loris3 commited on
Commit
a4bc489
·
verified ·
1 Parent(s): 2843587

Upload folder using huggingface_hub

Browse files
checkpoints/checkpoint-3132/config.json ADDED
@@ -0,0 +1,30 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "architectures": [
3
+ "LlamaForCausalLM"
4
+ ],
5
+ "attention_bias": false,
6
+ "attention_dropout": 0.0,
7
+ "bos_token_id": 1,
8
+ "eos_token_id": 2,
9
+ "head_dim": 64,
10
+ "hidden_act": "silu",
11
+ "hidden_size": 768,
12
+ "initializer_range": 0.02,
13
+ "intermediate_size": 2048,
14
+ "max_position_embeddings": 256,
15
+ "mlp_bias": false,
16
+ "model_type": "llama",
17
+ "num_attention_heads": 12,
18
+ "num_hidden_layers": 12,
19
+ "num_key_value_heads": 12,
20
+ "pad_token_id": 1,
21
+ "pretraining_tp": 1,
22
+ "rms_norm_eps": 1e-06,
23
+ "rope_scaling": null,
24
+ "rope_theta": 10000.0,
25
+ "tie_word_embeddings": true,
26
+ "torch_dtype": "float32",
27
+ "transformers_version": "4.47.0",
28
+ "use_cache": true,
29
+ "vocab_size": 16000
30
+ }
checkpoints/checkpoint-3132/generation_config.json ADDED
@@ -0,0 +1,7 @@
 
 
 
 
 
 
 
 
1
+ {
2
+ "_from_model_config": true,
3
+ "bos_token_id": 1,
4
+ "eos_token_id": 2,
5
+ "pad_token_id": 1,
6
+ "transformers_version": "4.47.0"
7
+ }
checkpoints/checkpoint-3132/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:cf1eb49da3f7ae1e510a532131f2c05492a784040acd20236181e6d2638de867
3
+ size 388979624
checkpoints/checkpoint-3132/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:992e4db67d00006308f249fde82c43bdf69bb9417d5b587288482b5c0dfbf667
3
+ size 778027770
checkpoints/checkpoint-3132/rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d2319dd9a8b23cecdfd669867c6907e25d7b349c6e32f2a455d59fdeb1efdc47
3
+ size 14244
checkpoints/checkpoint-3132/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:43eb36fcef750b931b702044250c1b0890cc3fd8c4a0c95211cf5a150d13afe0
3
+ size 1064
checkpoints/checkpoint-3132/trainer_state.json ADDED
@@ -0,0 +1,527 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": null,
3
+ "best_model_checkpoint": null,
4
+ "epoch": 5.998804113848362,
5
+ "eval_steps": 500,
6
+ "global_step": 3132,
7
+ "is_hyper_param_search": false,
8
+ "is_local_process_zero": true,
9
+ "is_world_process_zero": true,
10
+ "log_history": [
11
+ {
12
+ "epoch": 0.09567089213106912,
13
+ "grad_norm": 299.7386474609375,
14
+ "learning_rate": 0.00011666666666666667,
15
+ "loss": 129.9798,
16
+ "step": 50
17
+ },
18
+ {
19
+ "epoch": 0.19134178426213824,
20
+ "grad_norm": 28.020761489868164,
21
+ "learning_rate": 0.00023333333333333333,
22
+ "loss": 80.3567,
23
+ "step": 100
24
+ },
25
+ {
26
+ "epoch": 0.2870126763932074,
27
+ "grad_norm": 33.96213912963867,
28
+ "learning_rate": 0.00035,
29
+ "loss": 63.1931,
30
+ "step": 150
31
+ },
32
+ {
33
+ "epoch": 0.3826835685242765,
34
+ "grad_norm": 11.814678192138672,
35
+ "learning_rate": 0.00046666666666666666,
36
+ "loss": 61.9029,
37
+ "step": 200
38
+ },
39
+ {
40
+ "epoch": 0.4783544606553456,
41
+ "grad_norm": 12.915092468261719,
42
+ "learning_rate": 0.0005833333333333334,
43
+ "loss": 64.1859,
44
+ "step": 250
45
+ },
46
+ {
47
+ "epoch": 0.5740253527864148,
48
+ "grad_norm": 9.05835247039795,
49
+ "learning_rate": 0.0007,
50
+ "loss": 67.9854,
51
+ "step": 300
52
+ },
53
+ {
54
+ "epoch": 0.6696962449174838,
55
+ "grad_norm": 6.513669967651367,
56
+ "learning_rate": 0.000699821634561209,
57
+ "loss": 72.0232,
58
+ "step": 350
59
+ },
60
+ {
61
+ "epoch": 0.765367137048553,
62
+ "grad_norm": 7.418867588043213,
63
+ "learning_rate": 0.0006992867200404345,
64
+ "loss": 75.0032,
65
+ "step": 400
66
+ },
67
+ {
68
+ "epoch": 0.8610380291796221,
69
+ "grad_norm": 6.645226955413818,
70
+ "learning_rate": 0.0006983958016391807,
71
+ "loss": 77.4417,
72
+ "step": 450
73
+ },
74
+ {
75
+ "epoch": 0.9567089213106912,
76
+ "grad_norm": 7.699553489685059,
77
+ "learning_rate": 0.0006971497874091708,
78
+ "loss": 81.1586,
79
+ "step": 500
80
+ },
81
+ {
82
+ "epoch": 0.9988041138483617,
83
+ "eval_accuracy": 0.0,
84
+ "eval_loss": 5.180789947509766,
85
+ "eval_normalizer": 685885.0,
86
+ "eval_runtime": 107.5006,
87
+ "eval_samples_per_second": 497.272,
88
+ "eval_steps_per_second": 0.977,
89
+ "step": 522
90
+ },
91
+ {
92
+ "epoch": 1.0535756995933987,
93
+ "grad_norm": 69.71871185302734,
94
+ "learning_rate": 0.0006955499473268326,
95
+ "loss": 64.5289,
96
+ "step": 550
97
+ },
98
+ {
99
+ "epoch": 1.1492465917244679,
100
+ "grad_norm": 64.23685455322266,
101
+ "learning_rate": 0.0006935979119988993,
102
+ "loss": 48.1824,
103
+ "step": 600
104
+ },
105
+ {
106
+ "epoch": 1.244917483855537,
107
+ "grad_norm": 14.777277946472168,
108
+ "learning_rate": 0.0006912956710004438,
109
+ "loss": 47.3129,
110
+ "step": 650
111
+ },
112
+ {
113
+ "epoch": 1.3405883759866062,
114
+ "grad_norm": 14.003345489501953,
115
+ "learning_rate": 0.0006886455708470427,
116
+ "loss": 49.7599,
117
+ "step": 700
118
+ },
119
+ {
120
+ "epoch": 1.4362592681176751,
121
+ "grad_norm": 7.022165775299072,
122
+ "learning_rate": 0.0006856503126031346,
123
+ "loss": 53.7128,
124
+ "step": 750
125
+ },
126
+ {
127
+ "epoch": 1.5319301602487443,
128
+ "grad_norm": 9.524980545043945,
129
+ "learning_rate": 0.0006823129491290102,
130
+ "loss": 57.7798,
131
+ "step": 800
132
+ },
133
+ {
134
+ "epoch": 1.6276010523798135,
135
+ "grad_norm": 5.217013835906982,
136
+ "learning_rate": 0.0006786368819692442,
137
+ "loss": 62.1501,
138
+ "step": 850
139
+ },
140
+ {
141
+ "epoch": 1.7232719445108826,
142
+ "grad_norm": 4.990296363830566,
143
+ "learning_rate": 0.0006746258578857331,
144
+ "loss": 66.0142,
145
+ "step": 900
146
+ },
147
+ {
148
+ "epoch": 1.8189428366419516,
149
+ "grad_norm": 4.7475666999816895,
150
+ "learning_rate": 0.000670283965038881,
151
+ "loss": 69.0751,
152
+ "step": 950
153
+ },
154
+ {
155
+ "epoch": 1.9146137287730207,
156
+ "grad_norm": 4.947103023529053,
157
+ "learning_rate": 0.0006656156288208179,
158
+ "loss": 72.45,
159
+ "step": 1000
160
+ },
161
+ {
162
+ "epoch": 1.9988041138483617,
163
+ "eval_accuracy": 0.0,
164
+ "eval_loss": 4.9698710441589355,
165
+ "eval_normalizer": 685885.0,
166
+ "eval_runtime": 106.2071,
167
+ "eval_samples_per_second": 503.328,
168
+ "eval_steps_per_second": 0.989,
169
+ "step": 1044
170
+ },
171
+ {
172
+ "epoch": 2.0114805070557282,
173
+ "grad_norm": 0.008735532872378826,
174
+ "learning_rate": 0.000660625607344904,
175
+ "loss": 69.3964,
176
+ "step": 1050
177
+ },
178
+ {
179
+ "epoch": 2.1071513991867974,
180
+ "grad_norm": 131.39796447753906,
181
+ "learning_rate": 0.0006553189865961112,
182
+ "loss": 30.6883,
183
+ "step": 1100
184
+ },
185
+ {
186
+ "epoch": 2.2028222913178666,
187
+ "grad_norm": 24.75687026977539,
188
+ "learning_rate": 0.0006497011752472301,
189
+ "loss": 41.9627,
190
+ "step": 1150
191
+ },
192
+ {
193
+ "epoch": 2.2984931834489357,
194
+ "grad_norm": 12.107126235961914,
195
+ "learning_rate": 0.0006437778991461825,
196
+ "loss": 44.173,
197
+ "step": 1200
198
+ },
199
+ {
200
+ "epoch": 2.394164075580005,
201
+ "grad_norm": 11.241535186767578,
202
+ "learning_rate": 0.0006375551954800587,
203
+ "loss": 48.4226,
204
+ "step": 1250
205
+ },
206
+ {
207
+ "epoch": 2.489834967711074,
208
+ "grad_norm": 7.552574157714844,
209
+ "learning_rate": 0.0006310394066218296,
210
+ "loss": 52.2424,
211
+ "step": 1300
212
+ },
213
+ {
214
+ "epoch": 2.5855058598421428,
215
+ "grad_norm": 5.529368877410889,
216
+ "learning_rate": 0.0006242371736660025,
217
+ "loss": 56.5036,
218
+ "step": 1350
219
+ },
220
+ {
221
+ "epoch": 2.6811767519732124,
222
+ "grad_norm": 5.114018440246582,
223
+ "learning_rate": 0.000617155429659811,
224
+ "loss": 60.4936,
225
+ "step": 1400
226
+ },
227
+ {
228
+ "epoch": 2.776847644104281,
229
+ "grad_norm": 4.804865837097168,
230
+ "learning_rate": 0.0006098013925368385,
231
+ "loss": 63.7044,
232
+ "step": 1450
233
+ },
234
+ {
235
+ "epoch": 2.8725185362353503,
236
+ "grad_norm": 4.746634483337402,
237
+ "learning_rate": 0.0006021825577602754,
238
+ "loss": 66.5259,
239
+ "step": 1500
240
+ },
241
+ {
242
+ "epoch": 2.9681894283664194,
243
+ "grad_norm": 4.658045768737793,
244
+ "learning_rate": 0.0005943066906833104,
245
+ "loss": 70.2771,
246
+ "step": 1550
247
+ },
248
+ {
249
+ "epoch": 2.9988041138483617,
250
+ "eval_accuracy": 0.0,
251
+ "eval_loss": 4.887843132019043,
252
+ "eval_normalizer": 685885.0,
253
+ "eval_runtime": 107.3098,
254
+ "eval_samples_per_second": 498.156,
255
+ "eval_steps_per_second": 0.978,
256
+ "step": 1566
257
+ },
258
+ {
259
+ "epoch": 3.065056206649127,
260
+ "grad_norm": 320.45758056640625,
261
+ "learning_rate": 0.0005861818186344407,
262
+ "loss": 45.3344,
263
+ "step": 1600
264
+ },
265
+ {
266
+ "epoch": 3.160727098780196,
267
+ "grad_norm": 29.107763290405273,
268
+ "learning_rate": 0.00057781622273577,
269
+ "loss": 35.1219,
270
+ "step": 1650
271
+ },
272
+ {
273
+ "epoch": 3.2563979909112652,
274
+ "grad_norm": 14.5596342086792,
275
+ "learning_rate": 0.0005692184294626307,
276
+ "loss": 40.1945,
277
+ "step": 1700
278
+ },
279
+ {
280
+ "epoch": 3.3520688830423344,
281
+ "grad_norm": 19.97965431213379,
282
+ "learning_rate": 0.0005603972019531362,
283
+ "loss": 44.1522,
284
+ "step": 1750
285
+ },
286
+ {
287
+ "epoch": 3.4477397751734036,
288
+ "grad_norm": 9.178486824035645,
289
+ "learning_rate": 0.0005513615310765172,
290
+ "loss": 48.1065,
291
+ "step": 1800
292
+ },
293
+ {
294
+ "epoch": 3.5434106673044727,
295
+ "grad_norm": 5.76595401763916,
296
+ "learning_rate": 0.0005421206262693491,
297
+ "loss": 51.5121,
298
+ "step": 1850
299
+ },
300
+ {
301
+ "epoch": 3.639081559435542,
302
+ "grad_norm": 6.189362525939941,
303
+ "learning_rate": 0.0005326839061490078,
304
+ "loss": 55.4692,
305
+ "step": 1900
306
+ },
307
+ {
308
+ "epoch": 3.734752451566611,
309
+ "grad_norm": 5.159908771514893,
310
+ "learning_rate": 0.0005230609889139216,
311
+ "loss": 58.6133,
312
+ "step": 1950
313
+ },
314
+ {
315
+ "epoch": 3.83042334369768,
316
+ "grad_norm": 5.086872577667236,
317
+ "learning_rate": 0.0005132616825404055,
318
+ "loss": 61.4673,
319
+ "step": 2000
320
+ },
321
+ {
322
+ "epoch": 3.926094235828749,
323
+ "grad_norm": 5.382955074310303,
324
+ "learning_rate": 0.0005032959747860662,
325
+ "loss": 64.5968,
326
+ "step": 2050
327
+ },
328
+ {
329
+ "epoch": 3.9988041138483617,
330
+ "eval_accuracy": 0.0,
331
+ "eval_loss": 4.891599655151367,
332
+ "eval_normalizer": 685885.0,
333
+ "eval_runtime": 107.2829,
334
+ "eval_samples_per_second": 498.281,
335
+ "eval_steps_per_second": 0.979,
336
+ "step": 2088
337
+ },
338
+ {
339
+ "epoch": 4.0229610141114565,
340
+ "grad_norm": 0.001459582126699388,
341
+ "learning_rate": 0.000493174023009969,
342
+ "loss": 53.9448,
343
+ "step": 2100
344
+ },
345
+ {
346
+ "epoch": 4.118631906242526,
347
+ "grad_norm": 55.27322769165039,
348
+ "learning_rate": 0.00048290614381994235,
349
+ "loss": 28.4441,
350
+ "step": 2150
351
+ },
352
+ {
353
+ "epoch": 4.214302798373595,
354
+ "grad_norm": 20.95206642150879,
355
+ "learning_rate": 0.00047250280255757023,
356
+ "loss": 37.6623,
357
+ "step": 2200
358
+ },
359
+ {
360
+ "epoch": 4.309973690504664,
361
+ "grad_norm": 11.7781982421875,
362
+ "learning_rate": 0.0004619746026315906,
363
+ "loss": 40.4448,
364
+ "step": 2250
365
+ },
366
+ {
367
+ "epoch": 4.405644582635733,
368
+ "grad_norm": 10.008404731750488,
369
+ "learning_rate": 0.00045133227471057203,
370
+ "loss": 43.9661,
371
+ "step": 2300
372
+ },
373
+ {
374
+ "epoch": 4.501315474766802,
375
+ "grad_norm": 7.9073076248168945,
376
+ "learning_rate": 0.00044058666578588224,
377
+ "loss": 47.1353,
378
+ "step": 2350
379
+ },
380
+ {
381
+ "epoch": 4.596986366897871,
382
+ "grad_norm": 6.407111644744873,
383
+ "learning_rate": 0.0004297487281160982,
384
+ "loss": 50.1302,
385
+ "step": 2400
386
+ },
387
+ {
388
+ "epoch": 4.69265725902894,
389
+ "grad_norm": 8.047704696655273,
390
+ "learning_rate": 0.00041882950806412285,
391
+ "loss": 53.3268,
392
+ "step": 2450
393
+ },
394
+ {
395
+ "epoch": 4.78832815116001,
396
+ "grad_norm": 7.598127841949463,
397
+ "learning_rate": 0.0004078401348383897,
398
+ "loss": 56.3055,
399
+ "step": 2500
400
+ },
401
+ {
402
+ "epoch": 4.8839990432910785,
403
+ "grad_norm": 6.841054439544678,
404
+ "learning_rate": 0.00039679180914962693,
405
+ "loss": 59.1088,
406
+ "step": 2550
407
+ },
408
+ {
409
+ "epoch": 4.979669935422148,
410
+ "grad_norm": 5.6036577224731445,
411
+ "learning_rate": 0.00038569579179474536,
412
+ "loss": 63.7624,
413
+ "step": 2600
414
+ },
415
+ {
416
+ "epoch": 4.998804113848362,
417
+ "eval_accuracy": 0.0,
418
+ "eval_loss": 4.91996955871582,
419
+ "eval_normalizer": 685885.0,
420
+ "eval_runtime": 106.843,
421
+ "eval_samples_per_second": 500.332,
422
+ "eval_steps_per_second": 0.983,
423
+ "step": 2610
424
+ },
425
+ {
426
+ "epoch": 5.076536713704855,
427
+ "grad_norm": 56.89580154418945,
428
+ "learning_rate": 0.00037456339217948394,
429
+ "loss": 32.6245,
430
+ "step": 2650
431
+ },
432
+ {
433
+ "epoch": 5.172207605835925,
434
+ "grad_norm": 37.92051315307617,
435
+ "learning_rate": 0.0003634059567915124,
436
+ "loss": 32.9362,
437
+ "step": 2700
438
+ },
439
+ {
440
+ "epoch": 5.2678784979669935,
441
+ "grad_norm": 12.96127700805664,
442
+ "learning_rate": 0.00035223485763573775,
443
+ "loss": 37.0047,
444
+ "step": 2750
445
+ },
446
+ {
447
+ "epoch": 5.363549390098063,
448
+ "grad_norm": 10.119488716125488,
449
+ "learning_rate": 0.00034106148064360405,
450
+ "loss": 40.3371,
451
+ "step": 2800
452
+ },
453
+ {
454
+ "epoch": 5.459220282229132,
455
+ "grad_norm": 13.921025276184082,
456
+ "learning_rate": 0.0003298972140681969,
457
+ "loss": 42.8515,
458
+ "step": 2850
459
+ },
460
+ {
461
+ "epoch": 5.5548911743602005,
462
+ "grad_norm": 7.783252716064453,
463
+ "learning_rate": 0.00031875343687698203,
464
+ "loss": 44.7372,
465
+ "step": 2900
466
+ },
467
+ {
468
+ "epoch": 5.65056206649127,
469
+ "grad_norm": 7.753182411193848,
470
+ "learning_rate": 0.000307641507154008,
471
+ "loss": 47.9169,
472
+ "step": 2950
473
+ },
474
+ {
475
+ "epoch": 5.746232958622339,
476
+ "grad_norm": 7.485280513763428,
477
+ "learning_rate": 0.0002965727505233939,
478
+ "loss": 50.5956,
479
+ "step": 3000
480
+ },
481
+ {
482
+ "epoch": 5.8419038507534085,
483
+ "grad_norm": 7.22899866104126,
484
+ "learning_rate": 0.0002855584486059016,
485
+ "loss": 54.094,
486
+ "step": 3050
487
+ },
488
+ {
489
+ "epoch": 5.937574742884477,
490
+ "grad_norm": 7.100270748138428,
491
+ "learning_rate": 0.00027460982752035653,
492
+ "loss": 57.2679,
493
+ "step": 3100
494
+ },
495
+ {
496
+ "epoch": 5.998804113848362,
497
+ "eval_accuracy": 0.0,
498
+ "eval_loss": 4.958333492279053,
499
+ "eval_normalizer": 685885.0,
500
+ "eval_runtime": 108.5285,
501
+ "eval_samples_per_second": 492.562,
502
+ "eval_steps_per_second": 0.967,
503
+ "step": 3132
504
+ }
505
+ ],
506
+ "logging_steps": 50,
507
+ "max_steps": 5220,
508
+ "num_input_tokens_seen": 0,
509
+ "num_train_epochs": 10,
510
+ "save_steps": 500,
511
+ "stateful_callbacks": {
512
+ "TrainerControl": {
513
+ "args": {
514
+ "should_epoch_stop": false,
515
+ "should_evaluate": false,
516
+ "should_log": false,
517
+ "should_save": true,
518
+ "should_training_stop": false
519
+ },
520
+ "attributes": {}
521
+ }
522
+ },
523
+ "total_flos": 2.587043421778084e+17,
524
+ "train_batch_size": 128,
525
+ "trial_name": null,
526
+ "trial_params": null
527
+ }
checkpoints/checkpoint-3132/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:2ce7d2a89af981101c97b9dcfbbb18ef02080865f73f1485630adddf7702facd
3
+ size 5432