "grad_norm": 0.8475073575973511,
403
+ "learning_rate": 0.0005831624392876417,
404
+ "loss": 4.3724,
405
  "step": 2700
406
  },
407
  {
408
+ "epoch": 0.29649595687331537,
409
+ "grad_norm": 0.9192224144935608,
410
+ "learning_rate": 0.0005828386400431731,
411
+ "loss": 4.3824,
412
  "step": 2750
413
  },
414
  {
415
+ "epoch": 0.3018867924528302,
416
+ "grad_norm": 0.7423735857009888,
417
+ "learning_rate": 0.0005825148407987048,
418
+ "loss": 4.3731,
419
  "step": 2800
420
  },
421
  {
422
+ "epoch": 0.30727762803234504,
423
+ "grad_norm": 0.8002654314041138,
424
+ "learning_rate": 0.0005821910415542363,
425
+ "loss": 4.3368,
426
  "step": 2850
427
  },
428
  {
429
+ "epoch": 0.31266846361185985,
430
+ "grad_norm": 0.8412148356437683,
431
+ "learning_rate": 0.0005818672423097679,
432
+ "loss": 4.3234,
433
  "step": 2900
434
  },
435
  {
436
+ "epoch": 0.31805929919137466,
437
+ "grad_norm": 0.846697986125946,
438
+ "learning_rate": 0.0005815434430652994,
439
+ "loss": 4.3543,
440
  "step": 2950
441
  },
442
  {
443
+ "epoch": 0.32345013477088946,
444
+ "grad_norm": 0.7227766513824463,
445
+ "learning_rate": 0.0005812196438208311,
446
+ "loss": 4.318,
447
  "step": 3000
448
  },
449
  {
450
+ "epoch": 0.32345013477088946,
451
+ "eval_accuracy": 0.29839266523469404,
452
+ "eval_loss": 4.236840724945068,
453
+ "eval_runtime": 179.5407,
454
+ "eval_samples_per_second": 100.317,
455
+ "eval_steps_per_second": 6.272,
456
  "step": 3000
457
  },
458
  {
459
+ "epoch": 0.3288409703504043,
460
+ "grad_norm": 0.8503046631813049,
461
+ "learning_rate": 0.0005808958445763626,
462
+ "loss": 4.309,
463
  "step": 3050
464
  },
465
  {
466
+ "epoch": 0.33423180592991913,
467
+ "grad_norm": 0.8151911497116089,
468
+ "learning_rate": 0.0005805720453318942,
469
+ "loss": 4.2919,
470
  "step": 3100
471
  },
472
  {
473
+ "epoch": 0.33962264150943394,
474
+ "grad_norm": 0.8732330203056335,
475
+ "learning_rate": 0.0005802482460874257,
476
+ "loss": 4.2785,
477
  "step": 3150
478
  },
479
  {
480
+ "epoch": 0.3450134770889488,
481
+ "grad_norm": 0.7436563372612,
482
+ "learning_rate": 0.0005799244468429573,
483
+ "loss": 4.2827,
484
  "step": 3200
485
  },
486
  {
487
+ "epoch": 0.3504043126684636,
488
+ "grad_norm": 1.0856059789657593,
489
+ "learning_rate": 0.0005796006475984889,
490
+ "loss": 4.2564,
491
  "step": 3250
492
  },
493
  {
494
+ "epoch": 0.3557951482479784,
495
+ "grad_norm": 0.6840450167655945,
496
+ "learning_rate": 0.0005792768483540205,
497
+ "loss": 4.2669,
498
  "step": 3300
499
  },
500
  {
501
+ "epoch": 0.3611859838274933,
502
+ "grad_norm": 0.9430235028266907,
503
+ "learning_rate": 0.000578953049109552,
504
+ "loss": 4.2636,
505
  "step": 3350
506
  },
507
  {
508
+ "epoch": 0.3665768194070081,
509
+ "grad_norm": 0.7581385970115662,
510
+ "learning_rate": 0.0005786292498650836,
511
+ "loss": 4.2397,
512
  "step": 3400
513
  },
514
  {
515
+ "epoch": 0.3719676549865229,
516
+ "grad_norm": 0.9222409725189209,
517
+ "learning_rate": 0.0005783054506206152,
518
+ "loss": 4.2354,
519
  "step": 3450
520
  },
521
  {
522
+ "epoch": 0.37735849056603776,
523
+ "grad_norm": 0.6859261989593506,
524
+ "learning_rate": 0.0005779816513761467,
525
+ "loss": 4.2277,
526
  "step": 3500
527
  },
528
  {
529
+ "epoch": 0.38274932614555257,
530
+ "grad_norm": 0.8531029224395752,
531
+ "learning_rate": 0.0005776578521316782,
532
+ "loss": 4.2054,
533
  "step": 3550
534
  },
535
  {
536
+ "epoch": 0.3881401617250674,
537
+ "grad_norm": 0.8664888739585876,
538
+ "learning_rate": 0.0005773340528872099,
539
+ "loss": 4.2126,
540
  "step": 3600
541
  },
542
  {
543
+ "epoch": 0.3935309973045822,
544
+ "grad_norm": 0.7247611880302429,
545
+ "learning_rate": 0.0005770102536427414,
546
+ "loss": 4.2271,
547
  "step": 3650
548
  },
549
  {
550
+ "epoch": 0.39892183288409705,
551
+ "grad_norm": 0.7640480995178223,
552
+ "learning_rate": 0.000576686454398273,
553
+ "loss": 4.2139,
554
  "step": 3700
555
  },
556
  {
557
+ "epoch": 0.40431266846361186,
558
+ "grad_norm": 0.7702681422233582,
559
+ "learning_rate": 0.0005763626551538045,
560
+ "loss": 4.1992,
561
  "step": 3750
562
  },
563
  {
564
+ "epoch": 0.40970350404312667,
565
+ "grad_norm": 0.7853514552116394,
566
+ "learning_rate": 0.0005760388559093362,
567
+ "loss": 4.189,
568
  "step": 3800
569
  },
570
  {
571
+ "epoch": 0.41509433962264153,
572
+ "grad_norm": 0.7235543727874756,
573
+ "learning_rate": 0.0005757150566648678,
574
+ "loss": 4.1822,
575
  "step": 3850
576
  },
577
  {
578
+ "epoch": 0.42048517520215634,
579
+ "grad_norm": 0.7994705438613892,
580
+ "learning_rate": 0.0005753912574203993,
581
+ "loss": 4.176,
582
  "step": 3900
583
  },
584
  {
585
+ "epoch": 0.42587601078167114,
586
+ "grad_norm": 0.7909198999404907,
587
+ "learning_rate": 0.0005750674581759309,
588
+ "loss": 4.1754,
589
  "step": 3950
590
  },
591
  {
592
+ "epoch": 0.431266846361186,
593
+ "grad_norm": 0.8087334632873535,
594
+ "learning_rate": 0.0005747436589314624,
595
+ "loss": 4.1599,
596
  "step": 4000
597
  },
598
  {
599
+ "epoch": 0.431266846361186,
600
+ "eval_accuracy": 0.3121017260489105,
601
+ "eval_loss": 4.09109354019165,
602
+ "eval_runtime": 179.4252,
603
+ "eval_samples_per_second": 100.382,
604
+ "eval_steps_per_second": 6.276,
605
  "step": 4000
606
  },
607
  {
608
+ "epoch": 0.4366576819407008,
609
+ "grad_norm": 0.790748655796051,
610
+ "learning_rate": 0.0005744198596869941,
611
+ "loss": 4.1715,
612
  "step": 4050
613
  },
614
  {
615
+ "epoch": 0.4420485175202156,
616
+ "grad_norm": 0.6250526905059814,
617
+ "learning_rate": 0.0005740960604425255,
618
+ "loss": 4.1647,
619
  "step": 4100
620
  },
621
  {
622
+ "epoch": 0.4474393530997305,
623
+ "grad_norm": 0.7327976822853088,
624
+ "learning_rate": 0.0005737722611980572,
625
+ "loss": 4.1432,
626
  "step": 4150
627
  },
628
  {
629
+ "epoch": 0.4528301886792453,
630
+ "grad_norm": 0.6393489241600037,
631
+ "learning_rate": 0.0005734484619535887,
632
+ "loss": 4.1223,
633
  "step": 4200
634
  },
635
  {
636
+ "epoch": 0.4582210242587601,
637
+ "grad_norm": 0.6468517780303955,
638
+ "learning_rate": 0.0005731246627091203,
639
+ "loss": 4.1544,
640
  "step": 4250
641
  },
642
  {
643
+ "epoch": 0.4636118598382749,
644
+ "grad_norm": 0.631144106388092,
645
+ "learning_rate": 0.0005728008634646518,
646
+ "loss": 4.1296,
647
  "step": 4300
648
  },
649
  {
650
+ "epoch": 0.46900269541778977,
651
+ "grad_norm": 0.7250542640686035,
652
+ "learning_rate": 0.0005724770642201835,
653
+ "loss": 4.1325,
654
  "step": 4350
655
  },
656
  {
657
+ "epoch": 0.4743935309973046,
658
+ "grad_norm": 0.7701312303543091,
659
+ "learning_rate": 0.000572153264975715,
660
+ "loss": 4.128,
661
  "step": 4400
662
  },
663
  {
664
+ "epoch": 0.4797843665768194,
665
+ "grad_norm": 0.7531642317771912,
666
+ "learning_rate": 0.0005718294657312466,
667
+ "loss": 4.1011,
668
  "step": 4450
669
  },
670
  {
671
+ "epoch": 0.48517520215633425,
672
+ "grad_norm": 0.6209056377410889,
673
+ "learning_rate": 0.0005715056664867781,
674
+ "loss": 4.0979,
675
  "step": 4500
676
  },
677
  {
678
+ "epoch": 0.49056603773584906,
679
+ "grad_norm": 0.7149604558944702,
680
+ "learning_rate": 0.0005711818672423097,
681
+ "loss": 4.0913,
682
  "step": 4550
683
  },
684
  {
685
+ "epoch": 0.49595687331536387,
686
+ "grad_norm": 0.621547281742096,
687
+ "learning_rate": 0.0005708580679978413,
688
+ "loss": 4.0978,
689
  "step": 4600
690
  },
691
  {
692
+ "epoch": 0.5013477088948787,
693
+ "grad_norm": 0.8008401989936829,
694
+ "learning_rate": 0.0005705342687533729,
695
+ "loss": 4.0927,
696
  "step": 4650
697
  },
698
  {
699
+ "epoch": 0.5067385444743935,
700
+ "grad_norm": 0.7120444774627686,
701
+ "learning_rate": 0.0005702104695089044,
702
+ "loss": 4.0748,
703
  "step": 4700
704
  },
705
  {
706
+ "epoch": 0.5121293800539084,
707
+ "grad_norm": 0.8041061758995056,
708
+ "learning_rate": 0.000569886670264436,
709
+ "loss": 4.0861,
710
  "step": 4750
711
  },
712
  {
713
+ "epoch": 0.5175202156334232,
714
+ "grad_norm": 0.7002294063568115,
715
+ "learning_rate": 0.0005695628710199675,
716
+ "loss": 4.0773,
717
  "step": 4800
718
  },
719
  {
720
+ "epoch": 0.522911051212938,
721
+ "grad_norm": 0.5887237191200256,
722
+ "learning_rate": 0.0005692390717754991,
723
+ "loss": 4.0713,
724
  "step": 4850
725
  },
726
  {
727
+ "epoch": 0.5283018867924528,
728
+ "grad_norm": 0.6446994543075562,
729
+ "learning_rate": 0.0005689152725310306,
730
+ "loss": 4.0624,
731
  "step": 4900
732
  },
733
  {
734
+ "epoch": 0.5336927223719676,
735
+ "grad_norm": 0.6368472576141357,
736
+ "learning_rate": 0.0005685914732865623,
737
+ "loss": 4.0632,
738
  "step": 4950
739
  },
740
  {
741
+ "epoch": 0.5390835579514824,
742
+ "grad_norm": 0.6480772495269775,
743
+ "learning_rate": 0.0005682676740420939,
744
+ "loss": 4.0769,
745
  "step": 5000
746
  },
747
  {
748
+ "epoch": 0.5390835579514824,
749
+ "eval_accuracy": 0.321377857693184,
750
+ "eval_loss": 3.994414806365967,
751
+ "eval_runtime": 179.4261,
752
+ "eval_samples_per_second": 100.381,
753
+ "eval_steps_per_second": 6.276,
754
  "step": 5000
755
  },
756
  {
757
+ "epoch": 0.5444743935309974,
758
+ "grad_norm": 0.7073789834976196,
759
+ "learning_rate": 0.0005679438747976254,
760
+ "loss": 4.0512,
761
  "step": 5050
762
  },
763
  {
764
+ "epoch": 0.5498652291105122,
765
+ "grad_norm": 0.6471377015113831,
766
+ "learning_rate": 0.000567620075553157,
767
+ "loss": 4.0507,
768
  "step": 5100
769
  },
770
  {
771
+ "epoch": 0.555256064690027,
772
+ "grad_norm": 0.7392793297767639,
773
+ "learning_rate": 0.0005672962763086886,
774
+ "loss": 4.0555,
775
  "step": 5150
776
  },
777
  {
778
+ "epoch": 0.5606469002695418,
779
+ "grad_norm": 0.7037752866744995,
780
+ "learning_rate": 0.0005669724770642202,
781
+ "loss": 4.0371,
782
  "step": 5200
783
  },
784
  {
785
+ "epoch": 0.5660377358490566,
786
+ "grad_norm": 0.6513423323631287,
787
+ "learning_rate": 0.0005666486778197517,
788
+ "loss": 4.034,
789
  "step": 5250
790
  },
791
  {
792
+ "epoch": 0.5714285714285714,
793
+ "grad_norm": 0.6977173686027527,
794
+ "learning_rate": 0.0005663248785752833,
795
+ "loss": 4.0483,
796
  "step": 5300
797
  },
798
  {
799
+ "epoch": 0.5768194070080862,
800
+ "grad_norm": 0.6718908548355103,
801
+ "learning_rate": 0.0005660010793308148,
802
+ "loss": 4.0295,
803
  "step": 5350
804
  },
805
  {
806
+ "epoch": 0.5822102425876011,
807
+ "grad_norm": 0.6981878280639648,
808
+ "learning_rate": 0.0005656772800863465,
809
+ "loss": 4.0463,
810
  "step": 5400
811
  },
812
  {
813
+ "epoch": 0.5876010781671159,
814
+ "grad_norm": 0.6827294230461121,
815
+ "learning_rate": 0.0005653534808418779,
816
+ "loss": 4.0234,
817
  "step": 5450
818
  },
819
  {
820
+ "epoch": 0.5929919137466307,
821
+ "grad_norm": 0.6582399010658264,
822
+ "learning_rate": 0.0005650296815974096,
823
+ "loss": 4.0268,
824
  "step": 5500
825
  },
826
  {
827
+ "epoch": 0.5983827493261455,
828
+ "grad_norm": 0.6177818775177002,
829
+ "learning_rate": 0.0005647058823529411,
830
+ "loss": 4.0223,
831
  "step": 5550
832
  },
833
  {
834
+ "epoch": 0.6037735849056604,
835
+ "grad_norm": 0.5578612685203552,
836
+ "learning_rate": 0.0005643820831084727,
837
+ "loss": 4.0108,
838
  "step": 5600
839
  },
840
  {
841
+ "epoch": 0.6091644204851752,
842
+ "grad_norm": 0.6668821573257446,
843
+ "learning_rate": 0.0005640582838640042,
844
+ "loss": 4.0237,
845
  "step": 5650
846
  },
847
  {
848
+ "epoch": 0.6145552560646901,
849
+ "grad_norm": 0.742559552192688,
850
+ "learning_rate": 0.0005637344846195358,
851
+ "loss": 4.0175,
852
  "step": 5700
853
  },
854
  {
855
+ "epoch": 0.6199460916442049,
856
+ "grad_norm": 0.6238442063331604,
857
+ "learning_rate": 0.0005634106853750674,
858
+ "loss": 4.0245,
859
  "step": 5750
860
  },
861
  {
862
+ "epoch": 0.6253369272237197,
863
+ "grad_norm": 0.8193546533584595,
864
+ "learning_rate": 0.000563086886130599,
865
+ "loss": 3.9917,
866
  "step": 5800
867
  },
868
  {
869
+ "epoch": 0.6307277628032345,
870
+ "grad_norm": 0.729979395866394,
871
+ "learning_rate": 0.0005627630868861305,
872
+ "loss": 3.9921,
873
  "step": 5850
874
  },
875
  {
876
+ "epoch": 0.6361185983827493,
877
+ "grad_norm": 0.5888447165489197,
878
+ "learning_rate": 0.0005624392876416621,
879
+ "loss": 3.9782,
880
  "step": 5900
881
  },
882
  {
883
+ "epoch": 0.6415094339622641,
884
+ "grad_norm": 0.705590009689331,
885
+ "learning_rate": 0.0005621154883971937,
886
+ "loss": 3.9846,
887
  "step": 5950
888
  },
889
  {
890
+ "epoch": 0.6469002695417789,
891
+ "grad_norm": 0.7365871667861938,
892
+ "learning_rate": 0.0005617916891527253,
893
+ "loss": 3.9652,
894
  "step": 6000
895
  },
896
  {
897
+ "epoch": 0.6469002695417789,
898
+ "eval_accuracy": 0.32846506825954697,
899
+ "eval_loss": 3.9196362495422363,
900
+ "eval_runtime": 185.9381,
901
+ "eval_samples_per_second": 96.866,
902
+ "eval_steps_per_second": 6.056,
903
  "step": 6000
904
  },
905
  {
906
+ "epoch": 0.6522911051212938,
907
+ "grad_norm": 0.7241402864456177,
908
+ "learning_rate": 0.0005614678899082568,
909
+ "loss": 3.9989,
910
  "step": 6050
911
  },
912
  {
913
+ "epoch": 0.6576819407008087,
914
+ "grad_norm": 0.6035869121551514,
915
+ "learning_rate": 0.0005611440906637884,
916
+ "loss": 3.9857,
917
  "step": 6100
918
  },
919
  {
920
+ "epoch": 0.6630727762803235,
921
+ "grad_norm": 0.8426706194877625,
922
+ "learning_rate": 0.00056082029141932,
923
+ "loss": 3.9792,
924
  "step": 6150
925
  },
926
  {
927
+ "epoch": 0.6684636118598383,
928
+ "grad_norm": 0.5762391090393066,
929
+ "learning_rate": 0.0005604964921748515,
930
+ "loss": 3.9917,
931
  "step": 6200
932
  },
933
  {
934
+ "epoch": 0.6738544474393531,
935
+ "grad_norm": 0.5884817242622375,
936
+ "learning_rate": 0.000560172692930383,
937
+ "loss": 3.9506,
938
  "step": 6250
939
  },
940
  {
941
+ "epoch": 0.6792452830188679,
942
+ "grad_norm": 0.6221009492874146,
943
+ "learning_rate": 0.0005598488936859147,
944
+ "loss": 3.9624,
945
  "step": 6300
946
  },
947
  {
948
+ "epoch": 0.6846361185983828,
949
+ "grad_norm": 0.7877587080001831,
950
+ "learning_rate": 0.0005595250944414463,
951
+ "loss": 3.9535,
952
  "step": 6350
953
  },
954
  {
955
+ "epoch": 0.6900269541778976,
956
+ "grad_norm": 0.6189385652542114,
957
+ "learning_rate": 0.0005592012951969778,
958
+ "loss": 3.946,
959
  "step": 6400
960
  },
961
  {
962
+ "epoch": 0.6954177897574124,
963
+ "grad_norm": 0.7304018139839172,
964
+ "learning_rate": 0.0005588774959525094,
965
+ "loss": 3.9736,
966
  "step": 6450
967
  },
968
  {
969
+ "epoch": 0.7008086253369272,
970
+ "grad_norm": 0.6186996698379517,
971
+ "learning_rate": 0.000558553696708041,
972
+ "loss": 3.9576,
973
  "step": 6500
974
  },
975
  {
976
+ "epoch": 0.706199460916442,
977
+ "grad_norm": 0.5706896185874939,
978
+ "learning_rate": 0.0005582298974635726,
979
+ "loss": 3.9541,
980
  "step": 6550
981
  },
982
  {
983
+ "epoch": 0.7115902964959568,
984
+ "grad_norm": 0.5761089324951172,
985
+ "learning_rate": 0.0005579060982191041,
986
+ "loss": 3.9519,
987
  "step": 6600
988
  },
989
  {
990
+ "epoch": 0.7169811320754716,
991
+ "grad_norm": Infinity,
992
+ "learning_rate": 0.0005575887749595251,
993
+ "loss": 3.9458,
994
  "step": 6650
995
  },
996
  {
997
+ "epoch": 0.7223719676549866,
998
+ "grad_norm": 0.6021038889884949,
999
+ "learning_rate": 0.0005572649757150566,
1000
+ "loss": 3.9413,
1001
  "step": 6700
1002
  },
1003
  {
1004
+ "epoch": 0.7277628032345014,
1005
+ "grad_norm": 0.7275946140289307,
1006
+ "learning_rate": 0.0005569411764705882,
1007
+ "loss": 3.9426,
1008
  "step": 6750
1009
  },
1010
  {
1011
+ "epoch": 0.7331536388140162,
1012
+ "grad_norm": 0.5942234992980957,
1013
+ "learning_rate": 0.0005566173772261198,
1014
+ "loss": 3.9527,
1015
  "step": 6800
1016
  },
1017
  {
1018
+ "epoch": 0.738544474393531,
1019
+ "grad_norm": 0.5786459445953369,
1020
+ "learning_rate": 0.0005562935779816513,
1021
+ "loss": 3.9367,
1022
  "step": 6850
1023
  },
1024
  {
1025
+ "epoch": 0.7439353099730458,
1026
+ "grad_norm": 0.6273922920227051,
1027
+ "learning_rate": 0.0005559697787371828,
1028
+ "loss": 3.9311,
1029
  "step": 6900
1030
  },
1031
  {
1032
+ "epoch": 0.7493261455525606,
1033
+ "grad_norm": 0.6769258379936218,
1034
+ "learning_rate": 0.0005556459794927145,
1035
+ "loss": 3.9179,
1036
  "step": 6950
1037
  },
1038
  {
1039
+ "epoch": 0.7547169811320755,
1040
+ "grad_norm": 0.5907636284828186,
1041
+ "learning_rate": 0.000555322180248246,
1042
+ "loss": 3.9319,
1043
  "step": 7000
1044
  },
1045
  {
1046
+ "epoch": 0.7547169811320755,
1047
+ "eval_accuracy": 0.33351166893986617,
1048
+ "eval_loss": 3.861981153488159,
1049
+ "eval_runtime": 179.533,
1050
+ "eval_samples_per_second": 100.321,
1051
+ "eval_steps_per_second": 6.272,
1052
  "step": 7000
1053
  },
1054
  {
1055
+ "epoch": 0.7601078167115903,
1056
+ "grad_norm": 0.5700045824050903,
1057
+ "learning_rate": 0.0005549983810037776,
1058
+ "loss": 3.9355,
1059
  "step": 7050
1060
  },
1061
  {
1062
+ "epoch": 0.7654986522911051,
1063
+ "grad_norm": 0.5413492918014526,
1064
+ "learning_rate": 0.0005546745817593091,
1065
+ "loss": 3.9333,
1066
  "step": 7100
1067
  },
1068
  {
1069
+ "epoch": 0.77088948787062,
1070
+ "grad_norm": 0.6539698839187622,
1071
+ "learning_rate": 0.0005543507825148408,
1072
+ "loss": 3.9116,
1073
  "step": 7150
1074
  },
1075
  {
1076
+ "epoch": 0.7762803234501348,
1077
+ "grad_norm": 0.6693196892738342,
1078
+ "learning_rate": 0.0005540269832703723,
1079
+ "loss": 3.92,
1080
  "step": 7200
1081
  },
1082
  {
1083
+ "epoch": 0.7816711590296496,
1084
+ "grad_norm": 0.6446160078048706,
1085
+ "learning_rate": 0.0005537031840259039,
1086
+ "loss": 3.9226,
1087
  "step": 7250
1088
  },
1089
  {
1090
+ "epoch": 0.7870619946091644,
1091
+ "grad_norm": 0.7050884962081909,
1092
+ "learning_rate": 0.0005533793847814354,
1093
+ "loss": 3.9085,
1094
  "step": 7300
1095
  },
1096
  {
1097
+ "epoch": 0.7924528301886793,
1098
+ "grad_norm": 0.6198840141296387,
1099
+ "learning_rate": 0.000553055585536967,
1100
+ "loss": 3.9174,
1101
  "step": 7350
1102
  },
1103
  {
1104
+ "epoch": 0.7978436657681941,
1105
+ "grad_norm": 0.612169623374939,
1106
+ "learning_rate": 0.0005527317862924987,
1107
+ "loss": 3.8806,
1108
  "step": 7400
1109
  },
1110
  {
1111
+ "epoch": 0.8032345013477089,
1112
+ "grad_norm": 0.6216044425964355,
1113
+ "learning_rate": 0.0005524079870480301,
1114
+ "loss": 3.8935,
1115
  "step": 7450
1116
  },
1117
  {
1118
+ "epoch": 0.8086253369272237,
1119
+ "grad_norm": 0.7052186727523804,
1120
+ "learning_rate": 0.0005520841878035618,
1121
+ "loss": 3.9062,
1122
  "step": 7500
1123
  },
1124
  {
1125
+ "epoch": 0.8140161725067385,
1126
+ "grad_norm": 0.6003553867340088,
1127
+ "learning_rate": 0.0005517603885590933,
1128
+ "loss": 3.8995,
1129
  "step": 7550
1130
  },
1131
  {
1132
+ "epoch": 0.8194070080862533,
1133
+ "grad_norm": 0.5882673263549805,
1134
+ "learning_rate": 0.0005514365893146249,
1135
+ "loss": 3.9185,
1136
  "step": 7600
1137
  },
1138
  {
1139
+ "epoch": 0.8247978436657682,
1140
+ "grad_norm": 0.6166543960571289,
1141
+ "learning_rate": 0.0005511127900701564,
1142
+ "loss": 3.9029,
1143
  "step": 7650
1144
  },
1145
  {
1146
+ "epoch": 0.8301886792452831,
1147
+ "grad_norm": 0.6360199451446533,
1148
+ "learning_rate": 0.000550788990825688,
1149
+ "loss": 3.8742,
1150
  "step": 7700
1151
  },
1152
  {
1153
+ "epoch": 0.8355795148247979,
1154
+ "grad_norm": 0.7083805203437805,
1155
+ "learning_rate": 0.0005504651915812196,
1156
+ "loss": 3.8881,
1157
  "step": 7750
1158
  },
1159
  {
1160
+ "epoch": 0.8409703504043127,
1161
+ "grad_norm": 0.580138087272644,
1162
+ "learning_rate": 0.0005501413923367512,
1163
+ "loss": 3.904,
1164
  "step": 7800
1165
  },
1166
  {
1167
+ "epoch": 0.8463611859838275,
1168
+ "grad_norm": 0.6983124017715454,
1169
+ "learning_rate": 0.0005498175930922827,
1170
+ "loss": 3.8882,
1171
  "step": 7850
1172
  },
1173
  {
1174
+ "epoch": 0.8517520215633423,
1175
+ "grad_norm": 0.7539551854133606,
1176
+ "learning_rate": 0.0005494937938478143,
1177
+ "loss": 3.8887,
1178
  "step": 7900
1179
  },
1180
  {
1181
+ "epoch": 0.8571428571428571,
1182
+ "grad_norm": 0.589042603969574,
1183
+ "learning_rate": 0.0005491699946033459,
1184
+ "loss": 3.8886,
1185
  "step": 7950
1186
  },
1187
  {
1188
+ "epoch": 0.862533692722372,
1189
+ "grad_norm": 0.5650387406349182,
1190
+ "learning_rate": 0.0005488461953588775,
1191
+ "loss": 3.8678,
1192
  "step": 8000
1193
  },
1194
  {
1195
+ "epoch": 0.862533692722372,
1196
+ "eval_accuracy": 0.33783116449493084,
1197
+ "eval_loss": 3.815941572189331,
1198
+ "eval_runtime": 179.2962,
1199
+ "eval_samples_per_second": 100.454,
1200
+ "eval_steps_per_second": 6.28,
1201
  "step": 8000
1202
  },
1203
  {
1204
+ "epoch": 0.8679245283018868,
1205
+ "grad_norm": 0.587796151638031,
1206
+ "learning_rate": 0.000548522396114409,
1207
+ "loss": 3.8843,
1208
  "step": 8050
1209
  },
1210
  {
1211
+ "epoch": 0.8733153638814016,
1212
+ "grad_norm": 0.7792789936065674,
1213
+ "learning_rate": 0.0005481985968699406,
1214
+ "loss": 3.8637,
1215
  "step": 8100
1216
  },
1217
  {
1218
+ "epoch": 0.8787061994609164,
1219
+ "grad_norm": 0.5374471545219421,
1220
+ "learning_rate": 0.0005478747976254721,
1221
+ "loss": 3.8739,
1222
  "step": 8150
1223
  },
1224
  {
1225
+ "epoch": 0.8840970350404312,
1226
+ "grad_norm": 0.6031334400177002,
1227
+ "learning_rate": 0.0005475509983810037,
1228
+ "loss": 3.8634,
1229
  "step": 8200
1230
  },
1231
  {
1232
+ "epoch": 0.889487870619946,
1233
+ "grad_norm": 0.63655024766922,
1234
+ "learning_rate": 0.0005472271991365352,
1235
+ "loss": 3.8795,
1236
  "step": 8250
1237
  },
1238
  {
1239
+ "epoch": 0.894878706199461,
1240
+ "grad_norm": 0.5650659799575806,
1241
+ "learning_rate": 0.0005469033998920669,
1242
+ "loss": 3.8526,
1243
  "step": 8300
1244
  },
1245
  {
1246
+ "epoch": 0.9002695417789758,
1247
+ "grad_norm": 0.5938332676887512,
1248
+ "learning_rate": 0.0005465796006475984,
1249
+ "loss": 3.87,
1250
  "step": 8350
1251
  },
1252
  {
1253
+ "epoch": 0.9056603773584906,
1254
+ "grad_norm": 0.6128816604614258,
1255
+ "learning_rate": 0.00054625580140313,
1256
+ "loss": 3.87,
1257
  "step": 8400
1258
  },
1259
  {
1260
+ "epoch": 0.9110512129380054,
1261
+ "grad_norm": 0.6226370930671692,
1262
+ "learning_rate": 0.0005459320021586615,
1263
+ "loss": 3.8402,
1264
  "step": 8450
1265
  },
1266
  {
1267
+ "epoch": 0.9164420485175202,
1268
+ "grad_norm": 0.5840190052986145,
1269
+ "learning_rate": 0.0005456082029141932,
1270
+ "loss": 3.8508,
1271
  "step": 8500
1272
  },
1273
  {
1274
+ "epoch": 0.921832884097035,
1275
+ "grad_norm": 0.6211468577384949,
1276
+ "learning_rate": 0.0005452844036697248,
1277
+ "loss": 3.8534,
1278
  "step": 8550
1279
  },
1280
  {
1281
+ "epoch": 0.9272237196765498,
1282
+ "grad_norm": 0.5937489867210388,
1283
+ "learning_rate": 0.0005449606044252563,
1284
+ "loss": 3.8805,
1285
  "step": 8600
1286
  },
1287
  {
1288
+ "epoch": 0.9326145552560647,
1289
+ "grad_norm": 0.6649389266967773,
1290
+ "learning_rate": 0.0005446368051807879,
1291
+ "loss": 3.8549,
1292
  "step": 8650
1293
  },
1294
  {
1295
+ "epoch": 0.9380053908355795,
1296
+ "grad_norm": 0.5720226764678955,
1297
+ "learning_rate": 0.0005443130059363194,
1298
+ "loss": 3.8637,
1299
  "step": 8700
1300
  },
1301
  {
1302
+ "epoch": 0.9433962264150944,
1303
+ "grad_norm": 0.5358853936195374,
1304
+ "learning_rate": 0.0005439892066918511,
1305
+ "loss": 3.8465,
1306
  "step": 8750
1307
  },
1308
  {
1309
+ "epoch": 0.9487870619946092,
1310
+ "grad_norm": 0.631186842918396,
1311
+ "learning_rate": 0.0005436654074473825,
1312
+ "loss": 3.8466,
1313
  "step": 8800
1314
  },
1315
  {
1316
+ "epoch": 0.954177897574124,
1317
+ "grad_norm": 0.5419930815696716,
1318
+ "learning_rate": 0.0005433416082029142,
1319
+ "loss": 3.8474,
1320
  "step": 8850
1321
  },
1322
  {
1323
+ "epoch": 0.9595687331536388,
1324
+ "grad_norm": 0.661422848701477,
1325
+ "learning_rate": 0.0005430178089584457,
1326
+ "loss": 3.8489,
1327
  "step": 8900
1328
  },
1329
  {
1330
+ "epoch": 0.9649595687331537,
1331
+ "grad_norm": 0.7400330901145935,
1332
+ "learning_rate": 0.0005426940097139773,
1333
+ "loss": 3.8496,
1334
  "step": 8950
1335
  },
1336
  {
1337
+ "epoch": 0.9703504043126685,
1338
+ "grad_norm": 0.6218024492263794,
1339
+ "learning_rate": 0.0005423766864543982,
1340
+ "loss": 3.8659,
1341
  "step": 9000
1342
  },
1343
  {
1344
+ "epoch": 0.9703504043126685,
1345
+ "eval_accuracy": 0.34161282825531386,
1346
+ "eval_loss": 3.7764477729797363,
1347
+ "eval_runtime": 179.4617,
1348
+ "eval_samples_per_second": 100.361,
1349
+ "eval_steps_per_second": 6.274,
1350
  "step": 9000
1351
  },
1352
  {
1353
+ "epoch": 0.9757412398921833,
1354
+ "grad_norm": 0.6361964344978333,
1355
+ "learning_rate": 0.0005420528872099298,
1356
+ "loss": 3.8413,
1357
  "step": 9050
1358
  },
1359
  {
1360
+ "epoch": 0.9811320754716981,
1361
+ "grad_norm": 0.6142780184745789,
1362
+ "learning_rate": 0.0005417290879654613,
1363
+ "loss": 3.8326,
1364
  "step": 9100
1365
  },
1366
  {
1367
+ "epoch": 0.9865229110512129,
1368
+ "grad_norm": 0.6248621344566345,
1369
+ "learning_rate": 0.000541405288720993,
1370
+ "loss": 3.8448,
1371
  "step": 9150
1372
  },
1373
  {
1374
+ "epoch": 0.9919137466307277,
1375
+ "grad_norm": 0.5891006588935852,
1376
+ "learning_rate": 0.0005410814894765245,
1377
+ "loss": 3.8482,
1378
  "step": 9200
1379
  },
1380
  {
1381
+ "epoch": 0.9973045822102425,
1382
+ "grad_norm": 0.5713793635368347,
1383
+ "learning_rate": 0.0005407576902320561,
1384
+ "loss": 3.8387,
1385
  "step": 9250
1386
  },
1387
  {
1388
+ "epoch": 1.0026954177897573,
1389
+ "grad_norm": 0.5939117074012756,
1390
+ "learning_rate": 0.0005404338909875876,
1391
+ "loss": 3.8174,
1392
  "step": 9300
1393
  },
1394
  {
1395
+ "epoch": 1.0080862533692723,
1396
+ "grad_norm": 0.5924010276794434,
1397
+ "learning_rate": 0.0005401100917431192,
1398
+ "loss": 3.7836,
1399
  "step": 9350
1400
  },
1401
  {
1402
+ "epoch": 1.013477088948787,
1403
+ "grad_norm": 0.6352487206459045,
1404
+ "learning_rate": 0.0005397862924986508,
1405
+ "loss": 3.7956,
1406
  "step": 9400
1407
  },
1408
  {
1409
+ "epoch": 1.0188679245283019,
1410
+ "grad_norm": 0.5576720833778381,
1411
+ "learning_rate": 0.0005394624932541824,
1412
+ "loss": 3.7835,
1413
  "step": 9450
1414
  },
1415
  {
1416
+ "epoch": 1.0242587601078168,
1417
+ "grad_norm": 0.560811460018158,
1418
+ "learning_rate": 0.0005391386940097139,
1419
+ "loss": 3.781,
1420
  "step": 9500
1421
  },
1422
  {
1423
+ "epoch": 1.0296495956873315,
1424
+ "grad_norm": 0.5907502174377441,
1425
+ "learning_rate": 0.0005388148947652455,
1426
+ "loss": 3.7925,
1427
  "step": 9550
1428
  },
1429
  {
1430
+ "epoch": 1.0350404312668464,
1431
+ "grad_norm": 0.5446668863296509,
1432
+ "learning_rate": 0.000538491095520777,
1433
+ "loss": 3.7719,
1434
  "step": 9600
1435
  },
1436
  {
1437
+ "epoch": 1.0404312668463611,
1438
+ "grad_norm": 0.5410051941871643,
1439
+ "learning_rate": 0.0005381672962763086,
1440
+ "loss": 3.7634,
1441
  "step": 9650
1442
  },
1443
  {
1444
+ "epoch": 1.045822102425876,
1445
+ "grad_norm": 0.587929368019104,
1446
+ "learning_rate": 0.0005378434970318403,
1447
+ "loss": 3.7742,
1448
  "step": 9700
1449
  },
1450
  {
1451
+ "epoch": 1.0512129380053907,
1452
+ "grad_norm": 0.609221339225769,
1453
+ "learning_rate": 0.0005375196977873718,
1454
+ "loss": 3.7665,
1455
  "step": 9750
1456
  },
1457
  {
1458
+ "epoch": 1.0566037735849056,
1459
+ "grad_norm": 0.6146703958511353,
1460
+ "learning_rate": 0.0005371958985429034,
1461
+ "loss": 3.7712,
1462
  "step": 9800
1463
  },
1464
  {
1465
+ "epoch": 1.0619946091644206,
1466
+ "grad_norm": 0.6446210145950317,
1467
+ "learning_rate": 0.0005368720992984349,
1468
+ "loss": 3.7547,
1469
  "step": 9850
1470
  },
1471
  {
1472
+ "epoch": 1.0673854447439353,
1473
+ "grad_norm": 0.6286680102348328,
1474
+ "learning_rate": 0.0005365483000539665,
1475
+ "loss": 3.7662,
1476
  "step": 9900
1477
  },
1478
  {
1479
+ "epoch": 1.0727762803234502,
1480
+ "grad_norm": 0.5983867049217224,
1481
+ "learning_rate": 0.0005362245008094981,
1482
+ "loss": 3.7644,
1483
  "step": 9950
1484
  },
1485
  {
1486
+ "epoch": 1.0781671159029649,
1487
+ "grad_norm": 0.622414767742157,
1488
+ "learning_rate": 0.0005359007015650297,
1489
+ "loss": 3.7684,
1490
  "step": 10000
1491
  },
1492
  {
1493
+ "epoch": 1.0781671159029649,
1494
+ "eval_accuracy": 0.3455192255309079,
1495
+ "eval_loss": 3.7473487854003906,
1496
+ "eval_runtime": 179.6608,
1497
+ "eval_samples_per_second": 100.25,
1498
+ "eval_steps_per_second": 6.267,
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:cfee9d1ca2636462af1f5dd1f5348dda3ebb5ebff6f794ebb5861a958cddc3d9
3
  size 5304
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:17e61b03de797e84cad5a5a1386a832b5d860e232565c820b1f9056f83f9e849
3
  size 5304