kevinoli commited on
Commit
72d0151
1 Parent(s): 18ab3de

Training in progress, step 24500, checkpoint

Browse files
checkpoint-24500/config.json ADDED
@@ -0,0 +1,31 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "openai/clip-vit-large-patch14-336",
3
+ "architectures": [
4
+ "CLIPModel"
5
+ ],
6
+ "initializer_factor": 1.0,
7
+ "logit_scale_init_value": 2.6592,
8
+ "model_type": "clip",
9
+ "projection_dim": 768,
10
+ "text_config": {
11
+ "dropout": 0.0,
12
+ "hidden_size": 768,
13
+ "intermediate_size": 3072,
14
+ "model_type": "clip_text_model",
15
+ "num_attention_heads": 12,
16
+ "projection_dim": 768
17
+ },
18
+ "torch_dtype": "float32",
19
+ "transformers_version": "4.45.0.dev0",
20
+ "vision_config": {
21
+ "dropout": 0.0,
22
+ "hidden_size": 1024,
23
+ "image_size": 336,
24
+ "intermediate_size": 4096,
25
+ "model_type": "clip_vision_model",
26
+ "num_attention_heads": 16,
27
+ "num_hidden_layers": 24,
28
+ "patch_size": 14,
29
+ "projection_dim": 768
30
+ }
31
+ }
checkpoint-24500/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5331e390bb3f9c63c4f99eaf65256a6ac1ef9165bbdcf2837f56919c883e0e55
3
+ size 1711848436
checkpoint-24500/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:222d2f8f83c17d3a5971209f01861782e77daec39c77d979973be3278c8228fa
3
+ size 3424043887
checkpoint-24500/rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c5d4e1dd7f4c7bdfb91258c79712f110ad63d6950c728f09872c7fc2ae33ecaa
3
+ size 14503
checkpoint-24500/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:16d9550e293e0be02567991c406370e1618400811d6462ea8d8c126d3b1f08c3
3
+ size 623
checkpoint-24500/trainer_state.json ADDED
@@ -0,0 +1,768 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": 0.47645464539527893,
3
+ "best_model_checkpoint": "./output/clip-finetuned-csu-p14-336-e4l57-l/checkpoint-24500",
4
+ "epoch": 2.6130546075085324,
5
+ "eval_steps": 500,
6
+ "global_step": 24500,
7
+ "is_hyper_param_search": false,
8
+ "is_local_process_zero": true,
9
+ "is_world_process_zero": true,
10
+ "log_history": [
11
+ {
12
+ "epoch": 0.05332764505119454,
13
+ "grad_norm": 414.2552795410156,
14
+ "learning_rate": 4.911120591581342e-07,
15
+ "loss": 0.3809,
16
+ "step": 500
17
+ },
18
+ {
19
+ "epoch": 0.05332764505119454,
20
+ "eval_loss": 1.1109352111816406,
21
+ "eval_runtime": 63.6734,
22
+ "eval_samples_per_second": 15.501,
23
+ "eval_steps_per_second": 1.947,
24
+ "step": 500
25
+ },
26
+ {
27
+ "epoch": 0.10665529010238908,
28
+ "grad_norm": 45.880027770996094,
29
+ "learning_rate": 4.822241183162685e-07,
30
+ "loss": 0.2649,
31
+ "step": 1000
32
+ },
33
+ {
34
+ "epoch": 0.10665529010238908,
35
+ "eval_loss": 0.9588962197303772,
36
+ "eval_runtime": 62.6825,
37
+ "eval_samples_per_second": 15.746,
38
+ "eval_steps_per_second": 1.978,
39
+ "step": 1000
40
+ },
41
+ {
42
+ "epoch": 0.1599829351535836,
43
+ "grad_norm": 432.9915466308594,
44
+ "learning_rate": 4.733361774744027e-07,
45
+ "loss": 0.2104,
46
+ "step": 1500
47
+ },
48
+ {
49
+ "epoch": 0.1599829351535836,
50
+ "eval_loss": 0.90110844373703,
51
+ "eval_runtime": 62.6757,
52
+ "eval_samples_per_second": 15.748,
53
+ "eval_steps_per_second": 1.978,
54
+ "step": 1500
55
+ },
56
+ {
57
+ "epoch": 0.21331058020477817,
58
+ "grad_norm": 0.014208819717168808,
59
+ "learning_rate": 4.6444823663253695e-07,
60
+ "loss": 0.1849,
61
+ "step": 2000
62
+ },
63
+ {
64
+ "epoch": 0.21331058020477817,
65
+ "eval_loss": 0.8570474982261658,
66
+ "eval_runtime": 63.7296,
67
+ "eval_samples_per_second": 15.487,
68
+ "eval_steps_per_second": 1.946,
69
+ "step": 2000
70
+ },
71
+ {
72
+ "epoch": 0.2666382252559727,
73
+ "grad_norm": 174.73324584960938,
74
+ "learning_rate": 4.5556029579067116e-07,
75
+ "loss": 0.2056,
76
+ "step": 2500
77
+ },
78
+ {
79
+ "epoch": 0.2666382252559727,
80
+ "eval_loss": 0.7895939350128174,
81
+ "eval_runtime": 63.8387,
82
+ "eval_samples_per_second": 15.461,
83
+ "eval_steps_per_second": 1.942,
84
+ "step": 2500
85
+ },
86
+ {
87
+ "epoch": 0.3199658703071672,
88
+ "grad_norm": 9.852073823424234e-09,
89
+ "learning_rate": 4.4667235494880547e-07,
90
+ "loss": 0.159,
91
+ "step": 3000
92
+ },
93
+ {
94
+ "epoch": 0.3199658703071672,
95
+ "eval_loss": 0.7647623419761658,
96
+ "eval_runtime": 63.9155,
97
+ "eval_samples_per_second": 15.442,
98
+ "eval_steps_per_second": 1.94,
99
+ "step": 3000
100
+ },
101
+ {
102
+ "epoch": 0.37329351535836175,
103
+ "grad_norm": 0.0025215104687958956,
104
+ "learning_rate": 4.377844141069397e-07,
105
+ "loss": 0.1696,
106
+ "step": 3500
107
+ },
108
+ {
109
+ "epoch": 0.37329351535836175,
110
+ "eval_loss": 0.7638036012649536,
111
+ "eval_runtime": 63.8003,
112
+ "eval_samples_per_second": 15.47,
113
+ "eval_steps_per_second": 1.944,
114
+ "step": 3500
115
+ },
116
+ {
117
+ "epoch": 0.42662116040955633,
118
+ "grad_norm": 2.4611830711364746,
119
+ "learning_rate": 4.2889647326507393e-07,
120
+ "loss": 0.1125,
121
+ "step": 4000
122
+ },
123
+ {
124
+ "epoch": 0.42662116040955633,
125
+ "eval_loss": 0.7486374378204346,
126
+ "eval_runtime": 63.9847,
127
+ "eval_samples_per_second": 15.426,
128
+ "eval_steps_per_second": 1.938,
129
+ "step": 4000
130
+ },
131
+ {
132
+ "epoch": 0.47994880546075086,
133
+ "grad_norm": 5.230295658111572,
134
+ "learning_rate": 4.2000853242320814e-07,
135
+ "loss": 0.1322,
136
+ "step": 4500
137
+ },
138
+ {
139
+ "epoch": 0.47994880546075086,
140
+ "eval_loss": 0.7251659035682678,
141
+ "eval_runtime": 64.4782,
142
+ "eval_samples_per_second": 15.308,
143
+ "eval_steps_per_second": 1.923,
144
+ "step": 4500
145
+ },
146
+ {
147
+ "epoch": 0.5332764505119454,
148
+ "grad_norm": 3.821002974291332e-05,
149
+ "learning_rate": 4.1112059158134245e-07,
150
+ "loss": 0.1633,
151
+ "step": 5000
152
+ },
153
+ {
154
+ "epoch": 0.5332764505119454,
155
+ "eval_loss": 0.7035414576530457,
156
+ "eval_runtime": 62.6186,
157
+ "eval_samples_per_second": 15.762,
158
+ "eval_steps_per_second": 1.98,
159
+ "step": 5000
160
+ },
161
+ {
162
+ "epoch": 0.58660409556314,
163
+ "grad_norm": 5.916939244343666e-06,
164
+ "learning_rate": 4.0223265073947665e-07,
165
+ "loss": 0.125,
166
+ "step": 5500
167
+ },
168
+ {
169
+ "epoch": 0.58660409556314,
170
+ "eval_loss": 0.6792380809783936,
171
+ "eval_runtime": 65.0551,
172
+ "eval_samples_per_second": 15.172,
173
+ "eval_steps_per_second": 1.906,
174
+ "step": 5500
175
+ },
176
+ {
177
+ "epoch": 0.6399317406143344,
178
+ "grad_norm": 92.8386459350586,
179
+ "learning_rate": 3.933447098976109e-07,
180
+ "loss": 0.1297,
181
+ "step": 6000
182
+ },
183
+ {
184
+ "epoch": 0.6399317406143344,
185
+ "eval_loss": 0.6549482345581055,
186
+ "eval_runtime": 64.0827,
187
+ "eval_samples_per_second": 15.402,
188
+ "eval_steps_per_second": 1.935,
189
+ "step": 6000
190
+ },
191
+ {
192
+ "epoch": 0.693259385665529,
193
+ "grad_norm": 0.00014717792510055006,
194
+ "learning_rate": 3.8445676905574517e-07,
195
+ "loss": 0.1114,
196
+ "step": 6500
197
+ },
198
+ {
199
+ "epoch": 0.693259385665529,
200
+ "eval_loss": 0.6223254799842834,
201
+ "eval_runtime": 64.89,
202
+ "eval_samples_per_second": 15.21,
203
+ "eval_steps_per_second": 1.911,
204
+ "step": 6500
205
+ },
206
+ {
207
+ "epoch": 0.7465870307167235,
208
+ "grad_norm": 4.789559397977428e-07,
209
+ "learning_rate": 3.755688282138794e-07,
210
+ "loss": 0.144,
211
+ "step": 7000
212
+ },
213
+ {
214
+ "epoch": 0.7465870307167235,
215
+ "eval_loss": 0.6204637289047241,
216
+ "eval_runtime": 62.7674,
217
+ "eval_samples_per_second": 15.725,
218
+ "eval_steps_per_second": 1.976,
219
+ "step": 7000
220
+ },
221
+ {
222
+ "epoch": 0.7999146757679181,
223
+ "grad_norm": 1.7078508138656616,
224
+ "learning_rate": 3.6668088737201363e-07,
225
+ "loss": 0.1399,
226
+ "step": 7500
227
+ },
228
+ {
229
+ "epoch": 0.7999146757679181,
230
+ "eval_loss": 0.6074371933937073,
231
+ "eval_runtime": 63.4694,
232
+ "eval_samples_per_second": 15.551,
233
+ "eval_steps_per_second": 1.954,
234
+ "step": 7500
235
+ },
236
+ {
237
+ "epoch": 0.8532423208191127,
238
+ "grad_norm": 870.1183471679688,
239
+ "learning_rate": 3.5779294653014783e-07,
240
+ "loss": 0.141,
241
+ "step": 8000
242
+ },
243
+ {
244
+ "epoch": 0.8532423208191127,
245
+ "eval_loss": 0.5950366854667664,
246
+ "eval_runtime": 62.915,
247
+ "eval_samples_per_second": 15.688,
248
+ "eval_steps_per_second": 1.971,
249
+ "step": 8000
250
+ },
251
+ {
252
+ "epoch": 0.9065699658703071,
253
+ "grad_norm": 5.781071013188921e-05,
254
+ "learning_rate": 3.4890500568828214e-07,
255
+ "loss": 0.1291,
256
+ "step": 8500
257
+ },
258
+ {
259
+ "epoch": 0.9065699658703071,
260
+ "eval_loss": 0.5800224542617798,
261
+ "eval_runtime": 62.3336,
262
+ "eval_samples_per_second": 15.834,
263
+ "eval_steps_per_second": 1.989,
264
+ "step": 8500
265
+ },
266
+ {
267
+ "epoch": 0.9598976109215017,
268
+ "grad_norm": 1.3669992685317993,
269
+ "learning_rate": 3.4001706484641635e-07,
270
+ "loss": 0.1246,
271
+ "step": 9000
272
+ },
273
+ {
274
+ "epoch": 0.9598976109215017,
275
+ "eval_loss": 0.5702486038208008,
276
+ "eval_runtime": 64.055,
277
+ "eval_samples_per_second": 15.409,
278
+ "eval_steps_per_second": 1.936,
279
+ "step": 9000
280
+ },
281
+ {
282
+ "epoch": 1.0132252559726962,
283
+ "grad_norm": 644.68115234375,
284
+ "learning_rate": 3.311291240045506e-07,
285
+ "loss": 0.1257,
286
+ "step": 9500
287
+ },
288
+ {
289
+ "epoch": 1.0132252559726962,
290
+ "eval_loss": 0.5807780623435974,
291
+ "eval_runtime": 62.5312,
292
+ "eval_samples_per_second": 15.784,
293
+ "eval_steps_per_second": 1.983,
294
+ "step": 9500
295
+ },
296
+ {
297
+ "epoch": 1.0665529010238908,
298
+ "grad_norm": 27.854568481445312,
299
+ "learning_rate": 3.2224118316268486e-07,
300
+ "loss": 0.0336,
301
+ "step": 10000
302
+ },
303
+ {
304
+ "epoch": 1.0665529010238908,
305
+ "eval_loss": 0.5713164210319519,
306
+ "eval_runtime": 62.5622,
307
+ "eval_samples_per_second": 15.776,
308
+ "eval_steps_per_second": 1.982,
309
+ "step": 10000
310
+ },
311
+ {
312
+ "epoch": 1.1198805460750854,
313
+ "grad_norm": 0.00045933027286082506,
314
+ "learning_rate": 3.133532423208191e-07,
315
+ "loss": 0.0663,
316
+ "step": 10500
317
+ },
318
+ {
319
+ "epoch": 1.1198805460750854,
320
+ "eval_loss": 0.5601483583450317,
321
+ "eval_runtime": 63.4675,
322
+ "eval_samples_per_second": 15.551,
323
+ "eval_steps_per_second": 1.954,
324
+ "step": 10500
325
+ },
326
+ {
327
+ "epoch": 1.17320819112628,
328
+ "grad_norm": 0.14330387115478516,
329
+ "learning_rate": 3.044653014789533e-07,
330
+ "loss": 0.078,
331
+ "step": 11000
332
+ },
333
+ {
334
+ "epoch": 1.17320819112628,
335
+ "eval_loss": 0.5580261945724487,
336
+ "eval_runtime": 63.3927,
337
+ "eval_samples_per_second": 15.57,
338
+ "eval_steps_per_second": 1.956,
339
+ "step": 11000
340
+ },
341
+ {
342
+ "epoch": 1.2265358361774745,
343
+ "grad_norm": 4.066005518388316e-11,
344
+ "learning_rate": 2.955773606370876e-07,
345
+ "loss": 0.1118,
346
+ "step": 11500
347
+ },
348
+ {
349
+ "epoch": 1.2265358361774745,
350
+ "eval_loss": 0.5491370558738708,
351
+ "eval_runtime": 63.7293,
352
+ "eval_samples_per_second": 15.487,
353
+ "eval_steps_per_second": 1.946,
354
+ "step": 11500
355
+ },
356
+ {
357
+ "epoch": 1.2798634812286689,
358
+ "grad_norm": 7.519358769059181e-05,
359
+ "learning_rate": 2.8668941979522184e-07,
360
+ "loss": 0.0534,
361
+ "step": 12000
362
+ },
363
+ {
364
+ "epoch": 1.2798634812286689,
365
+ "eval_loss": 0.5478764772415161,
366
+ "eval_runtime": 62.7547,
367
+ "eval_samples_per_second": 15.728,
368
+ "eval_steps_per_second": 1.976,
369
+ "step": 12000
370
+ },
371
+ {
372
+ "epoch": 1.3331911262798635,
373
+ "grad_norm": 8.400617446113756e-08,
374
+ "learning_rate": 2.778014789533561e-07,
375
+ "loss": 0.0507,
376
+ "step": 12500
377
+ },
378
+ {
379
+ "epoch": 1.3331911262798635,
380
+ "eval_loss": 0.576600968837738,
381
+ "eval_runtime": 64.301,
382
+ "eval_samples_per_second": 15.35,
383
+ "eval_steps_per_second": 1.928,
384
+ "step": 12500
385
+ },
386
+ {
387
+ "epoch": 1.386518771331058,
388
+ "grad_norm": 2.5385464130084356e-12,
389
+ "learning_rate": 2.689135381114903e-07,
390
+ "loss": 0.0784,
391
+ "step": 13000
392
+ },
393
+ {
394
+ "epoch": 1.386518771331058,
395
+ "eval_loss": 0.5729931592941284,
396
+ "eval_runtime": 63.652,
397
+ "eval_samples_per_second": 15.506,
398
+ "eval_steps_per_second": 1.948,
399
+ "step": 13000
400
+ },
401
+ {
402
+ "epoch": 1.4398464163822526,
403
+ "grad_norm": 0.08409273624420166,
404
+ "learning_rate": 2.600255972696246e-07,
405
+ "loss": 0.079,
406
+ "step": 13500
407
+ },
408
+ {
409
+ "epoch": 1.4398464163822526,
410
+ "eval_loss": 0.5694039463996887,
411
+ "eval_runtime": 62.8247,
412
+ "eval_samples_per_second": 15.71,
413
+ "eval_steps_per_second": 1.974,
414
+ "step": 13500
415
+ },
416
+ {
417
+ "epoch": 1.493174061433447,
418
+ "grad_norm": 3.7334253022436314e-08,
419
+ "learning_rate": 2.511376564277588e-07,
420
+ "loss": 0.0507,
421
+ "step": 14000
422
+ },
423
+ {
424
+ "epoch": 1.493174061433447,
425
+ "eval_loss": 0.5713907480239868,
426
+ "eval_runtime": 62.3577,
427
+ "eval_samples_per_second": 15.828,
428
+ "eval_steps_per_second": 1.989,
429
+ "step": 14000
430
+ },
431
+ {
432
+ "epoch": 1.5465017064846416,
433
+ "grad_norm": 1.5219894647598267,
434
+ "learning_rate": 2.422497155858931e-07,
435
+ "loss": 0.0774,
436
+ "step": 14500
437
+ },
438
+ {
439
+ "epoch": 1.5465017064846416,
440
+ "eval_loss": 0.5674533843994141,
441
+ "eval_runtime": 62.3753,
442
+ "eval_samples_per_second": 15.824,
443
+ "eval_steps_per_second": 1.988,
444
+ "step": 14500
445
+ },
446
+ {
447
+ "epoch": 1.5998293515358362,
448
+ "grad_norm": 879.1193237304688,
449
+ "learning_rate": 2.333617747440273e-07,
450
+ "loss": 0.0376,
451
+ "step": 15000
452
+ },
453
+ {
454
+ "epoch": 1.5998293515358362,
455
+ "eval_loss": 0.561253547668457,
456
+ "eval_runtime": 62.4788,
457
+ "eval_samples_per_second": 15.797,
458
+ "eval_steps_per_second": 1.985,
459
+ "step": 15000
460
+ },
461
+ {
462
+ "epoch": 1.6531569965870307,
463
+ "grad_norm": 0.0006221308722160757,
464
+ "learning_rate": 2.2447383390216154e-07,
465
+ "loss": 0.0929,
466
+ "step": 15500
467
+ },
468
+ {
469
+ "epoch": 1.6531569965870307,
470
+ "eval_loss": 0.5395999550819397,
471
+ "eval_runtime": 62.5651,
472
+ "eval_samples_per_second": 15.776,
473
+ "eval_steps_per_second": 1.982,
474
+ "step": 15500
475
+ },
476
+ {
477
+ "epoch": 1.7064846416382253,
478
+ "grad_norm": 37.85109329223633,
479
+ "learning_rate": 2.155858930602958e-07,
480
+ "loss": 0.0442,
481
+ "step": 16000
482
+ },
483
+ {
484
+ "epoch": 1.7064846416382253,
485
+ "eval_loss": 0.5307884216308594,
486
+ "eval_runtime": 62.4681,
487
+ "eval_samples_per_second": 15.8,
488
+ "eval_steps_per_second": 1.985,
489
+ "step": 16000
490
+ },
491
+ {
492
+ "epoch": 1.75981228668942,
493
+ "grad_norm": 9.600337885773413e-11,
494
+ "learning_rate": 2.0669795221843002e-07,
495
+ "loss": 0.0593,
496
+ "step": 16500
497
+ },
498
+ {
499
+ "epoch": 1.75981228668942,
500
+ "eval_loss": 0.5230685472488403,
501
+ "eval_runtime": 63.8124,
502
+ "eval_samples_per_second": 15.467,
503
+ "eval_steps_per_second": 1.943,
504
+ "step": 16500
505
+ },
506
+ {
507
+ "epoch": 1.8131399317406145,
508
+ "grad_norm": 9.067448940129619e-13,
509
+ "learning_rate": 1.9781001137656428e-07,
510
+ "loss": 0.0578,
511
+ "step": 17000
512
+ },
513
+ {
514
+ "epoch": 1.8131399317406145,
515
+ "eval_loss": 0.5200281739234924,
516
+ "eval_runtime": 62.3767,
517
+ "eval_samples_per_second": 15.823,
518
+ "eval_steps_per_second": 1.988,
519
+ "step": 17000
520
+ },
521
+ {
522
+ "epoch": 1.8664675767918089,
523
+ "grad_norm": 0.00843018852174282,
524
+ "learning_rate": 1.889220705346985e-07,
525
+ "loss": 0.0928,
526
+ "step": 17500
527
+ },
528
+ {
529
+ "epoch": 1.8664675767918089,
530
+ "eval_loss": 0.509168803691864,
531
+ "eval_runtime": 62.4108,
532
+ "eval_samples_per_second": 15.815,
533
+ "eval_steps_per_second": 1.987,
534
+ "step": 17500
535
+ },
536
+ {
537
+ "epoch": 1.9197952218430034,
538
+ "grad_norm": 4.264631314754297e-09,
539
+ "learning_rate": 1.8003412969283277e-07,
540
+ "loss": 0.0633,
541
+ "step": 18000
542
+ },
543
+ {
544
+ "epoch": 1.9197952218430034,
545
+ "eval_loss": 0.4954444468021393,
546
+ "eval_runtime": 63.3639,
547
+ "eval_samples_per_second": 15.577,
548
+ "eval_steps_per_second": 1.957,
549
+ "step": 18000
550
+ },
551
+ {
552
+ "epoch": 1.9731228668941978,
553
+ "grad_norm": 459.5137939453125,
554
+ "learning_rate": 1.7114618885096697e-07,
555
+ "loss": 0.067,
556
+ "step": 18500
557
+ },
558
+ {
559
+ "epoch": 1.9731228668941978,
560
+ "eval_loss": 0.49844062328338623,
561
+ "eval_runtime": 63.6573,
562
+ "eval_samples_per_second": 15.505,
563
+ "eval_steps_per_second": 1.948,
564
+ "step": 18500
565
+ },
566
+ {
567
+ "epoch": 2.0264505119453924,
568
+ "grad_norm": 0.27590689063072205,
569
+ "learning_rate": 1.6225824800910123e-07,
570
+ "loss": 0.0525,
571
+ "step": 19000
572
+ },
573
+ {
574
+ "epoch": 2.0264505119453924,
575
+ "eval_loss": 0.4999145269393921,
576
+ "eval_runtime": 63.1727,
577
+ "eval_samples_per_second": 15.624,
578
+ "eval_steps_per_second": 1.963,
579
+ "step": 19000
580
+ },
581
+ {
582
+ "epoch": 2.079778156996587,
583
+ "grad_norm": 1.349623744317796e-06,
584
+ "learning_rate": 1.533703071672355e-07,
585
+ "loss": 0.0301,
586
+ "step": 19500
587
+ },
588
+ {
589
+ "epoch": 2.079778156996587,
590
+ "eval_loss": 0.4926412105560303,
591
+ "eval_runtime": 63.4421,
592
+ "eval_samples_per_second": 15.557,
593
+ "eval_steps_per_second": 1.955,
594
+ "step": 19500
595
+ },
596
+ {
597
+ "epoch": 2.1331058020477816,
598
+ "grad_norm": 4.928138878312893e-06,
599
+ "learning_rate": 1.4448236632536972e-07,
600
+ "loss": 0.0408,
601
+ "step": 20000
602
+ },
603
+ {
604
+ "epoch": 2.1331058020477816,
605
+ "eval_loss": 0.4872625172138214,
606
+ "eval_runtime": 63.2047,
607
+ "eval_samples_per_second": 15.616,
608
+ "eval_steps_per_second": 1.962,
609
+ "step": 20000
610
+ },
611
+ {
612
+ "epoch": 2.186433447098976,
613
+ "grad_norm": 0.003605826525017619,
614
+ "learning_rate": 1.3559442548350398e-07,
615
+ "loss": 0.0489,
616
+ "step": 20500
617
+ },
618
+ {
619
+ "epoch": 2.186433447098976,
620
+ "eval_loss": 0.48834049701690674,
621
+ "eval_runtime": 63.5831,
622
+ "eval_samples_per_second": 15.523,
623
+ "eval_steps_per_second": 1.95,
624
+ "step": 20500
625
+ },
626
+ {
627
+ "epoch": 2.2397610921501707,
628
+ "grad_norm": 0.0027391298208385706,
629
+ "learning_rate": 1.267064846416382e-07,
630
+ "loss": 0.0354,
631
+ "step": 21000
632
+ },
633
+ {
634
+ "epoch": 2.2397610921501707,
635
+ "eval_loss": 0.4887964129447937,
636
+ "eval_runtime": 63.0462,
637
+ "eval_samples_per_second": 15.655,
638
+ "eval_steps_per_second": 1.967,
639
+ "step": 21000
640
+ },
641
+ {
642
+ "epoch": 2.2930887372013653,
643
+ "grad_norm": 1.5921540352792363e-07,
644
+ "learning_rate": 1.1781854379977247e-07,
645
+ "loss": 0.0667,
646
+ "step": 21500
647
+ },
648
+ {
649
+ "epoch": 2.2930887372013653,
650
+ "eval_loss": 0.4866587221622467,
651
+ "eval_runtime": 63.4143,
652
+ "eval_samples_per_second": 15.564,
653
+ "eval_steps_per_second": 1.955,
654
+ "step": 21500
655
+ },
656
+ {
657
+ "epoch": 2.34641638225256,
658
+ "grad_norm": 3.954861771404694e-09,
659
+ "learning_rate": 1.0893060295790671e-07,
660
+ "loss": 0.0315,
661
+ "step": 22000
662
+ },
663
+ {
664
+ "epoch": 2.34641638225256,
665
+ "eval_loss": 0.48805058002471924,
666
+ "eval_runtime": 63.4731,
667
+ "eval_samples_per_second": 15.55,
668
+ "eval_steps_per_second": 1.954,
669
+ "step": 22000
670
+ },
671
+ {
672
+ "epoch": 2.399744027303754,
673
+ "grad_norm": 1.405540217547241e-07,
674
+ "learning_rate": 1.0004266211604096e-07,
675
+ "loss": 0.0579,
676
+ "step": 22500
677
+ },
678
+ {
679
+ "epoch": 2.399744027303754,
680
+ "eval_loss": 0.48956525325775146,
681
+ "eval_runtime": 63.5675,
682
+ "eval_samples_per_second": 15.527,
683
+ "eval_steps_per_second": 1.951,
684
+ "step": 22500
685
+ },
686
+ {
687
+ "epoch": 2.453071672354949,
688
+ "grad_norm": 2.2330443698592717e-06,
689
+ "learning_rate": 9.11547212741752e-08,
690
+ "loss": 0.0387,
691
+ "step": 23000
692
+ },
693
+ {
694
+ "epoch": 2.453071672354949,
695
+ "eval_loss": 0.48340705037117004,
696
+ "eval_runtime": 62.238,
697
+ "eval_samples_per_second": 15.858,
698
+ "eval_steps_per_second": 1.992,
699
+ "step": 23000
700
+ },
701
+ {
702
+ "epoch": 2.506399317406143,
703
+ "grad_norm": 0.000771363964304328,
704
+ "learning_rate": 8.226678043230943e-08,
705
+ "loss": 0.0495,
706
+ "step": 23500
707
+ },
708
+ {
709
+ "epoch": 2.506399317406143,
710
+ "eval_loss": 0.48572421073913574,
711
+ "eval_runtime": 62.5175,
712
+ "eval_samples_per_second": 15.788,
713
+ "eval_steps_per_second": 1.983,
714
+ "step": 23500
715
+ },
716
+ {
717
+ "epoch": 2.5597269624573378,
718
+ "grad_norm": 4.620137588062789e-06,
719
+ "learning_rate": 7.337883959044368e-08,
720
+ "loss": 0.0364,
721
+ "step": 24000
722
+ },
723
+ {
724
+ "epoch": 2.5597269624573378,
725
+ "eval_loss": 0.4860812723636627,
726
+ "eval_runtime": 64.1674,
727
+ "eval_samples_per_second": 15.382,
728
+ "eval_steps_per_second": 1.932,
729
+ "step": 24000
730
+ },
731
+ {
732
+ "epoch": 2.6130546075085324,
733
+ "grad_norm": 0.002093038521707058,
734
+ "learning_rate": 6.449089874857792e-08,
735
+ "loss": 0.0567,
736
+ "step": 24500
737
+ },
738
+ {
739
+ "epoch": 2.6130546075085324,
740
+ "eval_loss": 0.47645464539527893,
741
+ "eval_runtime": 62.5764,
742
+ "eval_samples_per_second": 15.773,
743
+ "eval_steps_per_second": 1.982,
744
+ "step": 24500
745
+ }
746
+ ],
747
+ "logging_steps": 500,
748
+ "max_steps": 28128,
749
+ "num_input_tokens_seen": 0,
750
+ "num_train_epochs": 3,
751
+ "save_steps": 500,
752
+ "stateful_callbacks": {
753
+ "TrainerControl": {
754
+ "args": {
755
+ "should_epoch_stop": false,
756
+ "should_evaluate": false,
757
+ "should_log": false,
758
+ "should_save": true,
759
+ "should_training_stop": false
760
+ },
761
+ "attributes": {}
762
+ }
763
+ },
764
+ "total_flos": 8814799122572520.0,
765
+ "train_batch_size": 2,
766
+ "trial_name": null,
767
+ "trial_params": null
768
+ }
checkpoint-24500/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:057f6bbf1ff3e34abda0863071bbf5fdcc7edb1102f04dba05c7a15087bc3396
3
+ size 4847