d071696 commited on
Commit
6d45468
·
verified ·
1 Parent(s): 7cd87e8

Training in progress, step 1000

Browse files
all_results.json CHANGED
@@ -5,9 +5,9 @@
5
  "eval_runtime": 1.234,
6
  "eval_samples_per_second": 24.312,
7
  "eval_steps_per_second": 3.242,
8
- "total_flos": 9299872197181440.0,
9
- "train_loss": 1.3567551374435425,
10
- "train_runtime": 8.4448,
11
- "train_samples_per_second": 14.21,
12
- "train_steps_per_second": 1.895
13
  }
 
5
  "eval_runtime": 1.234,
6
  "eval_samples_per_second": 24.312,
7
  "eval_steps_per_second": 3.242,
8
+ "total_flos": 9.63148132192297e+17,
9
+ "train_loss": 0.48303211886426173,
10
+ "train_runtime": 1041.8456,
11
+ "train_samples_per_second": 11.929,
12
+ "train_steps_per_second": 1.494
13
  }
model.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:975eb824ba74a090f881d73b03a5aa125239cf3bf3b7d88fbf2b3c326ec79338
3
  size 343251660
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d563541c04dbb7251d614a78ad4b90a6173da7438f314d9f54f7354ebee3ff11
3
  size 343251660
runs/Mar29_16-45-17_X5C922065N/events.out.tfevents.1711728701.X5C922065N.53009.4 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7529d8ae49ec5e28302f62b0487256bf877e9ca50f9724dd18497d7dd3b8cdf1
3
+ size 26498
train_results.json CHANGED
@@ -1,8 +1,8 @@
1
  {
2
  "epoch": 4.0,
3
- "total_flos": 9299872197181440.0,
4
- "train_loss": 1.3567551374435425,
5
- "train_runtime": 8.4448,
6
- "train_samples_per_second": 14.21,
7
- "train_steps_per_second": 1.895
8
  }
 
1
  {
2
  "epoch": 4.0,
3
+ "total_flos": 9.63148132192297e+17,
4
+ "train_loss": 0.48303211886426173,
5
+ "train_runtime": 1041.8456,
6
+ "train_samples_per_second": 11.929,
7
+ "train_steps_per_second": 1.494
8
  }
trainer_state.json CHANGED
@@ -1,36 +1,1123 @@
1
  {
2
- "best_metric": null,
3
- "best_model_checkpoint": null,
4
  "epoch": 4.0,
5
  "eval_steps": 1000,
6
- "global_step": 16,
7
  "is_hyper_param_search": false,
8
  "is_local_process_zero": true,
9
  "is_world_process_zero": true,
10
  "log_history": [
11
  {
12
- "epoch": 2.5,
13
- "grad_norm": 2.364220142364502,
14
- "learning_rate": 7.500000000000001e-05,
15
- "loss": 1.5278,
16
  "step": 10
17
  },
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
18
  {
19
  "epoch": 4.0,
20
- "step": 16,
21
- "total_flos": 9299872197181440.0,
22
- "train_loss": 1.3567551374435425,
23
- "train_runtime": 8.4448,
24
- "train_samples_per_second": 14.21,
25
- "train_steps_per_second": 1.895
26
  }
27
  ],
28
  "logging_steps": 10,
29
- "max_steps": 16,
30
  "num_input_tokens_seen": 0,
31
  "num_train_epochs": 4,
32
  "save_steps": 1000,
33
- "total_flos": 9299872197181440.0,
34
  "train_batch_size": 8,
35
  "trial_name": null,
36
  "trial_params": null
 
1
  {
2
+ "best_metric": 0.11155818402767181,
3
+ "best_model_checkpoint": "./vit-finetune-scrap/checkpoint-1000",
4
  "epoch": 4.0,
5
  "eval_steps": 1000,
6
+ "global_step": 1556,
7
  "is_hyper_param_search": false,
8
  "is_local_process_zero": true,
9
  "is_world_process_zero": true,
10
  "log_history": [
11
  {
12
+ "epoch": 0.03,
13
+ "grad_norm": 2.721052646636963,
14
+ "learning_rate": 0.0001987146529562982,
15
+ "loss": 2.3309,
16
  "step": 10
17
  },
18
+ {
19
+ "epoch": 0.05,
20
+ "grad_norm": 2.6302196979522705,
21
+ "learning_rate": 0.0001974293059125964,
22
+ "loss": 2.1693,
23
+ "step": 20
24
+ },
25
+ {
26
+ "epoch": 0.08,
27
+ "grad_norm": 3.2131187915802,
28
+ "learning_rate": 0.0001961439588688946,
29
+ "loss": 1.914,
30
+ "step": 30
31
+ },
32
+ {
33
+ "epoch": 0.1,
34
+ "grad_norm": 3.520822525024414,
35
+ "learning_rate": 0.00019485861182519281,
36
+ "loss": 1.6374,
37
+ "step": 40
38
+ },
39
+ {
40
+ "epoch": 0.13,
41
+ "grad_norm": 4.047000408172607,
42
+ "learning_rate": 0.000193573264781491,
43
+ "loss": 1.6214,
44
+ "step": 50
45
+ },
46
+ {
47
+ "epoch": 0.15,
48
+ "grad_norm": 3.1879024505615234,
49
+ "learning_rate": 0.0001922879177377892,
50
+ "loss": 1.6056,
51
+ "step": 60
52
+ },
53
+ {
54
+ "epoch": 0.18,
55
+ "grad_norm": 3.4971909523010254,
56
+ "learning_rate": 0.00019100257069408743,
57
+ "loss": 1.4073,
58
+ "step": 70
59
+ },
60
+ {
61
+ "epoch": 0.21,
62
+ "grad_norm": 5.953548908233643,
63
+ "learning_rate": 0.00018971722365038562,
64
+ "loss": 1.2913,
65
+ "step": 80
66
+ },
67
+ {
68
+ "epoch": 0.23,
69
+ "grad_norm": 3.639462471008301,
70
+ "learning_rate": 0.00018843187660668383,
71
+ "loss": 1.1548,
72
+ "step": 90
73
+ },
74
+ {
75
+ "epoch": 0.26,
76
+ "grad_norm": 5.055682182312012,
77
+ "learning_rate": 0.00018714652956298202,
78
+ "loss": 1.3169,
79
+ "step": 100
80
+ },
81
+ {
82
+ "epoch": 0.28,
83
+ "grad_norm": 2.787602186203003,
84
+ "learning_rate": 0.0001858611825192802,
85
+ "loss": 1.2729,
86
+ "step": 110
87
+ },
88
+ {
89
+ "epoch": 0.31,
90
+ "grad_norm": 5.495873928070068,
91
+ "learning_rate": 0.00018457583547557842,
92
+ "loss": 1.1431,
93
+ "step": 120
94
+ },
95
+ {
96
+ "epoch": 0.33,
97
+ "grad_norm": 2.6707160472869873,
98
+ "learning_rate": 0.0001832904884318766,
99
+ "loss": 1.0606,
100
+ "step": 130
101
+ },
102
+ {
103
+ "epoch": 0.36,
104
+ "grad_norm": 5.753376483917236,
105
+ "learning_rate": 0.00018200514138817483,
106
+ "loss": 1.0528,
107
+ "step": 140
108
+ },
109
+ {
110
+ "epoch": 0.39,
111
+ "grad_norm": 4.965968132019043,
112
+ "learning_rate": 0.000180719794344473,
113
+ "loss": 1.3918,
114
+ "step": 150
115
+ },
116
+ {
117
+ "epoch": 0.41,
118
+ "grad_norm": 4.477539539337158,
119
+ "learning_rate": 0.0001794344473007712,
120
+ "loss": 1.2681,
121
+ "step": 160
122
+ },
123
+ {
124
+ "epoch": 0.44,
125
+ "grad_norm": 6.264174938201904,
126
+ "learning_rate": 0.00017814910025706942,
127
+ "loss": 1.0713,
128
+ "step": 170
129
+ },
130
+ {
131
+ "epoch": 0.46,
132
+ "grad_norm": 3.289985179901123,
133
+ "learning_rate": 0.0001768637532133676,
134
+ "loss": 1.1408,
135
+ "step": 180
136
+ },
137
+ {
138
+ "epoch": 0.49,
139
+ "grad_norm": 4.64877986907959,
140
+ "learning_rate": 0.00017557840616966582,
141
+ "loss": 1.3037,
142
+ "step": 190
143
+ },
144
+ {
145
+ "epoch": 0.51,
146
+ "grad_norm": 3.4218943119049072,
147
+ "learning_rate": 0.000174293059125964,
148
+ "loss": 0.9621,
149
+ "step": 200
150
+ },
151
+ {
152
+ "epoch": 0.54,
153
+ "grad_norm": 5.507615566253662,
154
+ "learning_rate": 0.00017300771208226222,
155
+ "loss": 1.0243,
156
+ "step": 210
157
+ },
158
+ {
159
+ "epoch": 0.57,
160
+ "grad_norm": 4.309627532958984,
161
+ "learning_rate": 0.00017172236503856043,
162
+ "loss": 0.9739,
163
+ "step": 220
164
+ },
165
+ {
166
+ "epoch": 0.59,
167
+ "grad_norm": 4.056205749511719,
168
+ "learning_rate": 0.00017043701799485862,
169
+ "loss": 0.9158,
170
+ "step": 230
171
+ },
172
+ {
173
+ "epoch": 0.62,
174
+ "grad_norm": 3.1590564250946045,
175
+ "learning_rate": 0.00016915167095115684,
176
+ "loss": 0.8557,
177
+ "step": 240
178
+ },
179
+ {
180
+ "epoch": 0.64,
181
+ "grad_norm": 1.6367921829223633,
182
+ "learning_rate": 0.00016786632390745502,
183
+ "loss": 1.0898,
184
+ "step": 250
185
+ },
186
+ {
187
+ "epoch": 0.67,
188
+ "grad_norm": 5.508506774902344,
189
+ "learning_rate": 0.0001665809768637532,
190
+ "loss": 1.0173,
191
+ "step": 260
192
+ },
193
+ {
194
+ "epoch": 0.69,
195
+ "grad_norm": 5.602323532104492,
196
+ "learning_rate": 0.00016529562982005143,
197
+ "loss": 0.9706,
198
+ "step": 270
199
+ },
200
+ {
201
+ "epoch": 0.72,
202
+ "grad_norm": 8.27458381652832,
203
+ "learning_rate": 0.00016401028277634961,
204
+ "loss": 1.1064,
205
+ "step": 280
206
+ },
207
+ {
208
+ "epoch": 0.75,
209
+ "grad_norm": 3.5698864459991455,
210
+ "learning_rate": 0.00016272493573264783,
211
+ "loss": 0.9979,
212
+ "step": 290
213
+ },
214
+ {
215
+ "epoch": 0.77,
216
+ "grad_norm": 5.842220783233643,
217
+ "learning_rate": 0.00016143958868894602,
218
+ "loss": 1.0221,
219
+ "step": 300
220
+ },
221
+ {
222
+ "epoch": 0.8,
223
+ "grad_norm": 3.458761692047119,
224
+ "learning_rate": 0.00016015424164524423,
225
+ "loss": 0.9931,
226
+ "step": 310
227
+ },
228
+ {
229
+ "epoch": 0.82,
230
+ "grad_norm": 5.971825122833252,
231
+ "learning_rate": 0.00015886889460154242,
232
+ "loss": 1.1686,
233
+ "step": 320
234
+ },
235
+ {
236
+ "epoch": 0.85,
237
+ "grad_norm": 5.68731689453125,
238
+ "learning_rate": 0.0001575835475578406,
239
+ "loss": 0.9805,
240
+ "step": 330
241
+ },
242
+ {
243
+ "epoch": 0.87,
244
+ "grad_norm": 5.103214263916016,
245
+ "learning_rate": 0.00015629820051413882,
246
+ "loss": 0.8668,
247
+ "step": 340
248
+ },
249
+ {
250
+ "epoch": 0.9,
251
+ "grad_norm": 4.177506923675537,
252
+ "learning_rate": 0.00015501285347043704,
253
+ "loss": 1.1952,
254
+ "step": 350
255
+ },
256
+ {
257
+ "epoch": 0.93,
258
+ "grad_norm": 1.9655299186706543,
259
+ "learning_rate": 0.00015372750642673522,
260
+ "loss": 0.8981,
261
+ "step": 360
262
+ },
263
+ {
264
+ "epoch": 0.95,
265
+ "grad_norm": 4.982448577880859,
266
+ "learning_rate": 0.00015244215938303344,
267
+ "loss": 0.7721,
268
+ "step": 370
269
+ },
270
+ {
271
+ "epoch": 0.98,
272
+ "grad_norm": 5.1874775886535645,
273
+ "learning_rate": 0.00015115681233933163,
274
+ "loss": 0.97,
275
+ "step": 380
276
+ },
277
+ {
278
+ "epoch": 1.0,
279
+ "grad_norm": 4.942078590393066,
280
+ "learning_rate": 0.00014987146529562984,
281
+ "loss": 0.8415,
282
+ "step": 390
283
+ },
284
+ {
285
+ "epoch": 1.03,
286
+ "grad_norm": 3.160961389541626,
287
+ "learning_rate": 0.00014858611825192803,
288
+ "loss": 0.5367,
289
+ "step": 400
290
+ },
291
+ {
292
+ "epoch": 1.05,
293
+ "grad_norm": 5.394630432128906,
294
+ "learning_rate": 0.00014730077120822622,
295
+ "loss": 0.561,
296
+ "step": 410
297
+ },
298
+ {
299
+ "epoch": 1.08,
300
+ "grad_norm": 2.2095775604248047,
301
+ "learning_rate": 0.00014601542416452443,
302
+ "loss": 0.5548,
303
+ "step": 420
304
+ },
305
+ {
306
+ "epoch": 1.11,
307
+ "grad_norm": 4.9532575607299805,
308
+ "learning_rate": 0.00014473007712082262,
309
+ "loss": 0.6005,
310
+ "step": 430
311
+ },
312
+ {
313
+ "epoch": 1.13,
314
+ "grad_norm": 5.503066062927246,
315
+ "learning_rate": 0.00014344473007712083,
316
+ "loss": 0.514,
317
+ "step": 440
318
+ },
319
+ {
320
+ "epoch": 1.16,
321
+ "grad_norm": 5.952071189880371,
322
+ "learning_rate": 0.00014215938303341902,
323
+ "loss": 0.5386,
324
+ "step": 450
325
+ },
326
+ {
327
+ "epoch": 1.18,
328
+ "grad_norm": 4.198409557342529,
329
+ "learning_rate": 0.00014087403598971724,
330
+ "loss": 0.5937,
331
+ "step": 460
332
+ },
333
+ {
334
+ "epoch": 1.21,
335
+ "grad_norm": 4.768213272094727,
336
+ "learning_rate": 0.00013958868894601542,
337
+ "loss": 0.6504,
338
+ "step": 470
339
+ },
340
+ {
341
+ "epoch": 1.23,
342
+ "grad_norm": 4.068699359893799,
343
+ "learning_rate": 0.0001383033419023136,
344
+ "loss": 0.421,
345
+ "step": 480
346
+ },
347
+ {
348
+ "epoch": 1.26,
349
+ "grad_norm": 4.887763500213623,
350
+ "learning_rate": 0.00013701799485861185,
351
+ "loss": 0.5566,
352
+ "step": 490
353
+ },
354
+ {
355
+ "epoch": 1.29,
356
+ "grad_norm": 5.322113037109375,
357
+ "learning_rate": 0.00013573264781491004,
358
+ "loss": 0.514,
359
+ "step": 500
360
+ },
361
+ {
362
+ "epoch": 1.31,
363
+ "grad_norm": 3.480942726135254,
364
+ "learning_rate": 0.00013444730077120823,
365
+ "loss": 0.5012,
366
+ "step": 510
367
+ },
368
+ {
369
+ "epoch": 1.34,
370
+ "grad_norm": 3.2392122745513916,
371
+ "learning_rate": 0.00013316195372750644,
372
+ "loss": 0.5065,
373
+ "step": 520
374
+ },
375
+ {
376
+ "epoch": 1.36,
377
+ "grad_norm": 1.8148912191390991,
378
+ "learning_rate": 0.00013187660668380463,
379
+ "loss": 0.4932,
380
+ "step": 530
381
+ },
382
+ {
383
+ "epoch": 1.39,
384
+ "grad_norm": 1.9780988693237305,
385
+ "learning_rate": 0.00013059125964010284,
386
+ "loss": 0.6036,
387
+ "step": 540
388
+ },
389
+ {
390
+ "epoch": 1.41,
391
+ "grad_norm": 5.625373840332031,
392
+ "learning_rate": 0.00012930591259640103,
393
+ "loss": 0.564,
394
+ "step": 550
395
+ },
396
+ {
397
+ "epoch": 1.44,
398
+ "grad_norm": 9.524807929992676,
399
+ "learning_rate": 0.00012802056555269925,
400
+ "loss": 0.5695,
401
+ "step": 560
402
+ },
403
+ {
404
+ "epoch": 1.47,
405
+ "grad_norm": 1.463976263999939,
406
+ "learning_rate": 0.00012673521850899743,
407
+ "loss": 0.3198,
408
+ "step": 570
409
+ },
410
+ {
411
+ "epoch": 1.49,
412
+ "grad_norm": 6.108857154846191,
413
+ "learning_rate": 0.00012544987146529562,
414
+ "loss": 0.6759,
415
+ "step": 580
416
+ },
417
+ {
418
+ "epoch": 1.52,
419
+ "grad_norm": 1.5109316110610962,
420
+ "learning_rate": 0.00012416452442159384,
421
+ "loss": 0.4468,
422
+ "step": 590
423
+ },
424
+ {
425
+ "epoch": 1.54,
426
+ "grad_norm": 1.9603605270385742,
427
+ "learning_rate": 0.00012287917737789202,
428
+ "loss": 0.3569,
429
+ "step": 600
430
+ },
431
+ {
432
+ "epoch": 1.57,
433
+ "grad_norm": 7.527422904968262,
434
+ "learning_rate": 0.00012159383033419023,
435
+ "loss": 0.6518,
436
+ "step": 610
437
+ },
438
+ {
439
+ "epoch": 1.59,
440
+ "grad_norm": 5.3868255615234375,
441
+ "learning_rate": 0.00012030848329048843,
442
+ "loss": 0.5278,
443
+ "step": 620
444
+ },
445
+ {
446
+ "epoch": 1.62,
447
+ "grad_norm": 8.257445335388184,
448
+ "learning_rate": 0.00011902313624678665,
449
+ "loss": 0.6488,
450
+ "step": 630
451
+ },
452
+ {
453
+ "epoch": 1.65,
454
+ "grad_norm": 8.786994934082031,
455
+ "learning_rate": 0.00011773778920308484,
456
+ "loss": 0.6637,
457
+ "step": 640
458
+ },
459
+ {
460
+ "epoch": 1.67,
461
+ "grad_norm": 11.612885475158691,
462
+ "learning_rate": 0.00011645244215938304,
463
+ "loss": 0.5637,
464
+ "step": 650
465
+ },
466
+ {
467
+ "epoch": 1.7,
468
+ "grad_norm": 4.953100204467773,
469
+ "learning_rate": 0.00011516709511568124,
470
+ "loss": 0.3346,
471
+ "step": 660
472
+ },
473
+ {
474
+ "epoch": 1.72,
475
+ "grad_norm": 8.756507873535156,
476
+ "learning_rate": 0.00011388174807197945,
477
+ "loss": 0.5318,
478
+ "step": 670
479
+ },
480
+ {
481
+ "epoch": 1.75,
482
+ "grad_norm": 5.3309760093688965,
483
+ "learning_rate": 0.00011259640102827765,
484
+ "loss": 0.433,
485
+ "step": 680
486
+ },
487
+ {
488
+ "epoch": 1.77,
489
+ "grad_norm": 0.4981166422367096,
490
+ "learning_rate": 0.00011131105398457585,
491
+ "loss": 0.4548,
492
+ "step": 690
493
+ },
494
+ {
495
+ "epoch": 1.8,
496
+ "grad_norm": 7.036471366882324,
497
+ "learning_rate": 0.00011002570694087404,
498
+ "loss": 0.6301,
499
+ "step": 700
500
+ },
501
+ {
502
+ "epoch": 1.83,
503
+ "grad_norm": 5.0402021408081055,
504
+ "learning_rate": 0.00010874035989717224,
505
+ "loss": 0.6178,
506
+ "step": 710
507
+ },
508
+ {
509
+ "epoch": 1.85,
510
+ "grad_norm": 0.2094542682170868,
511
+ "learning_rate": 0.00010745501285347044,
512
+ "loss": 0.3818,
513
+ "step": 720
514
+ },
515
+ {
516
+ "epoch": 1.88,
517
+ "grad_norm": 5.399072647094727,
518
+ "learning_rate": 0.00010616966580976864,
519
+ "loss": 0.5482,
520
+ "step": 730
521
+ },
522
+ {
523
+ "epoch": 1.9,
524
+ "grad_norm": 9.017058372497559,
525
+ "learning_rate": 0.00010488431876606684,
526
+ "loss": 0.5286,
527
+ "step": 740
528
+ },
529
+ {
530
+ "epoch": 1.93,
531
+ "grad_norm": 2.5559568405151367,
532
+ "learning_rate": 0.00010359897172236503,
533
+ "loss": 0.4894,
534
+ "step": 750
535
+ },
536
+ {
537
+ "epoch": 1.95,
538
+ "grad_norm": 1.3460350036621094,
539
+ "learning_rate": 0.00010231362467866323,
540
+ "loss": 0.369,
541
+ "step": 760
542
+ },
543
+ {
544
+ "epoch": 1.98,
545
+ "grad_norm": 0.5879113078117371,
546
+ "learning_rate": 0.00010102827763496146,
547
+ "loss": 0.3088,
548
+ "step": 770
549
+ },
550
+ {
551
+ "epoch": 2.01,
552
+ "grad_norm": 1.1561224460601807,
553
+ "learning_rate": 9.974293059125965e-05,
554
+ "loss": 0.573,
555
+ "step": 780
556
+ },
557
+ {
558
+ "epoch": 2.03,
559
+ "grad_norm": 2.361337900161743,
560
+ "learning_rate": 9.845758354755785e-05,
561
+ "loss": 0.1348,
562
+ "step": 790
563
+ },
564
+ {
565
+ "epoch": 2.06,
566
+ "grad_norm": 2.3323395252227783,
567
+ "learning_rate": 9.717223650385605e-05,
568
+ "loss": 0.1292,
569
+ "step": 800
570
+ },
571
+ {
572
+ "epoch": 2.08,
573
+ "grad_norm": 0.5499300956726074,
574
+ "learning_rate": 9.588688946015425e-05,
575
+ "loss": 0.1817,
576
+ "step": 810
577
+ },
578
+ {
579
+ "epoch": 2.11,
580
+ "grad_norm": 0.2054494023323059,
581
+ "learning_rate": 9.460154241645245e-05,
582
+ "loss": 0.2232,
583
+ "step": 820
584
+ },
585
+ {
586
+ "epoch": 2.13,
587
+ "grad_norm": 2.15979266166687,
588
+ "learning_rate": 9.331619537275065e-05,
589
+ "loss": 0.2153,
590
+ "step": 830
591
+ },
592
+ {
593
+ "epoch": 2.16,
594
+ "grad_norm": 3.1036410331726074,
595
+ "learning_rate": 9.203084832904885e-05,
596
+ "loss": 0.1692,
597
+ "step": 840
598
+ },
599
+ {
600
+ "epoch": 2.19,
601
+ "grad_norm": 2.084644317626953,
602
+ "learning_rate": 9.074550128534704e-05,
603
+ "loss": 0.2034,
604
+ "step": 850
605
+ },
606
+ {
607
+ "epoch": 2.21,
608
+ "grad_norm": 2.1689724922180176,
609
+ "learning_rate": 8.946015424164524e-05,
610
+ "loss": 0.2217,
611
+ "step": 860
612
+ },
613
+ {
614
+ "epoch": 2.24,
615
+ "grad_norm": 1.0331225395202637,
616
+ "learning_rate": 8.817480719794346e-05,
617
+ "loss": 0.1232,
618
+ "step": 870
619
+ },
620
+ {
621
+ "epoch": 2.26,
622
+ "grad_norm": 3.129354953765869,
623
+ "learning_rate": 8.688946015424166e-05,
624
+ "loss": 0.1363,
625
+ "step": 880
626
+ },
627
+ {
628
+ "epoch": 2.29,
629
+ "grad_norm": 0.653751015663147,
630
+ "learning_rate": 8.560411311053986e-05,
631
+ "loss": 0.1845,
632
+ "step": 890
633
+ },
634
+ {
635
+ "epoch": 2.31,
636
+ "grad_norm": 0.20718339085578918,
637
+ "learning_rate": 8.431876606683805e-05,
638
+ "loss": 0.1638,
639
+ "step": 900
640
+ },
641
+ {
642
+ "epoch": 2.34,
643
+ "grad_norm": 5.0227274894714355,
644
+ "learning_rate": 8.303341902313625e-05,
645
+ "loss": 0.1423,
646
+ "step": 910
647
+ },
648
+ {
649
+ "epoch": 2.37,
650
+ "grad_norm": 0.7187924385070801,
651
+ "learning_rate": 8.174807197943445e-05,
652
+ "loss": 0.1702,
653
+ "step": 920
654
+ },
655
+ {
656
+ "epoch": 2.39,
657
+ "grad_norm": 1.2977266311645508,
658
+ "learning_rate": 8.046272493573265e-05,
659
+ "loss": 0.1354,
660
+ "step": 930
661
+ },
662
+ {
663
+ "epoch": 2.42,
664
+ "grad_norm": 2.3543667793273926,
665
+ "learning_rate": 7.917737789203086e-05,
666
+ "loss": 0.2209,
667
+ "step": 940
668
+ },
669
+ {
670
+ "epoch": 2.44,
671
+ "grad_norm": 0.8430781364440918,
672
+ "learning_rate": 7.789203084832905e-05,
673
+ "loss": 0.1487,
674
+ "step": 950
675
+ },
676
+ {
677
+ "epoch": 2.47,
678
+ "grad_norm": 0.08762349933385849,
679
+ "learning_rate": 7.660668380462725e-05,
680
+ "loss": 0.1038,
681
+ "step": 960
682
+ },
683
+ {
684
+ "epoch": 2.49,
685
+ "grad_norm": 8.408522605895996,
686
+ "learning_rate": 7.532133676092545e-05,
687
+ "loss": 0.2402,
688
+ "step": 970
689
+ },
690
+ {
691
+ "epoch": 2.52,
692
+ "grad_norm": 1.173913836479187,
693
+ "learning_rate": 7.403598971722365e-05,
694
+ "loss": 0.0641,
695
+ "step": 980
696
+ },
697
+ {
698
+ "epoch": 2.54,
699
+ "grad_norm": 7.908231735229492,
700
+ "learning_rate": 7.275064267352186e-05,
701
+ "loss": 0.2347,
702
+ "step": 990
703
+ },
704
+ {
705
+ "epoch": 2.57,
706
+ "grad_norm": 3.18058180809021,
707
+ "learning_rate": 7.146529562982006e-05,
708
+ "loss": 0.1326,
709
+ "step": 1000
710
+ },
711
+ {
712
+ "epoch": 2.57,
713
+ "eval_accuracy": 0.9694238815577728,
714
+ "eval_loss": 0.11155818402767181,
715
+ "eval_runtime": 52.5851,
716
+ "eval_samples_per_second": 59.085,
717
+ "eval_steps_per_second": 7.398,
718
+ "step": 1000
719
+ },
720
+ {
721
+ "epoch": 2.6,
722
+ "grad_norm": 0.4821953773498535,
723
+ "learning_rate": 7.017994858611826e-05,
724
+ "loss": 0.258,
725
+ "step": 1010
726
+ },
727
+ {
728
+ "epoch": 2.62,
729
+ "grad_norm": 4.647073268890381,
730
+ "learning_rate": 6.889460154241646e-05,
731
+ "loss": 0.1106,
732
+ "step": 1020
733
+ },
734
+ {
735
+ "epoch": 2.65,
736
+ "grad_norm": 0.07687141746282578,
737
+ "learning_rate": 6.760925449871466e-05,
738
+ "loss": 0.1768,
739
+ "step": 1030
740
+ },
741
+ {
742
+ "epoch": 2.67,
743
+ "grad_norm": 0.8537989854812622,
744
+ "learning_rate": 6.632390745501286e-05,
745
+ "loss": 0.1321,
746
+ "step": 1040
747
+ },
748
+ {
749
+ "epoch": 2.7,
750
+ "grad_norm": 1.6428909301757812,
751
+ "learning_rate": 6.503856041131106e-05,
752
+ "loss": 0.2679,
753
+ "step": 1050
754
+ },
755
+ {
756
+ "epoch": 2.72,
757
+ "grad_norm": 0.4707659184932709,
758
+ "learning_rate": 6.375321336760925e-05,
759
+ "loss": 0.192,
760
+ "step": 1060
761
+ },
762
+ {
763
+ "epoch": 2.75,
764
+ "grad_norm": 0.09739229083061218,
765
+ "learning_rate": 6.246786632390745e-05,
766
+ "loss": 0.2501,
767
+ "step": 1070
768
+ },
769
+ {
770
+ "epoch": 2.78,
771
+ "grad_norm": 2.0249221324920654,
772
+ "learning_rate": 6.118251928020567e-05,
773
+ "loss": 0.1988,
774
+ "step": 1080
775
+ },
776
+ {
777
+ "epoch": 2.8,
778
+ "grad_norm": 0.08042796701192856,
779
+ "learning_rate": 5.989717223650386e-05,
780
+ "loss": 0.04,
781
+ "step": 1090
782
+ },
783
+ {
784
+ "epoch": 2.83,
785
+ "grad_norm": 0.40489840507507324,
786
+ "learning_rate": 5.861182519280206e-05,
787
+ "loss": 0.1326,
788
+ "step": 1100
789
+ },
790
+ {
791
+ "epoch": 2.85,
792
+ "grad_norm": 8.32421875,
793
+ "learning_rate": 5.732647814910026e-05,
794
+ "loss": 0.1881,
795
+ "step": 1110
796
+ },
797
+ {
798
+ "epoch": 2.88,
799
+ "grad_norm": 0.3356345295906067,
800
+ "learning_rate": 5.604113110539846e-05,
801
+ "loss": 0.1638,
802
+ "step": 1120
803
+ },
804
+ {
805
+ "epoch": 2.9,
806
+ "grad_norm": 2.0262017250061035,
807
+ "learning_rate": 5.475578406169666e-05,
808
+ "loss": 0.0901,
809
+ "step": 1130
810
+ },
811
+ {
812
+ "epoch": 2.93,
813
+ "grad_norm": 5.13381290435791,
814
+ "learning_rate": 5.347043701799486e-05,
815
+ "loss": 0.1947,
816
+ "step": 1140
817
+ },
818
+ {
819
+ "epoch": 2.96,
820
+ "grad_norm": 4.401228904724121,
821
+ "learning_rate": 5.218508997429307e-05,
822
+ "loss": 0.1105,
823
+ "step": 1150
824
+ },
825
+ {
826
+ "epoch": 2.98,
827
+ "grad_norm": 3.711754083633423,
828
+ "learning_rate": 5.089974293059127e-05,
829
+ "loss": 0.082,
830
+ "step": 1160
831
+ },
832
+ {
833
+ "epoch": 3.01,
834
+ "grad_norm": 0.4783603847026825,
835
+ "learning_rate": 4.961439588688946e-05,
836
+ "loss": 0.1223,
837
+ "step": 1170
838
+ },
839
+ {
840
+ "epoch": 3.03,
841
+ "grad_norm": 6.101786136627197,
842
+ "learning_rate": 4.8329048843187664e-05,
843
+ "loss": 0.0386,
844
+ "step": 1180
845
+ },
846
+ {
847
+ "epoch": 3.06,
848
+ "grad_norm": 0.09219735115766525,
849
+ "learning_rate": 4.7043701799485865e-05,
850
+ "loss": 0.046,
851
+ "step": 1190
852
+ },
853
+ {
854
+ "epoch": 3.08,
855
+ "grad_norm": 0.09228511899709702,
856
+ "learning_rate": 4.5758354755784066e-05,
857
+ "loss": 0.0179,
858
+ "step": 1200
859
+ },
860
+ {
861
+ "epoch": 3.11,
862
+ "grad_norm": 0.06705611199140549,
863
+ "learning_rate": 4.447300771208227e-05,
864
+ "loss": 0.0401,
865
+ "step": 1210
866
+ },
867
+ {
868
+ "epoch": 3.14,
869
+ "grad_norm": 0.05702489614486694,
870
+ "learning_rate": 4.318766066838046e-05,
871
+ "loss": 0.0222,
872
+ "step": 1220
873
+ },
874
+ {
875
+ "epoch": 3.16,
876
+ "grad_norm": 0.051934726536273956,
877
+ "learning_rate": 4.190231362467866e-05,
878
+ "loss": 0.0249,
879
+ "step": 1230
880
+ },
881
+ {
882
+ "epoch": 3.19,
883
+ "grad_norm": 0.05382351949810982,
884
+ "learning_rate": 4.0616966580976864e-05,
885
+ "loss": 0.017,
886
+ "step": 1240
887
+ },
888
+ {
889
+ "epoch": 3.21,
890
+ "grad_norm": 0.10244094580411911,
891
+ "learning_rate": 3.9331619537275065e-05,
892
+ "loss": 0.1425,
893
+ "step": 1250
894
+ },
895
+ {
896
+ "epoch": 3.24,
897
+ "grad_norm": 0.04559057578444481,
898
+ "learning_rate": 3.8046272493573266e-05,
899
+ "loss": 0.0188,
900
+ "step": 1260
901
+ },
902
+ {
903
+ "epoch": 3.26,
904
+ "grad_norm": 1.9016327857971191,
905
+ "learning_rate": 3.676092544987147e-05,
906
+ "loss": 0.0196,
907
+ "step": 1270
908
+ },
909
+ {
910
+ "epoch": 3.29,
911
+ "grad_norm": 0.06497751176357269,
912
+ "learning_rate": 3.547557840616967e-05,
913
+ "loss": 0.0161,
914
+ "step": 1280
915
+ },
916
+ {
917
+ "epoch": 3.32,
918
+ "grad_norm": 0.05229075625538826,
919
+ "learning_rate": 3.419023136246787e-05,
920
+ "loss": 0.0165,
921
+ "step": 1290
922
+ },
923
+ {
924
+ "epoch": 3.34,
925
+ "grad_norm": 0.04599655419588089,
926
+ "learning_rate": 3.2904884318766064e-05,
927
+ "loss": 0.0338,
928
+ "step": 1300
929
+ },
930
+ {
931
+ "epoch": 3.37,
932
+ "grad_norm": 0.054148729890584946,
933
+ "learning_rate": 3.161953727506427e-05,
934
+ "loss": 0.0401,
935
+ "step": 1310
936
+ },
937
+ {
938
+ "epoch": 3.39,
939
+ "grad_norm": 0.135112926363945,
940
+ "learning_rate": 3.033419023136247e-05,
941
+ "loss": 0.0386,
942
+ "step": 1320
943
+ },
944
+ {
945
+ "epoch": 3.42,
946
+ "grad_norm": 0.05881468951702118,
947
+ "learning_rate": 2.9048843187660668e-05,
948
+ "loss": 0.0526,
949
+ "step": 1330
950
+ },
951
+ {
952
+ "epoch": 3.44,
953
+ "grad_norm": 0.11401781439781189,
954
+ "learning_rate": 2.7763496143958872e-05,
955
+ "loss": 0.0652,
956
+ "step": 1340
957
+ },
958
+ {
959
+ "epoch": 3.47,
960
+ "grad_norm": 0.6476575136184692,
961
+ "learning_rate": 2.647814910025707e-05,
962
+ "loss": 0.0772,
963
+ "step": 1350
964
+ },
965
+ {
966
+ "epoch": 3.5,
967
+ "grad_norm": 0.0521862767636776,
968
+ "learning_rate": 2.519280205655527e-05,
969
+ "loss": 0.0166,
970
+ "step": 1360
971
+ },
972
+ {
973
+ "epoch": 3.52,
974
+ "grad_norm": 0.05607061833143234,
975
+ "learning_rate": 2.3907455012853472e-05,
976
+ "loss": 0.0138,
977
+ "step": 1370
978
+ },
979
+ {
980
+ "epoch": 3.55,
981
+ "grad_norm": 0.05825699120759964,
982
+ "learning_rate": 2.262210796915167e-05,
983
+ "loss": 0.015,
984
+ "step": 1380
985
+ },
986
+ {
987
+ "epoch": 3.57,
988
+ "grad_norm": 4.6812334060668945,
989
+ "learning_rate": 2.133676092544987e-05,
990
+ "loss": 0.1053,
991
+ "step": 1390
992
+ },
993
+ {
994
+ "epoch": 3.6,
995
+ "grad_norm": 0.6198139786720276,
996
+ "learning_rate": 2.0051413881748076e-05,
997
+ "loss": 0.0178,
998
+ "step": 1400
999
+ },
1000
+ {
1001
+ "epoch": 3.62,
1002
+ "grad_norm": 0.05886732041835785,
1003
+ "learning_rate": 1.8766066838046273e-05,
1004
+ "loss": 0.0278,
1005
+ "step": 1410
1006
+ },
1007
+ {
1008
+ "epoch": 3.65,
1009
+ "grad_norm": 0.673959493637085,
1010
+ "learning_rate": 1.7480719794344475e-05,
1011
+ "loss": 0.0769,
1012
+ "step": 1420
1013
+ },
1014
+ {
1015
+ "epoch": 3.68,
1016
+ "grad_norm": 0.31164562702178955,
1017
+ "learning_rate": 1.6195372750642672e-05,
1018
+ "loss": 0.1499,
1019
+ "step": 1430
1020
+ },
1021
+ {
1022
+ "epoch": 3.7,
1023
+ "grad_norm": 0.2713916599750519,
1024
+ "learning_rate": 1.4910025706940875e-05,
1025
+ "loss": 0.0444,
1026
+ "step": 1440
1027
+ },
1028
+ {
1029
+ "epoch": 3.73,
1030
+ "grad_norm": 2.0257036685943604,
1031
+ "learning_rate": 1.3624678663239075e-05,
1032
+ "loss": 0.071,
1033
+ "step": 1450
1034
+ },
1035
+ {
1036
+ "epoch": 3.75,
1037
+ "grad_norm": 0.24306029081344604,
1038
+ "learning_rate": 1.2339331619537276e-05,
1039
+ "loss": 0.0114,
1040
+ "step": 1460
1041
+ },
1042
+ {
1043
+ "epoch": 3.78,
1044
+ "grad_norm": 0.2247108817100525,
1045
+ "learning_rate": 1.1053984575835475e-05,
1046
+ "loss": 0.0132,
1047
+ "step": 1470
1048
+ },
1049
+ {
1050
+ "epoch": 3.8,
1051
+ "grad_norm": 0.056268274784088135,
1052
+ "learning_rate": 9.768637532133676e-06,
1053
+ "loss": 0.0827,
1054
+ "step": 1480
1055
+ },
1056
+ {
1057
+ "epoch": 3.83,
1058
+ "grad_norm": 0.04434029012918472,
1059
+ "learning_rate": 8.483290488431877e-06,
1060
+ "loss": 0.0138,
1061
+ "step": 1490
1062
+ },
1063
+ {
1064
+ "epoch": 3.86,
1065
+ "grad_norm": 0.06419169157743454,
1066
+ "learning_rate": 7.197943444730078e-06,
1067
+ "loss": 0.077,
1068
+ "step": 1500
1069
+ },
1070
+ {
1071
+ "epoch": 3.88,
1072
+ "grad_norm": 0.11697979271411896,
1073
+ "learning_rate": 5.912596401028278e-06,
1074
+ "loss": 0.0206,
1075
+ "step": 1510
1076
+ },
1077
+ {
1078
+ "epoch": 3.91,
1079
+ "grad_norm": 1.25772225856781,
1080
+ "learning_rate": 4.627249357326478e-06,
1081
+ "loss": 0.02,
1082
+ "step": 1520
1083
+ },
1084
+ {
1085
+ "epoch": 3.93,
1086
+ "grad_norm": 0.12491010874509811,
1087
+ "learning_rate": 3.3419023136246787e-06,
1088
+ "loss": 0.0257,
1089
+ "step": 1530
1090
+ },
1091
+ {
1092
+ "epoch": 3.96,
1093
+ "grad_norm": 3.4478936195373535,
1094
+ "learning_rate": 2.056555269922879e-06,
1095
+ "loss": 0.0366,
1096
+ "step": 1540
1097
+ },
1098
+ {
1099
+ "epoch": 3.98,
1100
+ "grad_norm": 0.0410100519657135,
1101
+ "learning_rate": 7.712082262210797e-07,
1102
+ "loss": 0.0525,
1103
+ "step": 1550
1104
+ },
1105
  {
1106
  "epoch": 4.0,
1107
+ "step": 1556,
1108
+ "total_flos": 9.63148132192297e+17,
1109
+ "train_loss": 0.48303211886426173,
1110
+ "train_runtime": 1041.8456,
1111
+ "train_samples_per_second": 11.929,
1112
+ "train_steps_per_second": 1.494
1113
  }
1114
  ],
1115
  "logging_steps": 10,
1116
+ "max_steps": 1556,
1117
  "num_input_tokens_seen": 0,
1118
  "num_train_epochs": 4,
1119
  "save_steps": 1000,
1120
+ "total_flos": 9.63148132192297e+17,
1121
  "train_batch_size": 8,
1122
  "trial_name": null,
1123
  "trial_params": null