kevinoli commited on
Commit
f250f62
1 Parent(s): 3eaab21

Training in progress, step 16500, checkpoint

Browse files
checkpoint-16500/model.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:93d3ae5ea6bef794e51e17ea9f34f74005dc83d8307758a884be3eee2caed22a
3
  size 1711848436
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f43e1a379ee775f738f220107e5fb8a5771d977ec550eaafb7e350fc388d4fff
3
  size 1711848436
checkpoint-16500/optimizer.pt CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:5cfa9c06ddff11f1b47320d8e2d1c889d53bf76e522b441c64d8b8dffcf45f5d
3
  size 3424043887
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b716149a2b4038bfd649d0d9eb6b4bee9a4c5ae65b845c779d9d0a27d4575611
3
  size 3424043887
checkpoint-16500/rng_state.pth CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:f53870bad59ab7e8d2128ad156d2adcc3ce9798b4f3803ead355cb0eea68ae55
3
  size 14503
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f1c45c04b203caa5e64a0fe73a973b302ef17143590e9d4304433b96f734879b
3
  size 14503
checkpoint-16500/scheduler.pt CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:a504fb2a5adc1209f3fad3e226d3df1263af4b43ee9d44f1f404989e7e40c60d
3
  size 623
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f6e22603362cfa982e3758658e9f46c33c703572bab8b03143d3fcd63b97b691
3
  size 623
checkpoint-16500/trainer_state.json CHANGED
@@ -1,7 +1,7 @@
1
  {
2
- "best_metric": 0.8776962757110596,
3
  "best_model_checkpoint": "./output/clip-finetuned-csu-p14-336-e4l58-l/checkpoint-16500",
4
- "epoch": 3.0397936624907884,
5
  "eval_steps": 500,
6
  "global_step": 16500,
7
  "is_hyper_param_search": false,
@@ -9,503 +9,503 @@
9
  "is_world_process_zero": true,
10
  "log_history": [
11
  {
12
- "epoch": 0.09211495946941783,
13
- "grad_norm": 20.04043960571289,
14
- "learning_rate": 4.8848563006632274e-08,
15
- "loss": 0.3758,
16
  "step": 500
17
  },
18
  {
19
- "epoch": 0.09211495946941783,
20
- "eval_loss": 1.4184927940368652,
21
- "eval_runtime": 74.3562,
22
- "eval_samples_per_second": 16.233,
23
- "eval_steps_per_second": 2.031,
24
  "step": 500
25
  },
26
  {
27
- "epoch": 0.18422991893883567,
28
- "grad_norm": 74.25025939941406,
29
- "learning_rate": 4.7697126013264556e-08,
30
- "loss": 0.4103,
31
  "step": 1000
32
  },
33
  {
34
- "epoch": 0.18422991893883567,
35
- "eval_loss": 1.3500770330429077,
36
- "eval_runtime": 75.7013,
37
- "eval_samples_per_second": 15.944,
38
- "eval_steps_per_second": 1.995,
39
  "step": 1000
40
  },
41
  {
42
- "epoch": 0.2763448784082535,
43
- "grad_norm": 0.5102410912513733,
44
- "learning_rate": 4.6545689019896826e-08,
45
- "loss": 0.433,
46
  "step": 1500
47
  },
48
  {
49
- "epoch": 0.2763448784082535,
50
- "eval_loss": 1.2885024547576904,
51
- "eval_runtime": 75.197,
52
- "eval_samples_per_second": 16.051,
53
- "eval_steps_per_second": 2.008,
54
  "step": 1500
55
  },
56
  {
57
- "epoch": 0.36845983787767134,
58
- "grad_norm": 0.1667678952217102,
59
- "learning_rate": 4.539425202652911e-08,
60
- "loss": 0.3424,
61
  "step": 2000
62
  },
63
  {
64
- "epoch": 0.36845983787767134,
65
- "eval_loss": 1.239119052886963,
66
- "eval_runtime": 75.387,
67
- "eval_samples_per_second": 16.011,
68
- "eval_steps_per_second": 2.003,
69
  "step": 2000
70
  },
71
  {
72
- "epoch": 0.46057479734708917,
73
- "grad_norm": 343.8175048828125,
74
- "learning_rate": 4.4242815033161385e-08,
75
- "loss": 0.3645,
76
  "step": 2500
77
  },
78
  {
79
- "epoch": 0.46057479734708917,
80
- "eval_loss": 1.1902339458465576,
81
- "eval_runtime": 75.9899,
82
- "eval_samples_per_second": 15.884,
83
- "eval_steps_per_second": 1.987,
84
  "step": 2500
85
  },
86
  {
87
- "epoch": 0.552689756816507,
88
- "grad_norm": 305.5626525878906,
89
- "learning_rate": 4.309137803979366e-08,
90
- "loss": 0.3172,
91
  "step": 3000
92
  },
93
  {
94
- "epoch": 0.552689756816507,
95
- "eval_loss": 1.1506118774414062,
96
- "eval_runtime": 74.8655,
97
- "eval_samples_per_second": 16.122,
98
- "eval_steps_per_second": 2.017,
99
  "step": 3000
100
  },
101
  {
102
- "epoch": 0.6448047162859248,
103
- "grad_norm": 65.59048461914062,
104
- "learning_rate": 4.193994104642594e-08,
105
- "loss": 0.2751,
106
  "step": 3500
107
  },
108
  {
109
- "epoch": 0.6448047162859248,
110
- "eval_loss": 1.1169357299804688,
111
- "eval_runtime": 76.2936,
112
- "eval_samples_per_second": 15.82,
113
- "eval_steps_per_second": 1.979,
114
  "step": 3500
115
  },
116
  {
117
- "epoch": 0.7369196757553427,
118
- "grad_norm": 18.106201171875,
119
- "learning_rate": 4.078850405305821e-08,
120
- "loss": 0.2919,
121
  "step": 4000
122
  },
123
  {
124
- "epoch": 0.7369196757553427,
125
- "eval_loss": 1.0920671224594116,
126
- "eval_runtime": 74.9101,
127
- "eval_samples_per_second": 16.113,
128
- "eval_steps_per_second": 2.016,
129
  "step": 4000
130
  },
131
  {
132
- "epoch": 0.8290346352247605,
133
- "grad_norm": 506.40570068359375,
134
- "learning_rate": 3.9637067059690496e-08,
135
- "loss": 0.2583,
136
  "step": 4500
137
  },
138
  {
139
- "epoch": 0.8290346352247605,
140
- "eval_loss": 1.0721209049224854,
141
- "eval_runtime": 75.8053,
142
- "eval_samples_per_second": 15.922,
143
- "eval_steps_per_second": 1.992,
144
  "step": 4500
145
  },
146
  {
147
- "epoch": 0.9211495946941783,
148
- "grad_norm": 1.1508910655975342,
149
- "learning_rate": 3.848563006632277e-08,
150
- "loss": 0.2679,
151
  "step": 5000
152
  },
153
  {
154
- "epoch": 0.9211495946941783,
155
- "eval_loss": 1.0519349575042725,
156
- "eval_runtime": 75.226,
157
- "eval_samples_per_second": 16.045,
158
- "eval_steps_per_second": 2.007,
159
  "step": 5000
160
  },
161
  {
162
- "epoch": 1.013264554163596,
163
- "grad_norm": 0.1393290013074875,
164
- "learning_rate": 3.733419307295505e-08,
165
- "loss": 0.2472,
166
  "step": 5500
167
  },
168
  {
169
- "epoch": 1.013264554163596,
170
- "eval_loss": 1.0355629920959473,
171
- "eval_runtime": 74.6535,
172
- "eval_samples_per_second": 16.168,
173
- "eval_steps_per_second": 2.023,
174
  "step": 5500
175
  },
176
  {
177
- "epoch": 1.105379513633014,
178
- "grad_norm": 200.34664916992188,
179
- "learning_rate": 3.6182756079587324e-08,
180
- "loss": 0.26,
181
  "step": 6000
182
  },
183
  {
184
- "epoch": 1.105379513633014,
185
- "eval_loss": 1.0176944732666016,
186
- "eval_runtime": 74.4374,
187
- "eval_samples_per_second": 16.215,
188
- "eval_steps_per_second": 2.029,
189
  "step": 6000
190
  },
191
  {
192
- "epoch": 1.1974944731024317,
193
- "grad_norm": 431.1804504394531,
194
- "learning_rate": 3.50313190862196e-08,
195
- "loss": 0.2153,
196
  "step": 6500
197
  },
198
  {
199
- "epoch": 1.1974944731024317,
200
- "eval_loss": 1.0044586658477783,
201
- "eval_runtime": 74.3247,
202
- "eval_samples_per_second": 16.24,
203
- "eval_steps_per_second": 2.032,
204
  "step": 6500
205
  },
206
  {
207
- "epoch": 1.2896094325718497,
208
- "grad_norm": 0.6204941272735596,
209
- "learning_rate": 3.387988209285188e-08,
210
- "loss": 0.1791,
211
  "step": 7000
212
  },
213
  {
214
- "epoch": 1.2896094325718497,
215
- "eval_loss": 0.9926707148551941,
216
- "eval_runtime": 75.7317,
217
- "eval_samples_per_second": 15.938,
218
- "eval_steps_per_second": 1.994,
219
  "step": 7000
220
  },
221
  {
222
- "epoch": 1.3817243920412676,
223
- "grad_norm": 52.27273178100586,
224
- "learning_rate": 3.272844509948415e-08,
225
- "loss": 0.2082,
226
  "step": 7500
227
  },
228
  {
229
- "epoch": 1.3817243920412676,
230
- "eval_loss": 0.9803809523582458,
231
- "eval_runtime": 74.9936,
232
- "eval_samples_per_second": 16.095,
233
- "eval_steps_per_second": 2.014,
234
  "step": 7500
235
  },
236
  {
237
- "epoch": 1.4738393515106853,
238
- "grad_norm": 0.0976635068655014,
239
- "learning_rate": 3.1577008106116435e-08,
240
- "loss": 0.196,
241
  "step": 8000
242
  },
243
  {
244
- "epoch": 1.4738393515106853,
245
- "eval_loss": 0.9711655378341675,
246
- "eval_runtime": 74.3537,
247
- "eval_samples_per_second": 16.233,
248
- "eval_steps_per_second": 2.031,
249
  "step": 8000
250
  },
251
  {
252
- "epoch": 1.565954310980103,
253
- "grad_norm": 46.64162826538086,
254
- "learning_rate": 3.042557111274871e-08,
255
- "loss": 0.1946,
256
  "step": 8500
257
  },
258
  {
259
- "epoch": 1.565954310980103,
260
- "eval_loss": 0.9621181488037109,
261
- "eval_runtime": 75.7205,
262
- "eval_samples_per_second": 15.94,
263
- "eval_steps_per_second": 1.994,
264
  "step": 8500
265
  },
266
  {
267
- "epoch": 1.658069270449521,
268
- "grad_norm": 0.05690698325634003,
269
- "learning_rate": 2.927413411938099e-08,
270
- "loss": 0.2422,
271
  "step": 9000
272
  },
273
  {
274
- "epoch": 1.658069270449521,
275
- "eval_loss": 0.9536784291267395,
276
- "eval_runtime": 75.1745,
277
- "eval_samples_per_second": 16.056,
278
- "eval_steps_per_second": 2.009,
279
  "step": 9000
280
  },
281
  {
282
- "epoch": 1.750184229918939,
283
- "grad_norm": 0.08470064401626587,
284
- "learning_rate": 2.8122697126013263e-08,
285
- "loss": 0.2106,
286
  "step": 9500
287
  },
288
  {
289
- "epoch": 1.750184229918939,
290
- "eval_loss": 0.9457550644874573,
291
- "eval_runtime": 74.5062,
292
- "eval_samples_per_second": 16.2,
293
- "eval_steps_per_second": 2.027,
294
  "step": 9500
295
  },
296
  {
297
- "epoch": 1.8422991893883567,
298
- "grad_norm": 1.5745967626571655,
299
- "learning_rate": 2.697126013264554e-08,
300
- "loss": 0.1801,
301
  "step": 10000
302
  },
303
  {
304
- "epoch": 1.8422991893883567,
305
- "eval_loss": 0.9393072724342346,
306
- "eval_runtime": 75.9432,
307
- "eval_samples_per_second": 15.893,
308
- "eval_steps_per_second": 1.988,
309
  "step": 10000
310
  },
311
  {
312
- "epoch": 1.9344141488577744,
313
- "grad_norm": 0.06661002337932587,
314
- "learning_rate": 2.5819823139277818e-08,
315
- "loss": 0.2117,
316
  "step": 10500
317
  },
318
  {
319
- "epoch": 1.9344141488577744,
320
- "eval_loss": 0.9308408498764038,
321
- "eval_runtime": 75.0811,
322
- "eval_samples_per_second": 16.076,
323
- "eval_steps_per_second": 2.011,
324
  "step": 10500
325
  },
326
  {
327
- "epoch": 2.026529108327192,
328
- "grad_norm": 0.236691415309906,
329
- "learning_rate": 2.4668386145910094e-08,
330
- "loss": 0.2061,
331
  "step": 11000
332
  },
333
  {
334
- "epoch": 2.026529108327192,
335
- "eval_loss": 0.9237484335899353,
336
- "eval_runtime": 74.6134,
337
- "eval_samples_per_second": 16.177,
338
- "eval_steps_per_second": 2.024,
339
  "step": 11000
340
  },
341
  {
342
- "epoch": 2.1186440677966103,
343
- "grad_norm": 8.831416130065918,
344
- "learning_rate": 2.3516949152542374e-08,
345
- "loss": 0.1878,
346
  "step": 11500
347
  },
348
  {
349
- "epoch": 2.1186440677966103,
350
- "eval_loss": 0.9166750311851501,
351
- "eval_runtime": 75.6886,
352
- "eval_samples_per_second": 15.947,
353
- "eval_steps_per_second": 1.995,
354
  "step": 11500
355
  },
356
  {
357
- "epoch": 2.210759027266028,
358
- "grad_norm": 0.787421464920044,
359
- "learning_rate": 2.236551215917465e-08,
360
- "loss": 0.1655,
361
  "step": 12000
362
  },
363
  {
364
- "epoch": 2.210759027266028,
365
- "eval_loss": 0.9109458327293396,
366
- "eval_runtime": 74.9866,
367
- "eval_samples_per_second": 16.096,
368
- "eval_steps_per_second": 2.014,
369
  "step": 12000
370
  },
371
  {
372
- "epoch": 2.3028739867354457,
373
- "grad_norm": 535.4310913085938,
374
- "learning_rate": 2.1214075165806926e-08,
375
- "loss": 0.1946,
376
  "step": 12500
377
  },
378
  {
379
- "epoch": 2.3028739867354457,
380
- "eval_loss": 0.9070548415184021,
381
- "eval_runtime": 74.3796,
382
- "eval_samples_per_second": 16.228,
383
- "eval_steps_per_second": 2.03,
384
  "step": 12500
385
  },
386
  {
387
- "epoch": 2.3949889462048635,
388
- "grad_norm": 0.8420014381408691,
389
- "learning_rate": 2.0062638172439202e-08,
390
- "loss": 0.1882,
391
  "step": 13000
392
  },
393
  {
394
- "epoch": 2.3949889462048635,
395
- "eval_loss": 0.9021016359329224,
396
- "eval_runtime": 75.8812,
397
- "eval_samples_per_second": 15.906,
398
- "eval_steps_per_second": 1.99,
399
  "step": 13000
400
  },
401
  {
402
- "epoch": 2.4871039056742816,
403
- "grad_norm": 1.4041686058044434,
404
- "learning_rate": 1.891120117907148e-08,
405
- "loss": 0.1871,
406
  "step": 13500
407
  },
408
  {
409
- "epoch": 2.4871039056742816,
410
- "eval_loss": 0.8960274457931519,
411
- "eval_runtime": 75.2283,
412
- "eval_samples_per_second": 16.044,
413
- "eval_steps_per_second": 2.007,
414
  "step": 13500
415
  },
416
  {
417
- "epoch": 2.5792188651436994,
418
- "grad_norm": 327.2938537597656,
419
- "learning_rate": 1.7759764185703757e-08,
420
- "loss": 0.1419,
421
  "step": 14000
422
  },
423
  {
424
- "epoch": 2.5792188651436994,
425
- "eval_loss": 0.8913142085075378,
426
- "eval_runtime": 74.4503,
427
- "eval_samples_per_second": 16.212,
428
- "eval_steps_per_second": 2.028,
429
  "step": 14000
430
  },
431
  {
432
- "epoch": 2.671333824613117,
433
- "grad_norm": 0.01779676228761673,
434
- "learning_rate": 1.6608327192336037e-08,
435
- "loss": 0.1431,
436
  "step": 14500
437
  },
438
  {
439
- "epoch": 2.671333824613117,
440
- "eval_loss": 0.8879275321960449,
441
- "eval_runtime": 75.8636,
442
- "eval_samples_per_second": 15.91,
443
- "eval_steps_per_second": 1.99,
444
  "step": 14500
445
  },
446
  {
447
- "epoch": 2.7634487840825352,
448
- "grad_norm": 1.6428073644638062,
449
- "learning_rate": 1.5456890198968313e-08,
450
- "loss": 0.1811,
451
  "step": 15000
452
  },
453
  {
454
- "epoch": 2.7634487840825352,
455
- "eval_loss": 0.8848333954811096,
456
- "eval_runtime": 75.1911,
457
- "eval_samples_per_second": 16.052,
458
- "eval_steps_per_second": 2.008,
459
  "step": 15000
460
  },
461
  {
462
- "epoch": 2.855563743551953,
463
- "grad_norm": 0.0006026208866387606,
464
- "learning_rate": 1.430545320560059e-08,
465
- "loss": 0.1694,
466
  "step": 15500
467
  },
468
  {
469
- "epoch": 2.855563743551953,
470
- "eval_loss": 0.8826532959938049,
471
- "eval_runtime": 74.6044,
472
- "eval_samples_per_second": 16.179,
473
- "eval_steps_per_second": 2.024,
474
  "step": 15500
475
  },
476
  {
477
- "epoch": 2.9476787030213707,
478
- "grad_norm": 226.09982299804688,
479
- "learning_rate": 1.3154016212232866e-08,
480
- "loss": 0.1718,
481
  "step": 16000
482
  },
483
  {
484
- "epoch": 2.9476787030213707,
485
- "eval_loss": 0.8798421621322632,
486
- "eval_runtime": 76.0825,
487
- "eval_samples_per_second": 15.864,
488
- "eval_steps_per_second": 1.985,
489
  "step": 16000
490
  },
491
  {
492
- "epoch": 3.0397936624907884,
493
- "grad_norm": 5.1134706154698506e-05,
494
- "learning_rate": 1.2002579218865142e-08,
495
- "loss": 0.153,
496
  "step": 16500
497
  },
498
  {
499
- "epoch": 3.0397936624907884,
500
- "eval_loss": 0.8776962757110596,
501
- "eval_runtime": 75.3205,
502
- "eval_samples_per_second": 16.025,
503
- "eval_steps_per_second": 2.005,
504
  "step": 16500
505
  }
506
  ],
507
  "logging_steps": 500,
508
- "max_steps": 21712,
509
  "num_input_tokens_seen": 0,
510
  "num_train_epochs": 4,
511
  "save_steps": 500,
 
1
  {
2
+ "best_metric": 0.8900338411331177,
3
  "best_model_checkpoint": "./output/clip-finetuned-csu-p14-336-e4l58-l/checkpoint-16500",
4
+ "epoch": 1.75981228668942,
5
  "eval_steps": 500,
6
  "global_step": 16500,
7
  "is_hyper_param_search": false,
 
9
  "is_world_process_zero": true,
10
  "log_history": [
11
  {
12
+ "epoch": 0.05332764505119454,
13
+ "grad_norm": 217.55360412597656,
14
+ "learning_rate": 4.933340443686007e-08,
15
+ "loss": 0.4667,
16
  "step": 500
17
  },
18
  {
19
+ "epoch": 0.05332764505119454,
20
+ "eval_loss": 1.4426143169403076,
21
+ "eval_runtime": 61.9901,
22
+ "eval_samples_per_second": 15.922,
23
+ "eval_steps_per_second": 2.0,
24
  "step": 500
25
  },
26
  {
27
+ "epoch": 0.10665529010238908,
28
+ "grad_norm": 73.73340606689453,
29
+ "learning_rate": 4.8666808873720136e-08,
30
+ "loss": 0.4532,
31
  "step": 1000
32
  },
33
  {
34
+ "epoch": 0.10665529010238908,
35
+ "eval_loss": 1.3815597295761108,
36
+ "eval_runtime": 62.4624,
37
+ "eval_samples_per_second": 15.801,
38
+ "eval_steps_per_second": 1.985,
39
  "step": 1000
40
  },
41
  {
42
+ "epoch": 0.1599829351535836,
43
+ "grad_norm": 471.4011535644531,
44
+ "learning_rate": 4.80002133105802e-08,
45
+ "loss": 0.3749,
46
  "step": 1500
47
  },
48
  {
49
+ "epoch": 0.1599829351535836,
50
+ "eval_loss": 1.3310539722442627,
51
+ "eval_runtime": 63.3918,
52
+ "eval_samples_per_second": 15.57,
53
+ "eval_steps_per_second": 1.956,
54
  "step": 1500
55
  },
56
  {
57
+ "epoch": 0.21331058020477817,
58
+ "grad_norm": 11.962693214416504,
59
+ "learning_rate": 4.733361774744027e-08,
60
+ "loss": 0.336,
61
  "step": 2000
62
  },
63
  {
64
+ "epoch": 0.21331058020477817,
65
+ "eval_loss": 1.2890639305114746,
66
+ "eval_runtime": 63.177,
67
+ "eval_samples_per_second": 15.623,
68
+ "eval_steps_per_second": 1.963,
69
  "step": 2000
70
  },
71
  {
72
+ "epoch": 0.2666382252559727,
73
+ "grad_norm": 375.1609191894531,
74
+ "learning_rate": 4.666702218430034e-08,
75
+ "loss": 0.3585,
76
  "step": 2500
77
  },
78
  {
79
+ "epoch": 0.2666382252559727,
80
+ "eval_loss": 1.2536433935165405,
81
+ "eval_runtime": 63.1512,
82
+ "eval_samples_per_second": 15.629,
83
+ "eval_steps_per_second": 1.964,
84
  "step": 2500
85
  },
86
  {
87
+ "epoch": 0.3199658703071672,
88
+ "grad_norm": 0.0022936267778277397,
89
+ "learning_rate": 4.600042662116041e-08,
90
+ "loss": 0.303,
91
  "step": 3000
92
  },
93
  {
94
+ "epoch": 0.3199658703071672,
95
+ "eval_loss": 1.2202869653701782,
96
+ "eval_runtime": 63.5857,
97
+ "eval_samples_per_second": 15.522,
98
+ "eval_steps_per_second": 1.95,
99
  "step": 3000
100
  },
101
  {
102
+ "epoch": 0.37329351535836175,
103
+ "grad_norm": 15.805087089538574,
104
+ "learning_rate": 4.5333831058020476e-08,
105
+ "loss": 0.3242,
106
  "step": 3500
107
  },
108
  {
109
+ "epoch": 0.37329351535836175,
110
+ "eval_loss": 1.195627212524414,
111
+ "eval_runtime": 63.4024,
112
+ "eval_samples_per_second": 15.567,
113
+ "eval_steps_per_second": 1.956,
114
  "step": 3500
115
  },
116
  {
117
+ "epoch": 0.42662116040955633,
118
+ "grad_norm": 5.377908229827881,
119
+ "learning_rate": 4.4667235494880546e-08,
120
+ "loss": 0.2427,
121
  "step": 4000
122
  },
123
  {
124
+ "epoch": 0.42662116040955633,
125
+ "eval_loss": 1.169384241104126,
126
+ "eval_runtime": 63.5082,
127
+ "eval_samples_per_second": 15.541,
128
+ "eval_steps_per_second": 1.953,
129
  "step": 4000
130
  },
131
  {
132
+ "epoch": 0.47994880546075086,
133
+ "grad_norm": 150.5334930419922,
134
+ "learning_rate": 4.4000639931740615e-08,
135
+ "loss": 0.2993,
136
  "step": 4500
137
  },
138
  {
139
+ "epoch": 0.47994880546075086,
140
+ "eval_loss": 1.145558476448059,
141
+ "eval_runtime": 62.3888,
142
+ "eval_samples_per_second": 15.82,
143
+ "eval_steps_per_second": 1.988,
144
  "step": 4500
145
  },
146
  {
147
+ "epoch": 0.5332764505119454,
148
+ "grad_norm": 0.12246542423963547,
149
+ "learning_rate": 4.333404436860068e-08,
150
+ "loss": 0.3183,
151
  "step": 5000
152
  },
153
  {
154
+ "epoch": 0.5332764505119454,
155
+ "eval_loss": 1.1201218366622925,
156
+ "eval_runtime": 63.7154,
157
+ "eval_samples_per_second": 15.491,
158
+ "eval_steps_per_second": 1.946,
159
  "step": 5000
160
  },
161
  {
162
+ "epoch": 0.58660409556314,
163
+ "grad_norm": 0.3306196630001068,
164
+ "learning_rate": 4.266744880546075e-08,
165
+ "loss": 0.307,
166
  "step": 5500
167
  },
168
  {
169
+ "epoch": 0.58660409556314,
170
+ "eval_loss": 1.0982334613800049,
171
+ "eval_runtime": 62.4882,
172
+ "eval_samples_per_second": 15.795,
173
+ "eval_steps_per_second": 1.984,
174
  "step": 5500
175
  },
176
  {
177
+ "epoch": 0.6399317406143344,
178
+ "grad_norm": 149.38189697265625,
179
+ "learning_rate": 4.2000853242320816e-08,
180
+ "loss": 0.2638,
181
  "step": 6000
182
  },
183
  {
184
+ "epoch": 0.6399317406143344,
185
+ "eval_loss": 1.0779541730880737,
186
+ "eval_runtime": 63.4101,
187
+ "eval_samples_per_second": 15.565,
188
+ "eval_steps_per_second": 1.956,
189
  "step": 6000
190
  },
191
  {
192
+ "epoch": 0.693259385665529,
193
+ "grad_norm": 353.0036315917969,
194
+ "learning_rate": 4.1334257679180886e-08,
195
+ "loss": 0.2226,
196
  "step": 6500
197
  },
198
  {
199
+ "epoch": 0.693259385665529,
200
+ "eval_loss": 1.0613036155700684,
201
+ "eval_runtime": 62.8292,
202
+ "eval_samples_per_second": 15.709,
203
+ "eval_steps_per_second": 1.974,
204
  "step": 6500
205
  },
206
  {
207
+ "epoch": 0.7465870307167235,
208
+ "grad_norm": 0.13188917934894562,
209
+ "learning_rate": 4.0667662116040955e-08,
210
+ "loss": 0.2453,
211
  "step": 7000
212
  },
213
  {
214
+ "epoch": 0.7465870307167235,
215
+ "eval_loss": 1.0443507432937622,
216
+ "eval_runtime": 62.7457,
217
+ "eval_samples_per_second": 15.73,
218
+ "eval_steps_per_second": 1.976,
219
  "step": 7000
220
  },
221
  {
222
+ "epoch": 0.7999146757679181,
223
+ "grad_norm": 59.391639709472656,
224
+ "learning_rate": 4.000106655290102e-08,
225
+ "loss": 0.272,
226
  "step": 7500
227
  },
228
  {
229
+ "epoch": 0.7999146757679181,
230
+ "eval_loss": 1.0300519466400146,
231
+ "eval_runtime": 62.3719,
232
+ "eval_samples_per_second": 15.824,
233
+ "eval_steps_per_second": 1.988,
234
  "step": 7500
235
  },
236
  {
237
+ "epoch": 0.8532423208191127,
238
+ "grad_norm": 495.4797668457031,
239
+ "learning_rate": 3.9334470989761093e-08,
240
+ "loss": 0.283,
241
  "step": 8000
242
  },
243
  {
244
+ "epoch": 0.8532423208191127,
245
+ "eval_loss": 1.0166898965835571,
246
+ "eval_runtime": 62.0338,
247
+ "eval_samples_per_second": 15.911,
248
+ "eval_steps_per_second": 1.999,
249
  "step": 8000
250
  },
251
  {
252
+ "epoch": 0.9065699658703071,
253
+ "grad_norm": 159.0310821533203,
254
+ "learning_rate": 3.866787542662116e-08,
255
+ "loss": 0.2331,
256
  "step": 8500
257
  },
258
  {
259
+ "epoch": 0.9065699658703071,
260
+ "eval_loss": 1.003544807434082,
261
+ "eval_runtime": 62.9704,
262
+ "eval_samples_per_second": 15.674,
263
+ "eval_steps_per_second": 1.969,
264
  "step": 8500
265
  },
266
  {
267
+ "epoch": 0.9598976109215017,
268
+ "grad_norm": 45.40327835083008,
269
+ "learning_rate": 3.8001279863481225e-08,
270
+ "loss": 0.2362,
271
  "step": 9000
272
  },
273
  {
274
+ "epoch": 0.9598976109215017,
275
+ "eval_loss": 0.992453396320343,
276
+ "eval_runtime": 63.1405,
277
+ "eval_samples_per_second": 15.632,
278
+ "eval_steps_per_second": 1.964,
279
  "step": 9000
280
  },
281
  {
282
+ "epoch": 1.0132252559726962,
283
+ "grad_norm": 621.8538818359375,
284
+ "learning_rate": 3.7334684300341295e-08,
285
+ "loss": 0.2396,
286
  "step": 9500
287
  },
288
  {
289
+ "epoch": 1.0132252559726962,
290
+ "eval_loss": 0.9829725027084351,
291
+ "eval_runtime": 62.3244,
292
+ "eval_samples_per_second": 15.836,
293
+ "eval_steps_per_second": 1.99,
294
  "step": 9500
295
  },
296
  {
297
+ "epoch": 1.0665529010238908,
298
+ "grad_norm": 461.60736083984375,
299
+ "learning_rate": 3.6668088737201364e-08,
300
+ "loss": 0.2013,
301
  "step": 10000
302
  },
303
  {
304
+ "epoch": 1.0665529010238908,
305
+ "eval_loss": 0.9735796451568604,
306
+ "eval_runtime": 63.0763,
307
+ "eval_samples_per_second": 15.648,
308
+ "eval_steps_per_second": 1.966,
309
  "step": 10000
310
  },
311
  {
312
+ "epoch": 1.1198805460750854,
313
+ "grad_norm": 7.0309906005859375,
314
+ "learning_rate": 3.6001493174061433e-08,
315
+ "loss": 0.2082,
316
  "step": 10500
317
  },
318
  {
319
+ "epoch": 1.1198805460750854,
320
+ "eval_loss": 0.9639086127281189,
321
+ "eval_runtime": 63.0238,
322
+ "eval_samples_per_second": 15.661,
323
+ "eval_steps_per_second": 1.968,
324
  "step": 10500
325
  },
326
  {
327
+ "epoch": 1.17320819112628,
328
+ "grad_norm": 304.01995849609375,
329
+ "learning_rate": 3.5334897610921496e-08,
330
+ "loss": 0.2023,
331
  "step": 11000
332
  },
333
  {
334
+ "epoch": 1.17320819112628,
335
+ "eval_loss": 0.9558469653129578,
336
+ "eval_runtime": 62.1866,
337
+ "eval_samples_per_second": 15.872,
338
+ "eval_steps_per_second": 1.994,
339
  "step": 11000
340
  },
341
  {
342
+ "epoch": 1.2265358361774745,
343
+ "grad_norm": 0.0017981340643018484,
344
+ "learning_rate": 3.4668302047781565e-08,
345
+ "loss": 0.2331,
346
  "step": 11500
347
  },
348
  {
349
+ "epoch": 1.2265358361774745,
350
+ "eval_loss": 0.9465070962905884,
351
+ "eval_runtime": 63.1547,
352
+ "eval_samples_per_second": 15.628,
353
+ "eval_steps_per_second": 1.963,
354
  "step": 11500
355
  },
356
  {
357
+ "epoch": 1.2798634812286689,
358
+ "grad_norm": 0.8147886395454407,
359
+ "learning_rate": 3.4001706484641635e-08,
360
+ "loss": 0.1784,
361
  "step": 12000
362
  },
363
  {
364
+ "epoch": 1.2798634812286689,
365
+ "eval_loss": 0.9392343163490295,
366
+ "eval_runtime": 62.9221,
367
+ "eval_samples_per_second": 15.686,
368
+ "eval_steps_per_second": 1.971,
369
  "step": 12000
370
  },
371
  {
372
+ "epoch": 1.3331911262798635,
373
+ "grad_norm": 0.01276802271604538,
374
+ "learning_rate": 3.3335110921501704e-08,
375
+ "loss": 0.1953,
376
  "step": 12500
377
  },
378
  {
379
+ "epoch": 1.3331911262798635,
380
+ "eval_loss": 0.9316345453262329,
381
+ "eval_runtime": 62.2816,
382
+ "eval_samples_per_second": 15.847,
383
+ "eval_steps_per_second": 1.991,
384
  "step": 12500
385
  },
386
  {
387
+ "epoch": 1.386518771331058,
388
+ "grad_norm": 0.07341516762971878,
389
+ "learning_rate": 3.2668515358361773e-08,
390
+ "loss": 0.1867,
391
  "step": 13000
392
  },
393
  {
394
+ "epoch": 1.386518771331058,
395
+ "eval_loss": 0.9270105957984924,
396
+ "eval_runtime": 63.2619,
397
+ "eval_samples_per_second": 15.602,
398
+ "eval_steps_per_second": 1.96,
399
  "step": 13000
400
  },
401
  {
402
+ "epoch": 1.4398464163822526,
403
+ "grad_norm": 17.55596923828125,
404
+ "learning_rate": 3.200191979522184e-08,
405
+ "loss": 0.22,
406
  "step": 13500
407
  },
408
  {
409
+ "epoch": 1.4398464163822526,
410
+ "eval_loss": 0.919725775718689,
411
+ "eval_runtime": 63.3459,
412
+ "eval_samples_per_second": 15.581,
413
+ "eval_steps_per_second": 1.958,
414
  "step": 13500
415
  },
416
  {
417
+ "epoch": 1.493174061433447,
418
+ "grad_norm": 3.4780771732330322,
419
+ "learning_rate": 3.133532423208191e-08,
420
+ "loss": 0.1656,
421
  "step": 14000
422
  },
423
  {
424
+ "epoch": 1.493174061433447,
425
+ "eval_loss": 0.9148014783859253,
426
+ "eval_runtime": 62.257,
427
+ "eval_samples_per_second": 15.854,
428
+ "eval_steps_per_second": 1.992,
429
  "step": 14000
430
  },
431
  {
432
+ "epoch": 1.5465017064846416,
433
+ "grad_norm": 158.22390747070312,
434
+ "learning_rate": 3.066872866894198e-08,
435
+ "loss": 0.1968,
436
  "step": 14500
437
  },
438
  {
439
+ "epoch": 1.5465017064846416,
440
+ "eval_loss": 0.9096349477767944,
441
+ "eval_runtime": 63.3671,
442
+ "eval_samples_per_second": 15.576,
443
+ "eval_steps_per_second": 1.957,
444
  "step": 14500
445
  },
446
  {
447
+ "epoch": 1.5998293515358362,
448
+ "grad_norm": 621.3922729492188,
449
+ "learning_rate": 3.0002133105802044e-08,
450
+ "loss": 0.1676,
451
  "step": 15000
452
  },
453
  {
454
+ "epoch": 1.5998293515358362,
455
+ "eval_loss": 0.9056846499443054,
456
+ "eval_runtime": 63.2369,
457
+ "eval_samples_per_second": 15.608,
458
+ "eval_steps_per_second": 1.961,
459
  "step": 15000
460
  },
461
  {
462
+ "epoch": 1.6531569965870307,
463
+ "grad_norm": 26.23398208618164,
464
+ "learning_rate": 2.9335537542662113e-08,
465
+ "loss": 0.2074,
466
  "step": 15500
467
  },
468
  {
469
+ "epoch": 1.6531569965870307,
470
+ "eval_loss": 0.8993550539016724,
471
+ "eval_runtime": 62.3767,
472
+ "eval_samples_per_second": 15.823,
473
+ "eval_steps_per_second": 1.988,
474
  "step": 15500
475
  },
476
  {
477
+ "epoch": 1.7064846416382253,
478
+ "grad_norm": 283.3314514160156,
479
+ "learning_rate": 2.8668941979522183e-08,
480
+ "loss": 0.1847,
481
  "step": 16000
482
  },
483
  {
484
+ "epoch": 1.7064846416382253,
485
+ "eval_loss": 0.8954482078552246,
486
+ "eval_runtime": 63.2168,
487
+ "eval_samples_per_second": 15.613,
488
+ "eval_steps_per_second": 1.962,
489
  "step": 16000
490
  },
491
  {
492
+ "epoch": 1.75981228668942,
493
+ "grad_norm": 12.151030540466309,
494
+ "learning_rate": 2.800234641638225e-08,
495
+ "loss": 0.1845,
496
  "step": 16500
497
  },
498
  {
499
+ "epoch": 1.75981228668942,
500
+ "eval_loss": 0.8900338411331177,
501
+ "eval_runtime": 63.5581,
502
+ "eval_samples_per_second": 15.529,
503
+ "eval_steps_per_second": 1.951,
504
  "step": 16500
505
  }
506
  ],
507
  "logging_steps": 500,
508
+ "max_steps": 37504,
509
  "num_input_tokens_seen": 0,
510
  "num_train_epochs": 4,
511
  "save_steps": 500,
checkpoint-16500/training_args.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:9a29c2374e59ce3f57b34db9cb20d2d1bdc870c3dad128e9a4d462636c37e8b0
3
  size 4847
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:87e35dca972e13467c750106c19301b098fbd2918fda1087487e227bac3aca37
3
  size 4847