Hugging Face's logo

vislang
/
extractive_models

Safetensors
Model card Files Community
vislang commited on
Commit
84514b1
·
verified ·
1 Parent(s): ed49a0d

Upload folder using huggingface_hub

Browse files
This view is limited to 50 files because it contains too many changes.   See raw diff
Files changed (50) hide show
  1. cvnet_small_v1_global/checkpoint-1000/model.safetensors +3 -0
  2. cvnet_small_v1_global/checkpoint-1000/optimizer.pt +3 -0
  3. cvnet_small_v1_global/checkpoint-1000/rng_state_0.pth +3 -0
  4. cvnet_small_v1_global/checkpoint-1000/rng_state_1.pth +3 -0
  5. cvnet_small_v1_global/checkpoint-1000/rng_state_2.pth +3 -0
  6. cvnet_small_v1_global/checkpoint-1000/rng_state_3.pth +3 -0
  7. cvnet_small_v1_global/checkpoint-1000/rng_state_4.pth +3 -0
  8. cvnet_small_v1_global/checkpoint-1000/rng_state_5.pth +3 -0
  9. cvnet_small_v1_global/checkpoint-1000/rng_state_6.pth +3 -0
  10. cvnet_small_v1_global/checkpoint-1000/rng_state_7.pth +3 -0
  11. cvnet_small_v1_global/checkpoint-1000/scheduler.pt +3 -0
  12. cvnet_small_v1_global/checkpoint-1000/trainer_state.json +621 -0
  13. cvnet_small_v1_global/checkpoint-1000/training_args.bin +3 -0
  14. cvnet_small_v1_global/checkpoint-2000/model.safetensors +3 -0
  15. cvnet_small_v1_global/checkpoint-2000/optimizer.pt +3 -0
  16. cvnet_small_v1_global/checkpoint-2000/rng_state_0.pth +3 -0
  17. cvnet_small_v1_global/checkpoint-2000/rng_state_1.pth +3 -0
  18. cvnet_small_v1_global/checkpoint-2000/rng_state_2.pth +3 -0
  19. cvnet_small_v1_global/checkpoint-2000/rng_state_3.pth +3 -0
  20. cvnet_small_v1_global/checkpoint-2000/rng_state_4.pth +3 -0
  21. cvnet_small_v1_global/checkpoint-2000/rng_state_5.pth +3 -0
  22. cvnet_small_v1_global/checkpoint-2000/rng_state_6.pth +3 -0
  23. cvnet_small_v1_global/checkpoint-2000/rng_state_7.pth +3 -0
  24. cvnet_small_v1_global/checkpoint-2000/scheduler.pt +3 -0
  25. cvnet_small_v1_global/checkpoint-2000/trainer_state.json +1221 -0
  26. cvnet_small_v1_global/checkpoint-2000/training_args.bin +3 -0
  27. cvnet_small_v1_global/checkpoint-3000/model.safetensors +3 -0
  28. cvnet_small_v1_global/checkpoint-3000/optimizer.pt +3 -0
  29. cvnet_small_v1_global/checkpoint-3000/rng_state_0.pth +3 -0
  30. cvnet_small_v1_global/checkpoint-3000/rng_state_1.pth +3 -0
  31. cvnet_small_v1_global/checkpoint-3000/rng_state_2.pth +3 -0
  32. cvnet_small_v1_global/checkpoint-3000/rng_state_3.pth +3 -0
  33. cvnet_small_v1_global/checkpoint-3000/rng_state_4.pth +3 -0
  34. cvnet_small_v1_global/checkpoint-3000/rng_state_5.pth +3 -0
  35. cvnet_small_v1_global/checkpoint-3000/rng_state_6.pth +3 -0
  36. cvnet_small_v1_global/checkpoint-3000/rng_state_7.pth +3 -0
  37. cvnet_small_v1_global/checkpoint-3000/scheduler.pt +3 -0
  38. cvnet_small_v1_global/checkpoint-3000/trainer_state.json +1821 -0
  39. cvnet_small_v1_global/checkpoint-3000/training_args.bin +3 -0
  40. cvnet_small_v1_global/checkpoint-4000/model.safetensors +3 -0
  41. cvnet_small_v1_global/checkpoint-4000/optimizer.pt +3 -0
  42. cvnet_small_v1_global/checkpoint-4000/rng_state_0.pth +3 -0
  43. cvnet_small_v1_global/checkpoint-4000/rng_state_1.pth +3 -0
  44. cvnet_small_v1_global/checkpoint-4000/rng_state_2.pth +3 -0
  45. cvnet_small_v1_global/checkpoint-4000/rng_state_3.pth +3 -0
  46. cvnet_small_v1_global/checkpoint-4000/rng_state_4.pth +3 -0
  47. cvnet_small_v1_global/checkpoint-4000/rng_state_5.pth +3 -0
  48. cvnet_small_v1_global/checkpoint-4000/rng_state_6.pth +3 -0
  49. cvnet_small_v1_global/checkpoint-4000/rng_state_7.pth +3 -0
  50. cvnet_small_v1_global/checkpoint-4000/scheduler.pt +3 -0
cvnet_small_v1_global/checkpoint-1000/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:94b64e1aea865196e9b19a03a0b7e3ac2eec8b15357819313767513259f8a27d
3
+ size 237860264
cvnet_small_v1_global/checkpoint-1000/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e3087e6b510df04a6a96e86fb6a37ae576f3f26f7ce8e9068e3c726c8440dd3f
3
+ size 475804282
cvnet_small_v1_global/checkpoint-1000/rng_state_0.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7f68045b77ef059786b9d730ec7bfcab3532c91c560ed23338e0d4d518b2be5e
3
+ size 15920
cvnet_small_v1_global/checkpoint-1000/rng_state_1.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:164a39302aced8efcfa4d5b883ab7d4f208ce2d3aaf57fa0fff0dc5bc35e9b53
3
+ size 15920
cvnet_small_v1_global/checkpoint-1000/rng_state_2.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:74928fb7ef7944af9aedd43d3456cb7cc320ccd38e5d7989ea54cf81f2bdc4ac
3
+ size 15920
cvnet_small_v1_global/checkpoint-1000/rng_state_3.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7390fd7db05b7bd4f309fe3c16e3133e51c0b7bbf36881c88fd513b9ebb7b75a
3
+ size 15920
cvnet_small_v1_global/checkpoint-1000/rng_state_4.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:315c200f1cd4bb6a4a593accb8ba546796343253845919cb5b2b52089686830d
3
+ size 15920
cvnet_small_v1_global/checkpoint-1000/rng_state_5.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:561c368392fbdb6e8a7e5e086de59545625d80b821b5c72a4024b1f47b1390bb
3
+ size 15920
cvnet_small_v1_global/checkpoint-1000/rng_state_6.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:56a358b4a84de8c0fefda163466d1b792e3d23a409d2f7f54fdd9b3535793d39
3
+ size 15920
cvnet_small_v1_global/checkpoint-1000/rng_state_7.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:17ad99f8730fc68cc1e3438467270d0c73427583fe0bffeb5bca595169156c2f
3
+ size 15920
cvnet_small_v1_global/checkpoint-1000/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c34ba9e6f50c2313df84e91c7dfed2be6b77dec7394adba34d3bd12d40d5ba2d
3
+ size 1064
cvnet_small_v1_global/checkpoint-1000/trainer_state.json ADDED
@@ -0,0 +1,621 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": null,
3
+ "best_model_checkpoint": null,
4
+ "epoch": 0.3373534620899047,
5
+ "eval_steps": 500,
6
+ "global_step": 1000,
7
+ "is_hyper_param_search": false,
8
+ "is_local_process_zero": true,
9
+ "is_world_process_zero": true,
10
+ "log_history": [
11
+ {
12
+ "epoch": 0.0,
13
+ "learning_rate": 5e-05,
14
+ "loss": 0.7081,
15
+ "step": 10
16
+ },
17
+ {
18
+ "epoch": 0.01,
19
+ "learning_rate": 5e-05,
20
+ "loss": 0.6774,
21
+ "step": 20
22
+ },
23
+ {
24
+ "epoch": 0.01,
25
+ "learning_rate": 5e-05,
26
+ "loss": 0.6781,
27
+ "step": 30
28
+ },
29
+ {
30
+ "epoch": 0.01,
31
+ "learning_rate": 5e-05,
32
+ "loss": 0.6772,
33
+ "step": 40
34
+ },
35
+ {
36
+ "epoch": 0.02,
37
+ "learning_rate": 5e-05,
38
+ "loss": 0.6802,
39
+ "step": 50
40
+ },
41
+ {
42
+ "epoch": 0.02,
43
+ "learning_rate": 5e-05,
44
+ "loss": 0.6749,
45
+ "step": 60
46
+ },
47
+ {
48
+ "epoch": 0.02,
49
+ "learning_rate": 5e-05,
50
+ "loss": 0.6745,
51
+ "step": 70
52
+ },
53
+ {
54
+ "epoch": 0.03,
55
+ "learning_rate": 5e-05,
56
+ "loss": 0.6729,
57
+ "step": 80
58
+ },
59
+ {
60
+ "epoch": 0.03,
61
+ "learning_rate": 5e-05,
62
+ "loss": 0.6586,
63
+ "step": 90
64
+ },
65
+ {
66
+ "epoch": 0.03,
67
+ "learning_rate": 5e-05,
68
+ "loss": 0.636,
69
+ "step": 100
70
+ },
71
+ {
72
+ "epoch": 0.04,
73
+ "learning_rate": 5e-05,
74
+ "loss": 0.6237,
75
+ "step": 110
76
+ },
77
+ {
78
+ "epoch": 0.04,
79
+ "learning_rate": 5e-05,
80
+ "loss": 0.6098,
81
+ "step": 120
82
+ },
83
+ {
84
+ "epoch": 0.04,
85
+ "learning_rate": 5e-05,
86
+ "loss": 0.5978,
87
+ "step": 130
88
+ },
89
+ {
90
+ "epoch": 0.05,
91
+ "learning_rate": 5e-05,
92
+ "loss": 0.602,
93
+ "step": 140
94
+ },
95
+ {
96
+ "epoch": 0.05,
97
+ "learning_rate": 5e-05,
98
+ "loss": 0.5959,
99
+ "step": 150
100
+ },
101
+ {
102
+ "epoch": 0.05,
103
+ "learning_rate": 5e-05,
104
+ "loss": 0.5906,
105
+ "step": 160
106
+ },
107
+ {
108
+ "epoch": 0.06,
109
+ "learning_rate": 5e-05,
110
+ "loss": 0.5778,
111
+ "step": 170
112
+ },
113
+ {
114
+ "epoch": 0.06,
115
+ "learning_rate": 5e-05,
116
+ "loss": 0.5795,
117
+ "step": 180
118
+ },
119
+ {
120
+ "epoch": 0.06,
121
+ "learning_rate": 5e-05,
122
+ "loss": 0.5883,
123
+ "step": 190
124
+ },
125
+ {
126
+ "epoch": 0.07,
127
+ "learning_rate": 5e-05,
128
+ "loss": 0.604,
129
+ "step": 200
130
+ },
131
+ {
132
+ "epoch": 0.07,
133
+ "learning_rate": 5e-05,
134
+ "loss": 0.5882,
135
+ "step": 210
136
+ },
137
+ {
138
+ "epoch": 0.07,
139
+ "learning_rate": 5e-05,
140
+ "loss": 0.5747,
141
+ "step": 220
142
+ },
143
+ {
144
+ "epoch": 0.08,
145
+ "learning_rate": 5e-05,
146
+ "loss": 0.5745,
147
+ "step": 230
148
+ },
149
+ {
150
+ "epoch": 0.08,
151
+ "learning_rate": 5e-05,
152
+ "loss": 0.5846,
153
+ "step": 240
154
+ },
155
+ {
156
+ "epoch": 0.08,
157
+ "learning_rate": 5e-05,
158
+ "loss": 0.5823,
159
+ "step": 250
160
+ },
161
+ {
162
+ "epoch": 0.09,
163
+ "learning_rate": 5e-05,
164
+ "loss": 0.5722,
165
+ "step": 260
166
+ },
167
+ {
168
+ "epoch": 0.09,
169
+ "learning_rate": 5e-05,
170
+ "loss": 0.5599,
171
+ "step": 270
172
+ },
173
+ {
174
+ "epoch": 0.09,
175
+ "learning_rate": 5e-05,
176
+ "loss": 0.5716,
177
+ "step": 280
178
+ },
179
+ {
180
+ "epoch": 0.1,
181
+ "learning_rate": 5e-05,
182
+ "loss": 0.5832,
183
+ "step": 290
184
+ },
185
+ {
186
+ "epoch": 0.1,
187
+ "learning_rate": 5e-05,
188
+ "loss": 0.5745,
189
+ "step": 300
190
+ },
191
+ {
192
+ "epoch": 0.1,
193
+ "learning_rate": 5e-05,
194
+ "loss": 0.5691,
195
+ "step": 310
196
+ },
197
+ {
198
+ "epoch": 0.11,
199
+ "learning_rate": 5e-05,
200
+ "loss": 0.5711,
201
+ "step": 320
202
+ },
203
+ {
204
+ "epoch": 0.11,
205
+ "learning_rate": 5e-05,
206
+ "loss": 0.5821,
207
+ "step": 330
208
+ },
209
+ {
210
+ "epoch": 0.11,
211
+ "learning_rate": 5e-05,
212
+ "loss": 0.5557,
213
+ "step": 340
214
+ },
215
+ {
216
+ "epoch": 0.12,
217
+ "learning_rate": 5e-05,
218
+ "loss": 0.5581,
219
+ "step": 350
220
+ },
221
+ {
222
+ "epoch": 0.12,
223
+ "learning_rate": 5e-05,
224
+ "loss": 0.5761,
225
+ "step": 360
226
+ },
227
+ {
228
+ "epoch": 0.12,
229
+ "learning_rate": 5e-05,
230
+ "loss": 0.5643,
231
+ "step": 370
232
+ },
233
+ {
234
+ "epoch": 0.13,
235
+ "learning_rate": 5e-05,
236
+ "loss": 0.5745,
237
+ "step": 380
238
+ },
239
+ {
240
+ "epoch": 0.13,
241
+ "learning_rate": 5e-05,
242
+ "loss": 0.5702,
243
+ "step": 390
244
+ },
245
+ {
246
+ "epoch": 0.13,
247
+ "learning_rate": 5e-05,
248
+ "loss": 0.5648,
249
+ "step": 400
250
+ },
251
+ {
252
+ "epoch": 0.14,
253
+ "learning_rate": 5e-05,
254
+ "loss": 0.5691,
255
+ "step": 410
256
+ },
257
+ {
258
+ "epoch": 0.14,
259
+ "learning_rate": 5e-05,
260
+ "loss": 0.556,
261
+ "step": 420
262
+ },
263
+ {
264
+ "epoch": 0.15,
265
+ "learning_rate": 5e-05,
266
+ "loss": 0.5578,
267
+ "step": 430
268
+ },
269
+ {
270
+ "epoch": 0.15,
271
+ "learning_rate": 5e-05,
272
+ "loss": 0.564,
273
+ "step": 440
274
+ },
275
+ {
276
+ "epoch": 0.15,
277
+ "learning_rate": 5e-05,
278
+ "loss": 0.5562,
279
+ "step": 450
280
+ },
281
+ {
282
+ "epoch": 0.16,
283
+ "learning_rate": 5e-05,
284
+ "loss": 0.559,
285
+ "step": 460
286
+ },
287
+ {
288
+ "epoch": 0.16,
289
+ "learning_rate": 5e-05,
290
+ "loss": 0.565,
291
+ "step": 470
292
+ },
293
+ {
294
+ "epoch": 0.16,
295
+ "learning_rate": 5e-05,
296
+ "loss": 0.5508,
297
+ "step": 480
298
+ },
299
+ {
300
+ "epoch": 0.17,
301
+ "learning_rate": 5e-05,
302
+ "loss": 0.5588,
303
+ "step": 490
304
+ },
305
+ {
306
+ "epoch": 0.17,
307
+ "learning_rate": 5e-05,
308
+ "loss": 0.5584,
309
+ "step": 500
310
+ },
311
+ {
312
+ "epoch": 0.17,
313
+ "learning_rate": 5e-05,
314
+ "loss": 0.5555,
315
+ "step": 510
316
+ },
317
+ {
318
+ "epoch": 0.18,
319
+ "learning_rate": 5e-05,
320
+ "loss": 0.5597,
321
+ "step": 520
322
+ },
323
+ {
324
+ "epoch": 0.18,
325
+ "learning_rate": 5e-05,
326
+ "loss": 0.5483,
327
+ "step": 530
328
+ },
329
+ {
330
+ "epoch": 0.18,
331
+ "learning_rate": 5e-05,
332
+ "loss": 0.5621,
333
+ "step": 540
334
+ },
335
+ {
336
+ "epoch": 0.19,
337
+ "learning_rate": 5e-05,
338
+ "loss": 0.5568,
339
+ "step": 550
340
+ },
341
+ {
342
+ "epoch": 0.19,
343
+ "learning_rate": 5e-05,
344
+ "loss": 0.5527,
345
+ "step": 560
346
+ },
347
+ {
348
+ "epoch": 0.19,
349
+ "learning_rate": 5e-05,
350
+ "loss": 0.5524,
351
+ "step": 570
352
+ },
353
+ {
354
+ "epoch": 0.2,
355
+ "learning_rate": 5e-05,
356
+ "loss": 0.5458,
357
+ "step": 580
358
+ },
359
+ {
360
+ "epoch": 0.2,
361
+ "learning_rate": 5e-05,
362
+ "loss": 0.5426,
363
+ "step": 590
364
+ },
365
+ {
366
+ "epoch": 0.2,
367
+ "learning_rate": 5e-05,
368
+ "loss": 0.5524,
369
+ "step": 600
370
+ },
371
+ {
372
+ "epoch": 0.21,
373
+ "learning_rate": 5e-05,
374
+ "loss": 0.5553,
375
+ "step": 610
376
+ },
377
+ {
378
+ "epoch": 0.21,
379
+ "learning_rate": 5e-05,
380
+ "loss": 0.5582,
381
+ "step": 620
382
+ },
383
+ {
384
+ "epoch": 0.21,
385
+ "learning_rate": 5e-05,
386
+ "loss": 0.5518,
387
+ "step": 630
388
+ },
389
+ {
390
+ "epoch": 0.22,
391
+ "learning_rate": 5e-05,
392
+ "loss": 0.5485,
393
+ "step": 640
394
+ },
395
+ {
396
+ "epoch": 0.22,
397
+ "learning_rate": 5e-05,
398
+ "loss": 0.5511,
399
+ "step": 650
400
+ },
401
+ {
402
+ "epoch": 0.22,
403
+ "learning_rate": 5e-05,
404
+ "loss": 0.5444,
405
+ "step": 660
406
+ },
407
+ {
408
+ "epoch": 0.23,
409
+ "learning_rate": 5e-05,
410
+ "loss": 0.5522,
411
+ "step": 670
412
+ },
413
+ {
414
+ "epoch": 0.23,
415
+ "learning_rate": 5e-05,
416
+ "loss": 0.5417,
417
+ "step": 680
418
+ },
419
+ {
420
+ "epoch": 0.23,
421
+ "learning_rate": 5e-05,
422
+ "loss": 0.5543,
423
+ "step": 690
424
+ },
425
+ {
426
+ "epoch": 0.24,
427
+ "learning_rate": 5e-05,
428
+ "loss": 0.56,
429
+ "step": 700
430
+ },
431
+ {
432
+ "epoch": 0.24,
433
+ "learning_rate": 5e-05,
434
+ "loss": 0.5394,
435
+ "step": 710
436
+ },
437
+ {
438
+ "epoch": 0.24,
439
+ "learning_rate": 5e-05,
440
+ "loss": 0.5475,
441
+ "step": 720
442
+ },
443
+ {
444
+ "epoch": 0.25,
445
+ "learning_rate": 5e-05,
446
+ "loss": 0.5437,
447
+ "step": 730
448
+ },
449
+ {
450
+ "epoch": 0.25,
451
+ "learning_rate": 5e-05,
452
+ "loss": 0.5438,
453
+ "step": 740
454
+ },
455
+ {
456
+ "epoch": 0.25,
457
+ "learning_rate": 5e-05,
458
+ "loss": 0.5454,
459
+ "step": 750
460
+ },
461
+ {
462
+ "epoch": 0.26,
463
+ "learning_rate": 5e-05,
464
+ "loss": 0.5481,
465
+ "step": 760
466
+ },
467
+ {
468
+ "epoch": 0.26,
469
+ "learning_rate": 5e-05,
470
+ "loss": 0.5434,
471
+ "step": 770
472
+ },
473
+ {
474
+ "epoch": 0.26,
475
+ "learning_rate": 5e-05,
476
+ "loss": 0.5422,
477
+ "step": 780
478
+ },
479
+ {
480
+ "epoch": 0.27,
481
+ "learning_rate": 5e-05,
482
+ "loss": 0.5465,
483
+ "step": 790
484
+ },
485
+ {
486
+ "epoch": 0.27,
487
+ "learning_rate": 5e-05,
488
+ "loss": 0.5447,
489
+ "step": 800
490
+ },
491
+ {
492
+ "epoch": 0.27,
493
+ "learning_rate": 5e-05,
494
+ "loss": 0.5396,
495
+ "step": 810
496
+ },
497
+ {
498
+ "epoch": 0.28,
499
+ "learning_rate": 5e-05,
500
+ "loss": 0.543,
501
+ "step": 820
502
+ },
503
+ {
504
+ "epoch": 0.28,
505
+ "learning_rate": 5e-05,
506
+ "loss": 0.5461,
507
+ "step": 830
508
+ },
509
+ {
510
+ "epoch": 0.28,
511
+ "learning_rate": 5e-05,
512
+ "loss": 0.5482,
513
+ "step": 840
514
+ },
515
+ {
516
+ "epoch": 0.29,
517
+ "learning_rate": 5e-05,
518
+ "loss": 0.5426,
519
+ "step": 850
520
+ },
521
+ {
522
+ "epoch": 0.29,
523
+ "learning_rate": 5e-05,
524
+ "loss": 0.5453,
525
+ "step": 860
526
+ },
527
+ {
528
+ "epoch": 0.29,
529
+ "learning_rate": 5e-05,
530
+ "loss": 0.5493,
531
+ "step": 870
532
+ },
533
+ {
534
+ "epoch": 0.3,
535
+ "learning_rate": 5e-05,
536
+ "loss": 0.5477,
537
+ "step": 880
538
+ },
539
+ {
540
+ "epoch": 0.3,
541
+ "learning_rate": 5e-05,
542
+ "loss": 0.542,
543
+ "step": 890
544
+ },
545
+ {
546
+ "epoch": 0.3,
547
+ "learning_rate": 5e-05,
548
+ "loss": 0.5393,
549
+ "step": 900
550
+ },
551
+ {
552
+ "epoch": 0.31,
553
+ "learning_rate": 5e-05,
554
+ "loss": 0.5365,
555
+ "step": 910
556
+ },
557
+ {
558
+ "epoch": 0.31,
559
+ "learning_rate": 5e-05,
560
+ "loss": 0.5433,
561
+ "step": 920
562
+ },
563
+ {
564
+ "epoch": 0.31,
565
+ "learning_rate": 5e-05,
566
+ "loss": 0.542,
567
+ "step": 930
568
+ },
569
+ {
570
+ "epoch": 0.32,
571
+ "learning_rate": 5e-05,
572
+ "loss": 0.5368,
573
+ "step": 940
574
+ },
575
+ {
576
+ "epoch": 0.32,
577
+ "learning_rate": 5e-05,
578
+ "loss": 0.542,
579
+ "step": 950
580
+ },
581
+ {
582
+ "epoch": 0.32,
583
+ "learning_rate": 5e-05,
584
+ "loss": 0.5467,
585
+ "step": 960
586
+ },
587
+ {
588
+ "epoch": 0.33,
589
+ "learning_rate": 5e-05,
590
+ "loss": 0.5382,
591
+ "step": 970
592
+ },
593
+ {
594
+ "epoch": 0.33,
595
+ "learning_rate": 5e-05,
596
+ "loss": 0.5366,
597
+ "step": 980
598
+ },
599
+ {
600
+ "epoch": 0.33,
601
+ "learning_rate": 5e-05,
602
+ "loss": 0.5354,
603
+ "step": 990
604
+ },
605
+ {
606
+ "epoch": 0.34,
607
+ "learning_rate": 5e-05,
608
+ "loss": 0.5322,
609
+ "step": 1000
610
+ }
611
+ ],
612
+ "logging_steps": 10,
613
+ "max_steps": 8892,
614
+ "num_input_tokens_seen": 0,
615
+ "num_train_epochs": 3,
616
+ "save_steps": 1000,
617
+ "total_flos": 0.0,
618
+ "train_batch_size": 8,
619
+ "trial_name": null,
620
+ "trial_params": null
621
+ }
cvnet_small_v1_global/checkpoint-1000/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:dddfb786197274aad152ee4c2cc9787279a04f3c148146f1b57e85d4fb19f6fc
3
+ size 4856
cvnet_small_v1_global/checkpoint-2000/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:fbdd0ac1006e9b3741dea6c7d1cbaed5ad86865589792b5cd08c72bb887c128e
3
+ size 237860264
cvnet_small_v1_global/checkpoint-2000/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3dcf359b402ab155e141b07b029eddf98e5e63df26e4e808b51a68a0fa44c1bb
3
+ size 475804282
cvnet_small_v1_global/checkpoint-2000/rng_state_0.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:24545b6bb7493bdf35c20f30780f8e0793c67065421d64d7180f5a320abd8417
3
+ size 15920
cvnet_small_v1_global/checkpoint-2000/rng_state_1.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a0232b218d0ca1f5487f91d43e6fb49bb92e0051351056294e6e73f362f38dc4
3
+ size 15920
cvnet_small_v1_global/checkpoint-2000/rng_state_2.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3dffc06c0c6d363c38ffe8a64d78f552b7bfccf1d412a5e6d8b05a20badc3ab4
3
+ size 15920
cvnet_small_v1_global/checkpoint-2000/rng_state_3.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8c47dc8892b7412ec8da71c7704256b96b8469c5c7f3b6490b83502e4b2ee928
3
+ size 15920
cvnet_small_v1_global/checkpoint-2000/rng_state_4.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:905b45076269c4efc0bfdb46513ea882d1de5bce7ddd5bed4ce98092c7da2c0b
3
+ size 15920
cvnet_small_v1_global/checkpoint-2000/rng_state_5.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:97ee8e3283737d5d9165ab7bbed91fab8f8054e727830b0ac207e19ea1b998ca
3
+ size 15920
cvnet_small_v1_global/checkpoint-2000/rng_state_6.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f7205ffddd0d11808dcecbed7a8e9b2051147bffa0bcf02b1ee5ef28b9e65162
3
+ size 15920
cvnet_small_v1_global/checkpoint-2000/rng_state_7.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:07bde508a1939450d8ec45b79bab2568f6dd5d0ac0cf2add96faab85481697da
3
+ size 15920
cvnet_small_v1_global/checkpoint-2000/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:115fc5d3c71fc8b112d5c1701359950aa4e48ea559451a73de1d1d8ca1116d89
3
+ size 1064
cvnet_small_v1_global/checkpoint-2000/trainer_state.json ADDED
@@ -0,0 +1,1221 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": null,
3
+ "best_model_checkpoint": null,
4
+ "epoch": 0.6747069241798094,
5
+ "eval_steps": 500,
6
+ "global_step": 2000,
7
+ "is_hyper_param_search": false,
8
+ "is_local_process_zero": true,
9
+ "is_world_process_zero": true,
10
+ "log_history": [
11
+ {
12
+ "epoch": 0.0,
13
+ "learning_rate": 5e-05,
14
+ "loss": 0.7081,
15
+ "step": 10
16
+ },
17
+ {
18
+ "epoch": 0.01,
19
+ "learning_rate": 5e-05,
20
+ "loss": 0.6774,
21
+ "step": 20
22
+ },
23
+ {
24
+ "epoch": 0.01,
25
+ "learning_rate": 5e-05,
26
+ "loss": 0.6781,
27
+ "step": 30
28
+ },
29
+ {
30
+ "epoch": 0.01,
31
+ "learning_rate": 5e-05,
32
+ "loss": 0.6772,
33
+ "step": 40
34
+ },
35
+ {
36
+ "epoch": 0.02,
37
+ "learning_rate": 5e-05,
38
+ "loss": 0.6802,
39
+ "step": 50
40
+ },
41
+ {
42
+ "epoch": 0.02,
43
+ "learning_rate": 5e-05,
44
+ "loss": 0.6749,
45
+ "step": 60
46
+ },
47
+ {
48
+ "epoch": 0.02,
49
+ "learning_rate": 5e-05,
50
+ "loss": 0.6745,
51
+ "step": 70
52
+ },
53
+ {
54
+ "epoch": 0.03,
55
+ "learning_rate": 5e-05,
56
+ "loss": 0.6729,
57
+ "step": 80
58
+ },
59
+ {
60
+ "epoch": 0.03,
61
+ "learning_rate": 5e-05,
62
+ "loss": 0.6586,
63
+ "step": 90
64
+ },
65
+ {
66
+ "epoch": 0.03,
67
+ "learning_rate": 5e-05,
68
+ "loss": 0.636,
69
+ "step": 100
70
+ },
71
+ {
72
+ "epoch": 0.04,
73
+ "learning_rate": 5e-05,
74
+ "loss": 0.6237,
75
+ "step": 110
76
+ },
77
+ {
78
+ "epoch": 0.04,
79
+ "learning_rate": 5e-05,
80
+ "loss": 0.6098,
81
+ "step": 120
82
+ },
83
+ {
84
+ "epoch": 0.04,
85
+ "learning_rate": 5e-05,
86
+ "loss": 0.5978,
87
+ "step": 130
88
+ },
89
+ {
90
+ "epoch": 0.05,
91
+ "learning_rate": 5e-05,
92
+ "loss": 0.602,
93
+ "step": 140
94
+ },
95
+ {
96
+ "epoch": 0.05,
97
+ "learning_rate": 5e-05,
98
+ "loss": 0.5959,
99
+ "step": 150
100
+ },
101
+ {
102
+ "epoch": 0.05,
103
+ "learning_rate": 5e-05,
104
+ "loss": 0.5906,
105
+ "step": 160
106
+ },
107
+ {
108
+ "epoch": 0.06,
109
+ "learning_rate": 5e-05,
110
+ "loss": 0.5778,
111
+ "step": 170
112
+ },
113
+ {
114
+ "epoch": 0.06,
115
+ "learning_rate": 5e-05,
116
+ "loss": 0.5795,
117
+ "step": 180
118
+ },
119
+ {
120
+ "epoch": 0.06,
121
+ "learning_rate": 5e-05,
122
+ "loss": 0.5883,
123
+ "step": 190
124
+ },
125
+ {
126
+ "epoch": 0.07,
127
+ "learning_rate": 5e-05,
128
+ "loss": 0.604,
129
+ "step": 200
130
+ },
131
+ {
132
+ "epoch": 0.07,
133
+ "learning_rate": 5e-05,
134
+ "loss": 0.5882,
135
+ "step": 210
136
+ },
137
+ {
138
+ "epoch": 0.07,
139
+ "learning_rate": 5e-05,
140
+ "loss": 0.5747,
141
+ "step": 220
142
+ },
143
+ {
144
+ "epoch": 0.08,
145
+ "learning_rate": 5e-05,
146
+ "loss": 0.5745,
147
+ "step": 230
148
+ },
149
+ {
150
+ "epoch": 0.08,
151
+ "learning_rate": 5e-05,
152
+ "loss": 0.5846,
153
+ "step": 240
154
+ },
155
+ {
156
+ "epoch": 0.08,
157
+ "learning_rate": 5e-05,
158
+ "loss": 0.5823,
159
+ "step": 250
160
+ },
161
+ {
162
+ "epoch": 0.09,
163
+ "learning_rate": 5e-05,
164
+ "loss": 0.5722,
165
+ "step": 260
166
+ },
167
+ {
168
+ "epoch": 0.09,
169
+ "learning_rate": 5e-05,
170
+ "loss": 0.5599,
171
+ "step": 270
172
+ },
173
+ {
174
+ "epoch": 0.09,
175
+ "learning_rate": 5e-05,
176
+ "loss": 0.5716,
177
+ "step": 280
178
+ },
179
+ {
180
+ "epoch": 0.1,
181
+ "learning_rate": 5e-05,
182
+ "loss": 0.5832,
183
+ "step": 290
184
+ },
185
+ {
186
+ "epoch": 0.1,
187
+ "learning_rate": 5e-05,
188
+ "loss": 0.5745,
189
+ "step": 300
190
+ },
191
+ {
192
+ "epoch": 0.1,
193
+ "learning_rate": 5e-05,
194
+ "loss": 0.5691,
195
+ "step": 310
196
+ },
197
+ {
198
+ "epoch": 0.11,
199
+ "learning_rate": 5e-05,
200
+ "loss": 0.5711,
201
+ "step": 320
202
+ },
203
+ {
204
+ "epoch": 0.11,
205
+ "learning_rate": 5e-05,
206
+ "loss": 0.5821,
207
+ "step": 330
208
+ },
209
+ {
210
+ "epoch": 0.11,
211
+ "learning_rate": 5e-05,
212
+ "loss": 0.5557,
213
+ "step": 340
214
+ },
215
+ {
216
+ "epoch": 0.12,
217
+ "learning_rate": 5e-05,
218
+ "loss": 0.5581,
219
+ "step": 350
220
+ },
221
+ {
222
+ "epoch": 0.12,
223
+ "learning_rate": 5e-05,
224
+ "loss": 0.5761,
225
+ "step": 360
226
+ },
227
+ {
228
+ "epoch": 0.12,
229
+ "learning_rate": 5e-05,
230
+ "loss": 0.5643,
231
+ "step": 370
232
+ },
233
+ {
234
+ "epoch": 0.13,
235
+ "learning_rate": 5e-05,
236
+ "loss": 0.5745,
237
+ "step": 380
238
+ },
239
+ {
240
+ "epoch": 0.13,
241
+ "learning_rate": 5e-05,
242
+ "loss": 0.5702,
243
+ "step": 390
244
+ },
245
+ {
246
+ "epoch": 0.13,
247
+ "learning_rate": 5e-05,
248
+ "loss": 0.5648,
249
+ "step": 400
250
+ },
251
+ {
252
+ "epoch": 0.14,
253
+ "learning_rate": 5e-05,
254
+ "loss": 0.5691,
255
+ "step": 410
256
+ },
257
+ {
258
+ "epoch": 0.14,
259
+ "learning_rate": 5e-05,
260
+ "loss": 0.556,
261
+ "step": 420
262
+ },
263
+ {
264
+ "epoch": 0.15,
265
+ "learning_rate": 5e-05,
266
+ "loss": 0.5578,
267
+ "step": 430
268
+ },
269
+ {
270
+ "epoch": 0.15,
271
+ "learning_rate": 5e-05,
272
+ "loss": 0.564,
273
+ "step": 440
274
+ },
275
+ {
276
+ "epoch": 0.15,
277
+ "learning_rate": 5e-05,
278
+ "loss": 0.5562,
279
+ "step": 450
280
+ },
281
+ {
282
+ "epoch": 0.16,
283
+ "learning_rate": 5e-05,
284
+ "loss": 0.559,
285
+ "step": 460
286
+ },
287
+ {
288
+ "epoch": 0.16,
289
+ "learning_rate": 5e-05,
290
+ "loss": 0.565,
291
+ "step": 470
292
+ },
293
+ {
294
+ "epoch": 0.16,
295
+ "learning_rate": 5e-05,
296
+ "loss": 0.5508,
297
+ "step": 480
298
+ },
299
+ {
300
+ "epoch": 0.17,
301
+ "learning_rate": 5e-05,
302
+ "loss": 0.5588,
303
+ "step": 490
304
+ },
305
+ {
306
+ "epoch": 0.17,
307
+ "learning_rate": 5e-05,
308
+ "loss": 0.5584,
309
+ "step": 500
310
+ },
311
+ {
312
+ "epoch": 0.17,
313
+ "learning_rate": 5e-05,
314
+ "loss": 0.5555,
315
+ "step": 510
316
+ },
317
+ {
318
+ "epoch": 0.18,
319
+ "learning_rate": 5e-05,
320
+ "loss": 0.5597,
321
+ "step": 520
322
+ },
323
+ {
324
+ "epoch": 0.18,
325
+ "learning_rate": 5e-05,
326
+ "loss": 0.5483,
327
+ "step": 530
328
+ },
329
+ {
330
+ "epoch": 0.18,
331
+ "learning_rate": 5e-05,
332
+ "loss": 0.5621,
333
+ "step": 540
334
+ },
335
+ {
336
+ "epoch": 0.19,
337
+ "learning_rate": 5e-05,
338
+ "loss": 0.5568,
339
+ "step": 550
340
+ },
341
+ {
342
+ "epoch": 0.19,
343
+ "learning_rate": 5e-05,
344
+ "loss": 0.5527,
345
+ "step": 560
346
+ },
347
+ {
348
+ "epoch": 0.19,
349
+ "learning_rate": 5e-05,
350
+ "loss": 0.5524,
351
+ "step": 570
352
+ },
353
+ {
354
+ "epoch": 0.2,
355
+ "learning_rate": 5e-05,
356
+ "loss": 0.5458,
357
+ "step": 580
358
+ },
359
+ {
360
+ "epoch": 0.2,
361
+ "learning_rate": 5e-05,
362
+ "loss": 0.5426,
363
+ "step": 590
364
+ },
365
+ {
366
+ "epoch": 0.2,
367
+ "learning_rate": 5e-05,
368
+ "loss": 0.5524,
369
+ "step": 600
370
+ },
371
+ {
372
+ "epoch": 0.21,
373
+ "learning_rate": 5e-05,
374
+ "loss": 0.5553,
375
+ "step": 610
376
+ },
377
+ {
378
+ "epoch": 0.21,
379
+ "learning_rate": 5e-05,
380
+ "loss": 0.5582,
381
+ "step": 620
382
+ },
383
+ {
384
+ "epoch": 0.21,
385
+ "learning_rate": 5e-05,
386
+ "loss": 0.5518,
387
+ "step": 630
388
+ },
389
+ {
390
+ "epoch": 0.22,
391
+ "learning_rate": 5e-05,
392
+ "loss": 0.5485,
393
+ "step": 640
394
+ },
395
+ {
396
+ "epoch": 0.22,
397
+ "learning_rate": 5e-05,
398
+ "loss": 0.5511,
399
+ "step": 650
400
+ },
401
+ {
402
+ "epoch": 0.22,
403
+ "learning_rate": 5e-05,
404
+ "loss": 0.5444,
405
+ "step": 660
406
+ },
407
+ {
408
+ "epoch": 0.23,
409
+ "learning_rate": 5e-05,
410
+ "loss": 0.5522,
411
+ "step": 670
412
+ },
413
+ {
414
+ "epoch": 0.23,
415
+ "learning_rate": 5e-05,
416
+ "loss": 0.5417,
417
+ "step": 680
418
+ },
419
+ {
420
+ "epoch": 0.23,
421
+ "learning_rate": 5e-05,
422
+ "loss": 0.5543,
423
+ "step": 690
424
+ },
425
+ {
426
+ "epoch": 0.24,
427
+ "learning_rate": 5e-05,
428
+ "loss": 0.56,
429
+ "step": 700
430
+ },
431
+ {
432
+ "epoch": 0.24,
433
+ "learning_rate": 5e-05,
434
+ "loss": 0.5394,
435
+ "step": 710
436
+ },
437
+ {
438
+ "epoch": 0.24,
439
+ "learning_rate": 5e-05,
440
+ "loss": 0.5475,
441
+ "step": 720
442
+ },
443
+ {
444
+ "epoch": 0.25,
445
+ "learning_rate": 5e-05,
446
+ "loss": 0.5437,
447
+ "step": 730
448
+ },
449
+ {
450
+ "epoch": 0.25,
451
+ "learning_rate": 5e-05,
452
+ "loss": 0.5438,
453
+ "step": 740
454
+ },
455
+ {
456
+ "epoch": 0.25,
457
+ "learning_rate": 5e-05,
458
+ "loss": 0.5454,
459
+ "step": 750
460
+ },
461
+ {
462
+ "epoch": 0.26,
463
+ "learning_rate": 5e-05,
464
+ "loss": 0.5481,
465
+ "step": 760
466
+ },
467
+ {
468
+ "epoch": 0.26,
469
+ "learning_rate": 5e-05,
470
+ "loss": 0.5434,
471
+ "step": 770
472
+ },
473
+ {
474
+ "epoch": 0.26,
475
+ "learning_rate": 5e-05,
476
+ "loss": 0.5422,
477
+ "step": 780
478
+ },
479
+ {
480
+ "epoch": 0.27,
481
+ "learning_rate": 5e-05,
482
+ "loss": 0.5465,
483
+ "step": 790
484
+ },
485
+ {
486
+ "epoch": 0.27,
487
+ "learning_rate": 5e-05,
488
+ "loss": 0.5447,
489
+ "step": 800
490
+ },
491
+ {
492
+ "epoch": 0.27,
493
+ "learning_rate": 5e-05,
494
+ "loss": 0.5396,
495
+ "step": 810
496
+ },
497
+ {
498
+ "epoch": 0.28,
499
+ "learning_rate": 5e-05,
500
+ "loss": 0.543,
501
+ "step": 820
502
+ },
503
+ {
504
+ "epoch": 0.28,
505
+ "learning_rate": 5e-05,
506
+ "loss": 0.5461,
507
+ "step": 830
508
+ },
509
+ {
510
+ "epoch": 0.28,
511
+ "learning_rate": 5e-05,
512
+ "loss": 0.5482,
513
+ "step": 840
514
+ },
515
+ {
516
+ "epoch": 0.29,
517
+ "learning_rate": 5e-05,
518
+ "loss": 0.5426,
519
+ "step": 850
520
+ },
521
+ {
522
+ "epoch": 0.29,
523
+ "learning_rate": 5e-05,
524
+ "loss": 0.5453,
525
+ "step": 860
526
+ },
527
+ {
528
+ "epoch": 0.29,
529
+ "learning_rate": 5e-05,
530
+ "loss": 0.5493,
531
+ "step": 870
532
+ },
533
+ {
534
+ "epoch": 0.3,
535
+ "learning_rate": 5e-05,
536
+ "loss": 0.5477,
537
+ "step": 880
538
+ },
539
+ {
540
+ "epoch": 0.3,
541
+ "learning_rate": 5e-05,
542
+ "loss": 0.542,
543
+ "step": 890
544
+ },
545
+ {
546
+ "epoch": 0.3,
547
+ "learning_rate": 5e-05,
548
+ "loss": 0.5393,
549
+ "step": 900
550
+ },
551
+ {
552
+ "epoch": 0.31,
553
+ "learning_rate": 5e-05,
554
+ "loss": 0.5365,
555
+ "step": 910
556
+ },
557
+ {
558
+ "epoch": 0.31,
559
+ "learning_rate": 5e-05,
560
+ "loss": 0.5433,
561
+ "step": 920
562
+ },
563
+ {
564
+ "epoch": 0.31,
565
+ "learning_rate": 5e-05,
566
+ "loss": 0.542,
567
+ "step": 930
568
+ },
569
+ {
570
+ "epoch": 0.32,
571
+ "learning_rate": 5e-05,
572
+ "loss": 0.5368,
573
+ "step": 940
574
+ },
575
+ {
576
+ "epoch": 0.32,
577
+ "learning_rate": 5e-05,
578
+ "loss": 0.542,
579
+ "step": 950
580
+ },
581
+ {
582
+ "epoch": 0.32,
583
+ "learning_rate": 5e-05,
584
+ "loss": 0.5467,
585
+ "step": 960
586
+ },
587
+ {
588
+ "epoch": 0.33,
589
+ "learning_rate": 5e-05,
590
+ "loss": 0.5382,
591
+ "step": 970
592
+ },
593
+ {
594
+ "epoch": 0.33,
595
+ "learning_rate": 5e-05,
596
+ "loss": 0.5366,
597
+ "step": 980
598
+ },
599
+ {
600
+ "epoch": 0.33,
601
+ "learning_rate": 5e-05,
602
+ "loss": 0.5354,
603
+ "step": 990
604
+ },
605
+ {
606
+ "epoch": 0.34,
607
+ "learning_rate": 5e-05,
608
+ "loss": 0.5322,
609
+ "step": 1000
610
+ },
611
+ {
612
+ "epoch": 0.34,
613
+ "learning_rate": 5e-05,
614
+ "loss": 0.5379,
615
+ "step": 1010
616
+ },
617
+ {
618
+ "epoch": 0.34,
619
+ "learning_rate": 5e-05,
620
+ "loss": 0.5398,
621
+ "step": 1020
622
+ },
623
+ {
624
+ "epoch": 0.35,
625
+ "learning_rate": 5e-05,
626
+ "loss": 0.5337,
627
+ "step": 1030
628
+ },
629
+ {
630
+ "epoch": 0.35,
631
+ "learning_rate": 5e-05,
632
+ "loss": 0.5384,
633
+ "step": 1040
634
+ },
635
+ {
636
+ "epoch": 0.35,
637
+ "learning_rate": 5e-05,
638
+ "loss": 0.5256,
639
+ "step": 1050
640
+ },
641
+ {
642
+ "epoch": 0.36,
643
+ "learning_rate": 5e-05,
644
+ "loss": 0.5353,
645
+ "step": 1060
646
+ },
647
+ {
648
+ "epoch": 0.36,
649
+ "learning_rate": 5e-05,
650
+ "loss": 0.5233,
651
+ "step": 1070
652
+ },
653
+ {
654
+ "epoch": 0.36,
655
+ "learning_rate": 5e-05,
656
+ "loss": 0.5483,
657
+ "step": 1080
658
+ },
659
+ {
660
+ "epoch": 0.37,
661
+ "learning_rate": 5e-05,
662
+ "loss": 0.5362,
663
+ "step": 1090
664
+ },
665
+ {
666
+ "epoch": 0.37,
667
+ "learning_rate": 5e-05,
668
+ "loss": 0.5426,
669
+ "step": 1100
670
+ },
671
+ {
672
+ "epoch": 0.37,
673
+ "learning_rate": 5e-05,
674
+ "loss": 0.5347,
675
+ "step": 1110
676
+ },
677
+ {
678
+ "epoch": 0.38,
679
+ "learning_rate": 5e-05,
680
+ "loss": 0.5377,
681
+ "step": 1120
682
+ },
683
+ {
684
+ "epoch": 0.38,
685
+ "learning_rate": 5e-05,
686
+ "loss": 0.5374,
687
+ "step": 1130
688
+ },
689
+ {
690
+ "epoch": 0.38,
691
+ "learning_rate": 5e-05,
692
+ "loss": 0.532,
693
+ "step": 1140
694
+ },
695
+ {
696
+ "epoch": 0.39,
697
+ "learning_rate": 5e-05,
698
+ "loss": 0.5315,
699
+ "step": 1150
700
+ },
701
+ {
702
+ "epoch": 0.39,
703
+ "learning_rate": 5e-05,
704
+ "loss": 0.5274,
705
+ "step": 1160
706
+ },
707
+ {
708
+ "epoch": 0.39,
709
+ "learning_rate": 5e-05,
710
+ "loss": 0.537,
711
+ "step": 1170
712
+ },
713
+ {
714
+ "epoch": 0.4,
715
+ "learning_rate": 5e-05,
716
+ "loss": 0.5427,
717
+ "step": 1180
718
+ },
719
+ {
720
+ "epoch": 0.4,
721
+ "learning_rate": 5e-05,
722
+ "loss": 0.5288,
723
+ "step": 1190
724
+ },
725
+ {
726
+ "epoch": 0.4,
727
+ "learning_rate": 5e-05,
728
+ "loss": 0.5259,
729
+ "step": 1200
730
+ },
731
+ {
732
+ "epoch": 0.41,
733
+ "learning_rate": 5e-05,
734
+ "loss": 0.531,
735
+ "step": 1210
736
+ },
737
+ {
738
+ "epoch": 0.41,
739
+ "learning_rate": 5e-05,
740
+ "loss": 0.5263,
741
+ "step": 1220
742
+ },
743
+ {
744
+ "epoch": 0.41,
745
+ "learning_rate": 5e-05,
746
+ "loss": 0.5305,
747
+ "step": 1230
748
+ },
749
+ {
750
+ "epoch": 0.42,
751
+ "learning_rate": 5e-05,
752
+ "loss": 0.5245,
753
+ "step": 1240
754
+ },
755
+ {
756
+ "epoch": 0.42,
757
+ "learning_rate": 5e-05,
758
+ "loss": 0.5306,
759
+ "step": 1250
760
+ },
761
+ {
762
+ "epoch": 0.43,
763
+ "learning_rate": 5e-05,
764
+ "loss": 0.526,
765
+ "step": 1260
766
+ },
767
+ {
768
+ "epoch": 0.43,
769
+ "learning_rate": 5e-05,
770
+ "loss": 0.5308,
771
+ "step": 1270
772
+ },
773
+ {
774
+ "epoch": 0.43,
775
+ "learning_rate": 5e-05,
776
+ "loss": 0.5226,
777
+ "step": 1280
778
+ },
779
+ {
780
+ "epoch": 0.44,
781
+ "learning_rate": 5e-05,
782
+ "loss": 0.5295,
783
+ "step": 1290
784
+ },
785
+ {
786
+ "epoch": 0.44,
787
+ "learning_rate": 5e-05,
788
+ "loss": 0.5297,
789
+ "step": 1300
790
+ },
791
+ {
792
+ "epoch": 0.44,
793
+ "learning_rate": 5e-05,
794
+ "loss": 0.5293,
795
+ "step": 1310
796
+ },
797
+ {
798
+ "epoch": 0.45,
799
+ "learning_rate": 5e-05,
800
+ "loss": 0.5387,
801
+ "step": 1320
802
+ },
803
+ {
804
+ "epoch": 0.45,
805
+ "learning_rate": 5e-05,
806
+ "loss": 0.5252,
807
+ "step": 1330
808
+ },
809
+ {
810
+ "epoch": 0.45,
811
+ "learning_rate": 5e-05,
812
+ "loss": 0.524,
813
+ "step": 1340
814
+ },
815
+ {
816
+ "epoch": 0.46,
817
+ "learning_rate": 5e-05,
818
+ "loss": 0.5231,
819
+ "step": 1350
820
+ },
821
+ {
822
+ "epoch": 0.46,
823
+ "learning_rate": 5e-05,
824
+ "loss": 0.5248,
825
+ "step": 1360
826
+ },
827
+ {
828
+ "epoch": 0.46,
829
+ "learning_rate": 5e-05,
830
+ "loss": 0.5332,
831
+ "step": 1370
832
+ },
833
+ {
834
+ "epoch": 0.47,
835
+ "learning_rate": 5e-05,
836
+ "loss": 0.5366,
837
+ "step": 1380
838
+ },
839
+ {
840
+ "epoch": 0.47,
841
+ "learning_rate": 5e-05,
842
+ "loss": 0.5231,
843
+ "step": 1390
844
+ },
845
+ {
846
+ "epoch": 0.47,
847
+ "learning_rate": 5e-05,
848
+ "loss": 0.5206,
849
+ "step": 1400
850
+ },
851
+ {
852
+ "epoch": 0.48,
853
+ "learning_rate": 5e-05,
854
+ "loss": 0.5237,
855
+ "step": 1410
856
+ },
857
+ {
858
+ "epoch": 0.48,
859
+ "learning_rate": 5e-05,
860
+ "loss": 0.5201,
861
+ "step": 1420
862
+ },
863
+ {
864
+ "epoch": 0.48,
865
+ "learning_rate": 5e-05,
866
+ "loss": 0.5366,
867
+ "step": 1430
868
+ },
869
+ {
870
+ "epoch": 0.49,
871
+ "learning_rate": 5e-05,
872
+ "loss": 0.5336,
873
+ "step": 1440
874
+ },
875
+ {
876
+ "epoch": 0.49,
877
+ "learning_rate": 5e-05,
878
+ "loss": 0.5282,
879
+ "step": 1450
880
+ },
881
+ {
882
+ "epoch": 0.49,
883
+ "learning_rate": 5e-05,
884
+ "loss": 0.5269,
885
+ "step": 1460
886
+ },
887
+ {
888
+ "epoch": 0.5,
889
+ "learning_rate": 5e-05,
890
+ "loss": 0.5245,
891
+ "step": 1470
892
+ },
893
+ {
894
+ "epoch": 0.5,
895
+ "learning_rate": 5e-05,
896
+ "loss": 0.5203,
897
+ "step": 1480
898
+ },
899
+ {
900
+ "epoch": 0.5,
901
+ "learning_rate": 5e-05,
902
+ "loss": 0.5225,
903
+ "step": 1490
904
+ },
905
+ {
906
+ "epoch": 0.51,
907
+ "learning_rate": 5e-05,
908
+ "loss": 0.5311,
909
+ "step": 1500
910
+ },
911
+ {
912
+ "epoch": 0.51,
913
+ "learning_rate": 5e-05,
914
+ "loss": 0.5229,
915
+ "step": 1510
916
+ },
917
+ {
918
+ "epoch": 0.51,
919
+ "learning_rate": 5e-05,
920
+ "loss": 0.5266,
921
+ "step": 1520
922
+ },
923
+ {
924
+ "epoch": 0.52,
925
+ "learning_rate": 5e-05,
926
+ "loss": 0.5283,
927
+ "step": 1530
928
+ },
929
+ {
930
+ "epoch": 0.52,
931
+ "learning_rate": 5e-05,
932
+ "loss": 0.5279,
933
+ "step": 1540
934
+ },
935
+ {
936
+ "epoch": 0.52,
937
+ "learning_rate": 5e-05,
938
+ "loss": 0.5299,
939
+ "step": 1550
940
+ },
941
+ {
942
+ "epoch": 0.53,
943
+ "learning_rate": 5e-05,
944
+ "loss": 0.5267,
945
+ "step": 1560
946
+ },
947
+ {
948
+ "epoch": 0.53,
949
+ "learning_rate": 5e-05,
950
+ "loss": 0.5186,
951
+ "step": 1570
952
+ },
953
+ {
954
+ "epoch": 0.53,
955
+ "learning_rate": 5e-05,
956
+ "loss": 0.5238,
957
+ "step": 1580
958
+ },
959
+ {
960
+ "epoch": 0.54,
961
+ "learning_rate": 5e-05,
962
+ "loss": 0.5249,
963
+ "step": 1590
964
+ },
965
+ {
966
+ "epoch": 0.54,
967
+ "learning_rate": 5e-05,
968
+ "loss": 0.5225,
969
+ "step": 1600
970
+ },
971
+ {
972
+ "epoch": 0.54,
973
+ "learning_rate": 5e-05,
974
+ "loss": 0.5226,
975
+ "step": 1610
976
+ },
977
+ {
978
+ "epoch": 0.55,
979
+ "learning_rate": 5e-05,
980
+ "loss": 0.5188,
981
+ "step": 1620
982
+ },
983
+ {
984
+ "epoch": 0.55,
985
+ "learning_rate": 5e-05,
986
+ "loss": 0.525,
987
+ "step": 1630
988
+ },
989
+ {
990
+ "epoch": 0.55,
991
+ "learning_rate": 5e-05,
992
+ "loss": 0.5249,
993
+ "step": 1640
994
+ },
995
+ {
996
+ "epoch": 0.56,
997
+ "learning_rate": 5e-05,
998
+ "loss": 0.5225,
999
+ "step": 1650
1000
+ },
1001
+ {
1002
+ "epoch": 0.56,
1003
+ "learning_rate": 5e-05,
1004
+ "loss": 0.5281,
1005
+ "step": 1660
1006
+ },
1007
+ {
1008
+ "epoch": 0.56,
1009
+ "learning_rate": 5e-05,
1010
+ "loss": 0.5204,
1011
+ "step": 1670
1012
+ },
1013
+ {
1014
+ "epoch": 0.57,
1015
+ "learning_rate": 5e-05,
1016
+ "loss": 0.5193,
1017
+ "step": 1680
1018
+ },
1019
+ {
1020
+ "epoch": 0.57,
1021
+ "learning_rate": 5e-05,
1022
+ "loss": 0.5278,
1023
+ "step": 1690
1024
+ },
1025
+ {
1026
+ "epoch": 0.57,
1027
+ "learning_rate": 5e-05,
1028
+ "loss": 0.5224,
1029
+ "step": 1700
1030
+ },
1031
+ {
1032
+ "epoch": 0.58,
1033
+ "learning_rate": 5e-05,
1034
+ "loss": 0.5218,
1035
+ "step": 1710
1036
+ },
1037
+ {
1038
+ "epoch": 0.58,
1039
+ "learning_rate": 5e-05,
1040
+ "loss": 0.5171,
1041
+ "step": 1720
1042
+ },
1043
+ {
1044
+ "epoch": 0.58,
1045
+ "learning_rate": 5e-05,
1046
+ "loss": 0.5239,
1047
+ "step": 1730
1048
+ },
1049
+ {
1050
+ "epoch": 0.59,
1051
+ "learning_rate": 5e-05,
1052
+ "loss": 0.5231,
1053
+ "step": 1740
1054
+ },
1055
+ {
1056
+ "epoch": 0.59,
1057
+ "learning_rate": 5e-05,
1058
+ "loss": 0.5164,
1059
+ "step": 1750
1060
+ },
1061
+ {
1062
+ "epoch": 0.59,
1063
+ "learning_rate": 5e-05,
1064
+ "loss": 0.5181,
1065
+ "step": 1760
1066
+ },
1067
+ {
1068
+ "epoch": 0.6,
1069
+ "learning_rate": 5e-05,
1070
+ "loss": 0.5166,
1071
+ "step": 1770
1072
+ },
1073
+ {
1074
+ "epoch": 0.6,
1075
+ "learning_rate": 5e-05,
1076
+ "loss": 0.5179,
1077
+ "step": 1780
1078
+ },
1079
+ {
1080
+ "epoch": 0.6,
1081
+ "learning_rate": 5e-05,
1082
+ "loss": 0.5169,
1083
+ "step": 1790
1084
+ },
1085
+ {
1086
+ "epoch": 0.61,
1087
+ "learning_rate": 5e-05,
1088
+ "loss": 0.5275,
1089
+ "step": 1800
1090
+ },
1091
+ {
1092
+ "epoch": 0.61,
1093
+ "learning_rate": 5e-05,
1094
+ "loss": 0.5179,
1095
+ "step": 1810
1096
+ },
1097
+ {
1098
+ "epoch": 0.61,
1099
+ "learning_rate": 5e-05,
1100
+ "loss": 0.5216,
1101
+ "step": 1820
1102
+ },
1103
+ {
1104
+ "epoch": 0.62,
1105
+ "learning_rate": 5e-05,
1106
+ "loss": 0.5236,
1107
+ "step": 1830
1108
+ },
1109
+ {
1110
+ "epoch": 0.62,
1111
+ "learning_rate": 5e-05,
1112
+ "loss": 0.5125,
1113
+ "step": 1840
1114
+ },
1115
+ {
1116
+ "epoch": 0.62,
1117
+ "learning_rate": 5e-05,
1118
+ "loss": 0.527,
1119
+ "step": 1850
1120
+ },
1121
+ {
1122
+ "epoch": 0.63,
1123
+ "learning_rate": 5e-05,
1124
+ "loss": 0.5224,
1125
+ "step": 1860
1126
+ },
1127
+ {
1128
+ "epoch": 0.63,
1129
+ "learning_rate": 5e-05,
1130
+ "loss": 0.5304,
1131
+ "step": 1870
1132
+ },
1133
+ {
1134
+ "epoch": 0.63,
1135
+ "learning_rate": 5e-05,
1136
+ "loss": 0.5242,
1137
+ "step": 1880
1138
+ },
1139
+ {
1140
+ "epoch": 0.64,
1141
+ "learning_rate": 5e-05,
1142
+ "loss": 0.5351,
1143
+ "step": 1890
1144
+ },
1145
+ {
1146
+ "epoch": 0.64,
1147
+ "learning_rate": 5e-05,
1148
+ "loss": 0.5262,
1149
+ "step": 1900
1150
+ },
1151
+ {
1152
+ "epoch": 0.64,
1153
+ "learning_rate": 5e-05,
1154
+ "loss": 0.5178,
1155
+ "step": 1910
1156
+ },
1157
+ {
1158
+ "epoch": 0.65,
1159
+ "learning_rate": 5e-05,
1160
+ "loss": 0.5104,
1161
+ "step": 1920
1162
+ },
1163
+ {
1164
+ "epoch": 0.65,
1165
+ "learning_rate": 5e-05,
1166
+ "loss": 0.5172,
1167
+ "step": 1930
1168
+ },
1169
+ {
1170
+ "epoch": 0.65,
1171
+ "learning_rate": 5e-05,
1172
+ "loss": 0.517,
1173
+ "step": 1940
1174
+ },
1175
+ {
1176
+ "epoch": 0.66,
1177
+ "learning_rate": 5e-05,
1178
+ "loss": 0.5154,
1179
+ "step": 1950
1180
+ },
1181
+ {
1182
+ "epoch": 0.66,
1183
+ "learning_rate": 5e-05,
1184
+ "loss": 0.5285,
1185
+ "step": 1960
1186
+ },
1187
+ {
1188
+ "epoch": 0.66,
1189
+ "learning_rate": 5e-05,
1190
+ "loss": 0.5163,
1191
+ "step": 1970
1192
+ },
1193
+ {
1194
+ "epoch": 0.67,
1195
+ "learning_rate": 5e-05,
1196
+ "loss": 0.5104,
1197
+ "step": 1980
1198
+ },
1199
+ {
1200
+ "epoch": 0.67,
1201
+ "learning_rate": 5e-05,
1202
+ "loss": 0.5228,
1203
+ "step": 1990
1204
+ },
1205
+ {
1206
+ "epoch": 0.67,
1207
+ "learning_rate": 5e-05,
1208
+ "loss": 0.5131,
1209
+ "step": 2000
1210
+ }
1211
+ ],
1212
+ "logging_steps": 10,
1213
+ "max_steps": 8892,
1214
+ "num_input_tokens_seen": 0,
1215
+ "num_train_epochs": 3,
1216
+ "save_steps": 1000,
1217
+ "total_flos": 0.0,
1218
+ "train_batch_size": 8,
1219
+ "trial_name": null,
1220
+ "trial_params": null
1221
+ }
cvnet_small_v1_global/checkpoint-2000/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:dddfb786197274aad152ee4c2cc9787279a04f3c148146f1b57e85d4fb19f6fc
3
+ size 4856
cvnet_small_v1_global/checkpoint-3000/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f6f885f65a815e27442367ce964bc2c0559b43681759c08c7f00bb28cb1d07e1
3
+ size 237860264
cvnet_small_v1_global/checkpoint-3000/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:97471d9b30876e8c0dbc7c483bcba3c96168b58bf31900a8605c5dc276a793de
3
+ size 475804282
cvnet_small_v1_global/checkpoint-3000/rng_state_0.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7431609fc9d10c784464e14fe45f5a89b22573e0769c9c21d7d0a2dead49ec58
3
+ size 15920
cvnet_small_v1_global/checkpoint-3000/rng_state_1.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5081bec13a1a7ac2ad3917351c18540a7bae0cfef6e0df05d537d5c0163fce12
3
+ size 15920
cvnet_small_v1_global/checkpoint-3000/rng_state_2.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:535456fe6d77127c3883ae9c325dccec963de01449e321e9c704f78c50992939
3
+ size 15920
cvnet_small_v1_global/checkpoint-3000/rng_state_3.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:880e7012bf446334aa49027d1791b52f192a6dc11d44fea74095b153fd34bc97
3
+ size 15920
cvnet_small_v1_global/checkpoint-3000/rng_state_4.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:40038cca0e1cc4c41ee9eb36cfd70f6e868697faa2d2d79afba994d00387b725
3
+ size 15920
cvnet_small_v1_global/checkpoint-3000/rng_state_5.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:86ec016c5b320b9210f7edc5e4b4d309aa8c19437ccbdf2fd68012814e7b0be6
3
+ size 15920
cvnet_small_v1_global/checkpoint-3000/rng_state_6.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:049da174472b1bfd728972b360f9b2b80d3b4842c4c00b3019ff285417d7b3ca
3
+ size 15920
cvnet_small_v1_global/checkpoint-3000/rng_state_7.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b2dcd919cf3c5f19cfa3f8a17ee1d2a7803a15af43f3917c55ec05ed6ca5858d
3
+ size 15920
cvnet_small_v1_global/checkpoint-3000/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4df43d46412429a4bda78060914e68357cbee658ac819784da79443fe1ce6982
3
+ size 1064
cvnet_small_v1_global/checkpoint-3000/trainer_state.json ADDED
@@ -0,0 +1,1821 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": null,
3
+ "best_model_checkpoint": null,
4
+ "epoch": 1.0120603862697142,
5
+ "eval_steps": 500,
6
+ "global_step": 3000,
7
+ "is_hyper_param_search": false,
8
+ "is_local_process_zero": true,
9
+ "is_world_process_zero": true,
10
+ "log_history": [
11
+ {
12
+ "epoch": 0.0,
13
+ "learning_rate": 5e-05,
14
+ "loss": 0.7081,
15
+ "step": 10
16
+ },
17
+ {
18
+ "epoch": 0.01,
19
+ "learning_rate": 5e-05,
20
+ "loss": 0.6774,
21
+ "step": 20
22
+ },
23
+ {
24
+ "epoch": 0.01,
25
+ "learning_rate": 5e-05,
26
+ "loss": 0.6781,
27
+ "step": 30
28
+ },
29
+ {
30
+ "epoch": 0.01,
31
+ "learning_rate": 5e-05,
32
+ "loss": 0.6772,
33
+ "step": 40
34
+ },
35
+ {
36
+ "epoch": 0.02,
37
+ "learning_rate": 5e-05,
38
+ "loss": 0.6802,
39
+ "step": 50
40
+ },
41
+ {
42
+ "epoch": 0.02,
43
+ "learning_rate": 5e-05,
44
+ "loss": 0.6749,
45
+ "step": 60
46
+ },
47
+ {
48
+ "epoch": 0.02,
49
+ "learning_rate": 5e-05,
50
+ "loss": 0.6745,
51
+ "step": 70
52
+ },
53
+ {
54
+ "epoch": 0.03,
55
+ "learning_rate": 5e-05,
56
+ "loss": 0.6729,
57
+ "step": 80
58
+ },
59
+ {
60
+ "epoch": 0.03,
61
+ "learning_rate": 5e-05,
62
+ "loss": 0.6586,
63
+ "step": 90
64
+ },
65
+ {
66
+ "epoch": 0.03,
67
+ "learning_rate": 5e-05,
68
+ "loss": 0.636,
69
+ "step": 100
70
+ },
71
+ {
72
+ "epoch": 0.04,
73
+ "learning_rate": 5e-05,
74
+ "loss": 0.6237,
75
+ "step": 110
76
+ },
77
+ {
78
+ "epoch": 0.04,
79
+ "learning_rate": 5e-05,
80
+ "loss": 0.6098,
81
+ "step": 120
82
+ },
83
+ {
84
+ "epoch": 0.04,
85
+ "learning_rate": 5e-05,
86
+ "loss": 0.5978,
87
+ "step": 130
88
+ },
89
+ {
90
+ "epoch": 0.05,
91
+ "learning_rate": 5e-05,
92
+ "loss": 0.602,
93
+ "step": 140
94
+ },
95
+ {
96
+ "epoch": 0.05,
97
+ "learning_rate": 5e-05,
98
+ "loss": 0.5959,
99
+ "step": 150
100
+ },
101
+ {
102
+ "epoch": 0.05,
103
+ "learning_rate": 5e-05,
104
+ "loss": 0.5906,
105
+ "step": 160
106
+ },
107
+ {
108
+ "epoch": 0.06,
109
+ "learning_rate": 5e-05,
110
+ "loss": 0.5778,
111
+ "step": 170
112
+ },
113
+ {
114
+ "epoch": 0.06,
115
+ "learning_rate": 5e-05,
116
+ "loss": 0.5795,
117
+ "step": 180
118
+ },
119
+ {
120
+ "epoch": 0.06,
121
+ "learning_rate": 5e-05,
122
+ "loss": 0.5883,
123
+ "step": 190
124
+ },
125
+ {
126
+ "epoch": 0.07,
127
+ "learning_rate": 5e-05,
128
+ "loss": 0.604,
129
+ "step": 200
130
+ },
131
+ {
132
+ "epoch": 0.07,
133
+ "learning_rate": 5e-05,
134
+ "loss": 0.5882,
135
+ "step": 210
136
+ },
137
+ {
138
+ "epoch": 0.07,
139
+ "learning_rate": 5e-05,
140
+ "loss": 0.5747,
141
+ "step": 220
142
+ },
143
+ {
144
+ "epoch": 0.08,
145
+ "learning_rate": 5e-05,
146
+ "loss": 0.5745,
147
+ "step": 230
148
+ },
149
+ {
150
+ "epoch": 0.08,
151
+ "learning_rate": 5e-05,
152
+ "loss": 0.5846,
153
+ "step": 240
154
+ },
155
+ {
156
+ "epoch": 0.08,
157
+ "learning_rate": 5e-05,
158
+ "loss": 0.5823,
159
+ "step": 250
160
+ },
161
+ {
162
+ "epoch": 0.09,
163
+ "learning_rate": 5e-05,
164
+ "loss": 0.5722,
165
+ "step": 260
166
+ },
167
+ {
168
+ "epoch": 0.09,
169
+ "learning_rate": 5e-05,
170
+ "loss": 0.5599,
171
+ "step": 270
172
+ },
173
+ {
174
+ "epoch": 0.09,
175
+ "learning_rate": 5e-05,
176
+ "loss": 0.5716,
177
+ "step": 280
178
+ },
179
+ {
180
+ "epoch": 0.1,
181
+ "learning_rate": 5e-05,
182
+ "loss": 0.5832,
183
+ "step": 290
184
+ },
185
+ {
186
+ "epoch": 0.1,
187
+ "learning_rate": 5e-05,
188
+ "loss": 0.5745,
189
+ "step": 300
190
+ },
191
+ {
192
+ "epoch": 0.1,
193
+ "learning_rate": 5e-05,
194
+ "loss": 0.5691,
195
+ "step": 310
196
+ },
197
+ {
198
+ "epoch": 0.11,
199
+ "learning_rate": 5e-05,
200
+ "loss": 0.5711,
201
+ "step": 320
202
+ },
203
+ {
204
+ "epoch": 0.11,
205
+ "learning_rate": 5e-05,
206
+ "loss": 0.5821,
207
+ "step": 330
208
+ },
209
+ {
210
+ "epoch": 0.11,
211
+ "learning_rate": 5e-05,
212
+ "loss": 0.5557,
213
+ "step": 340
214
+ },
215
+ {
216
+ "epoch": 0.12,
217
+ "learning_rate": 5e-05,
218
+ "loss": 0.5581,
219
+ "step": 350
220
+ },
221
+ {
222
+ "epoch": 0.12,
223
+ "learning_rate": 5e-05,
224
+ "loss": 0.5761,
225
+ "step": 360
226
+ },
227
+ {
228
+ "epoch": 0.12,
229
+ "learning_rate": 5e-05,
230
+ "loss": 0.5643,
231
+ "step": 370
232
+ },
233
+ {
234
+ "epoch": 0.13,
235
+ "learning_rate": 5e-05,
236
+ "loss": 0.5745,
237
+ "step": 380
238
+ },
239
+ {
240
+ "epoch": 0.13,
241
+ "learning_rate": 5e-05,
242
+ "loss": 0.5702,
243
+ "step": 390
244
+ },
245
+ {
246
+ "epoch": 0.13,
247
+ "learning_rate": 5e-05,
248
+ "loss": 0.5648,
249
+ "step": 400
250
+ },
251
+ {
252
+ "epoch": 0.14,
253
+ "learning_rate": 5e-05,
254
+ "loss": 0.5691,
255
+ "step": 410
256
+ },
257
+ {
258
+ "epoch": 0.14,
259
+ "learning_rate": 5e-05,
260
+ "loss": 0.556,
261
+ "step": 420
262
+ },
263
+ {
264
+ "epoch": 0.15,
265
+ "learning_rate": 5e-05,
266
+ "loss": 0.5578,
267
+ "step": 430
268
+ },
269
+ {
270
+ "epoch": 0.15,
271
+ "learning_rate": 5e-05,
272
+ "loss": 0.564,
273
+ "step": 440
274
+ },
275
+ {
276
+ "epoch": 0.15,
277
+ "learning_rate": 5e-05,
278
+ "loss": 0.5562,
279
+ "step": 450
280
+ },
281
+ {
282
+ "epoch": 0.16,
283
+ "learning_rate": 5e-05,
284
+ "loss": 0.559,
285
+ "step": 460
286
+ },
287
+ {
288
+ "epoch": 0.16,
289
+ "learning_rate": 5e-05,
290
+ "loss": 0.565,
291
+ "step": 470
292
+ },
293
+ {
294
+ "epoch": 0.16,
295
+ "learning_rate": 5e-05,
296
+ "loss": 0.5508,
297
+ "step": 480
298
+ },
299
+ {
300
+ "epoch": 0.17,
301
+ "learning_rate": 5e-05,
302
+ "loss": 0.5588,
303
+ "step": 490
304
+ },
305
+ {
306
+ "epoch": 0.17,
307
+ "learning_rate": 5e-05,
308
+ "loss": 0.5584,
309
+ "step": 500
310
+ },
311
+ {
312
+ "epoch": 0.17,
313
+ "learning_rate": 5e-05,
314
+ "loss": 0.5555,
315
+ "step": 510
316
+ },
317
+ {
318
+ "epoch": 0.18,
319
+ "learning_rate": 5e-05,
320
+ "loss": 0.5597,
321
+ "step": 520
322
+ },
323
+ {
324
+ "epoch": 0.18,
325
+ "learning_rate": 5e-05,
326
+ "loss": 0.5483,
327
+ "step": 530
328
+ },
329
+ {
330
+ "epoch": 0.18,
331
+ "learning_rate": 5e-05,
332
+ "loss": 0.5621,
333
+ "step": 540
334
+ },
335
+ {
336
+ "epoch": 0.19,
337
+ "learning_rate": 5e-05,
338
+ "loss": 0.5568,
339
+ "step": 550
340
+ },
341
+ {
342
+ "epoch": 0.19,
343
+ "learning_rate": 5e-05,
344
+ "loss": 0.5527,
345
+ "step": 560
346
+ },
347
+ {
348
+ "epoch": 0.19,
349
+ "learning_rate": 5e-05,
350
+ "loss": 0.5524,
351
+ "step": 570
352
+ },
353
+ {
354
+ "epoch": 0.2,
355
+ "learning_rate": 5e-05,
356
+ "loss": 0.5458,
357
+ "step": 580
358
+ },
359
+ {
360
+ "epoch": 0.2,
361
+ "learning_rate": 5e-05,
362
+ "loss": 0.5426,
363
+ "step": 590
364
+ },
365
+ {
366
+ "epoch": 0.2,
367
+ "learning_rate": 5e-05,
368
+ "loss": 0.5524,
369
+ "step": 600
370
+ },
371
+ {
372
+ "epoch": 0.21,
373
+ "learning_rate": 5e-05,
374
+ "loss": 0.5553,
375
+ "step": 610
376
+ },
377
+ {
378
+ "epoch": 0.21,
379
+ "learning_rate": 5e-05,
380
+ "loss": 0.5582,
381
+ "step": 620
382
+ },
383
+ {
384
+ "epoch": 0.21,
385
+ "learning_rate": 5e-05,
386
+ "loss": 0.5518,
387
+ "step": 630
388
+ },
389
+ {
390
+ "epoch": 0.22,
391
+ "learning_rate": 5e-05,
392
+ "loss": 0.5485,
393
+ "step": 640
394
+ },
395
+ {
396
+ "epoch": 0.22,
397
+ "learning_rate": 5e-05,
398
+ "loss": 0.5511,
399
+ "step": 650
400
+ },
401
+ {
402
+ "epoch": 0.22,
403
+ "learning_rate": 5e-05,
404
+ "loss": 0.5444,
405
+ "step": 660
406
+ },
407
+ {
408
+ "epoch": 0.23,
409
+ "learning_rate": 5e-05,
410
+ "loss": 0.5522,
411
+ "step": 670
412
+ },
413
+ {
414
+ "epoch": 0.23,
415
+ "learning_rate": 5e-05,
416
+ "loss": 0.5417,
417
+ "step": 680
418
+ },
419
+ {
420
+ "epoch": 0.23,
421
+ "learning_rate": 5e-05,
422
+ "loss": 0.5543,
423
+ "step": 690
424
+ },
425
+ {
426
+ "epoch": 0.24,
427
+ "learning_rate": 5e-05,
428
+ "loss": 0.56,
429
+ "step": 700
430
+ },
431
+ {
432
+ "epoch": 0.24,
433
+ "learning_rate": 5e-05,
434
+ "loss": 0.5394,
435
+ "step": 710
436
+ },
437
+ {
438
+ "epoch": 0.24,
439
+ "learning_rate": 5e-05,
440
+ "loss": 0.5475,
441
+ "step": 720
442
+ },
443
+ {
444
+ "epoch": 0.25,
445
+ "learning_rate": 5e-05,
446
+ "loss": 0.5437,
447
+ "step": 730
448
+ },
449
+ {
450
+ "epoch": 0.25,
451
+ "learning_rate": 5e-05,
452
+ "loss": 0.5438,
453
+ "step": 740
454
+ },
455
+ {
456
+ "epoch": 0.25,
457
+ "learning_rate": 5e-05,
458
+ "loss": 0.5454,
459
+ "step": 750
460
+ },
461
+ {
462
+ "epoch": 0.26,
463
+ "learning_rate": 5e-05,
464
+ "loss": 0.5481,
465
+ "step": 760
466
+ },
467
+ {
468
+ "epoch": 0.26,
469
+ "learning_rate": 5e-05,
470
+ "loss": 0.5434,
471
+ "step": 770
472
+ },
473
+ {
474
+ "epoch": 0.26,
475
+ "learning_rate": 5e-05,
476
+ "loss": 0.5422,
477
+ "step": 780
478
+ },
479
+ {
480
+ "epoch": 0.27,
481
+ "learning_rate": 5e-05,
482
+ "loss": 0.5465,
483
+ "step": 790
484
+ },
485
+ {
486
+ "epoch": 0.27,
487
+ "learning_rate": 5e-05,
488
+ "loss": 0.5447,
489
+ "step": 800
490
+ },
491
+ {
492
+ "epoch": 0.27,
493
+ "learning_rate": 5e-05,
494
+ "loss": 0.5396,
495
+ "step": 810
496
+ },
497
+ {
498
+ "epoch": 0.28,
499
+ "learning_rate": 5e-05,
500
+ "loss": 0.543,
501
+ "step": 820
502
+ },
503
+ {
504
+ "epoch": 0.28,
505
+ "learning_rate": 5e-05,
506
+ "loss": 0.5461,
507
+ "step": 830
508
+ },
509
+ {
510
+ "epoch": 0.28,
511
+ "learning_rate": 5e-05,
512
+ "loss": 0.5482,
513
+ "step": 840
514
+ },
515
+ {
516
+ "epoch": 0.29,
517
+ "learning_rate": 5e-05,
518
+ "loss": 0.5426,
519
+ "step": 850
520
+ },
521
+ {
522
+ "epoch": 0.29,
523
+ "learning_rate": 5e-05,
524
+ "loss": 0.5453,
525
+ "step": 860
526
+ },
527
+ {
528
+ "epoch": 0.29,
529
+ "learning_rate": 5e-05,
530
+ "loss": 0.5493,
531
+ "step": 870
532
+ },
533
+ {
534
+ "epoch": 0.3,
535
+ "learning_rate": 5e-05,
536
+ "loss": 0.5477,
537
+ "step": 880
538
+ },
539
+ {
540
+ "epoch": 0.3,
541
+ "learning_rate": 5e-05,
542
+ "loss": 0.542,
543
+ "step": 890
544
+ },
545
+ {
546
+ "epoch": 0.3,
547
+ "learning_rate": 5e-05,
548
+ "loss": 0.5393,
549
+ "step": 900
550
+ },
551
+ {
552
+ "epoch": 0.31,
553
+ "learning_rate": 5e-05,
554
+ "loss": 0.5365,
555
+ "step": 910
556
+ },
557
+ {
558
+ "epoch": 0.31,
559
+ "learning_rate": 5e-05,
560
+ "loss": 0.5433,
561
+ "step": 920
562
+ },
563
+ {
564
+ "epoch": 0.31,
565
+ "learning_rate": 5e-05,
566
+ "loss": 0.542,
567
+ "step": 930
568
+ },
569
+ {
570
+ "epoch": 0.32,
571
+ "learning_rate": 5e-05,
572
+ "loss": 0.5368,
573
+ "step": 940
574
+ },
575
+ {
576
+ "epoch": 0.32,
577
+ "learning_rate": 5e-05,
578
+ "loss": 0.542,
579
+ "step": 950
580
+ },
581
+ {
582
+ "epoch": 0.32,
583
+ "learning_rate": 5e-05,
584
+ "loss": 0.5467,
585
+ "step": 960
586
+ },
587
+ {
588
+ "epoch": 0.33,
589
+ "learning_rate": 5e-05,
590
+ "loss": 0.5382,
591
+ "step": 970
592
+ },
593
+ {
594
+ "epoch": 0.33,
595
+ "learning_rate": 5e-05,
596
+ "loss": 0.5366,
597
+ "step": 980
598
+ },
599
+ {
600
+ "epoch": 0.33,
601
+ "learning_rate": 5e-05,
602
+ "loss": 0.5354,
603
+ "step": 990
604
+ },
605
+ {
606
+ "epoch": 0.34,
607
+ "learning_rate": 5e-05,
608
+ "loss": 0.5322,
609
+ "step": 1000
610
+ },
611
+ {
612
+ "epoch": 0.34,
613
+ "learning_rate": 5e-05,
614
+ "loss": 0.5379,
615
+ "step": 1010
616
+ },
617
+ {
618
+ "epoch": 0.34,
619
+ "learning_rate": 5e-05,
620
+ "loss": 0.5398,
621
+ "step": 1020
622
+ },
623
+ {
624
+ "epoch": 0.35,
625
+ "learning_rate": 5e-05,
626
+ "loss": 0.5337,
627
+ "step": 1030
628
+ },
629
+ {
630
+ "epoch": 0.35,
631
+ "learning_rate": 5e-05,
632
+ "loss": 0.5384,
633
+ "step": 1040
634
+ },
635
+ {
636
+ "epoch": 0.35,
637
+ "learning_rate": 5e-05,
638
+ "loss": 0.5256,
639
+ "step": 1050
640
+ },
641
+ {
642
+ "epoch": 0.36,
643
+ "learning_rate": 5e-05,
644
+ "loss": 0.5353,
645
+ "step": 1060
646
+ },
647
+ {
648
+ "epoch": 0.36,
649
+ "learning_rate": 5e-05,
650
+ "loss": 0.5233,
651
+ "step": 1070
652
+ },
653
+ {
654
+ "epoch": 0.36,
655
+ "learning_rate": 5e-05,
656
+ "loss": 0.5483,
657
+ "step": 1080
658
+ },
659
+ {
660
+ "epoch": 0.37,
661
+ "learning_rate": 5e-05,
662
+ "loss": 0.5362,
663
+ "step": 1090
664
+ },
665
+ {
666
+ "epoch": 0.37,
667
+ "learning_rate": 5e-05,
668
+ "loss": 0.5426,
669
+ "step": 1100
670
+ },
671
+ {
672
+ "epoch": 0.37,
673
+ "learning_rate": 5e-05,
674
+ "loss": 0.5347,
675
+ "step": 1110
676
+ },
677
+ {
678
+ "epoch": 0.38,
679
+ "learning_rate": 5e-05,
680
+ "loss": 0.5377,
681
+ "step": 1120
682
+ },
683
+ {
684
+ "epoch": 0.38,
685
+ "learning_rate": 5e-05,
686
+ "loss": 0.5374,
687
+ "step": 1130
688
+ },
689
+ {
690
+ "epoch": 0.38,
691
+ "learning_rate": 5e-05,
692
+ "loss": 0.532,
693
+ "step": 1140
694
+ },
695
+ {
696
+ "epoch": 0.39,
697
+ "learning_rate": 5e-05,
698
+ "loss": 0.5315,
699
+ "step": 1150
700
+ },
701
+ {
702
+ "epoch": 0.39,
703
+ "learning_rate": 5e-05,
704
+ "loss": 0.5274,
705
+ "step": 1160
706
+ },
707
+ {
708
+ "epoch": 0.39,
709
+ "learning_rate": 5e-05,
710
+ "loss": 0.537,
711
+ "step": 1170
712
+ },
713
+ {
714
+ "epoch": 0.4,
715
+ "learning_rate": 5e-05,
716
+ "loss": 0.5427,
717
+ "step": 1180
718
+ },
719
+ {
720
+ "epoch": 0.4,
721
+ "learning_rate": 5e-05,
722
+ "loss": 0.5288,
723
+ "step": 1190
724
+ },
725
+ {
726
+ "epoch": 0.4,
727
+ "learning_rate": 5e-05,
728
+ "loss": 0.5259,
729
+ "step": 1200
730
+ },
731
+ {
732
+ "epoch": 0.41,
733
+ "learning_rate": 5e-05,
734
+ "loss": 0.531,
735
+ "step": 1210
736
+ },
737
+ {
738
+ "epoch": 0.41,
739
+ "learning_rate": 5e-05,
740
+ "loss": 0.5263,
741
+ "step": 1220
742
+ },
743
+ {
744
+ "epoch": 0.41,
745
+ "learning_rate": 5e-05,
746
+ "loss": 0.5305,
747
+ "step": 1230
748
+ },
749
+ {
750
+ "epoch": 0.42,
751
+ "learning_rate": 5e-05,
752
+ "loss": 0.5245,
753
+ "step": 1240
754
+ },
755
+ {
756
+ "epoch": 0.42,
757
+ "learning_rate": 5e-05,
758
+ "loss": 0.5306,
759
+ "step": 1250
760
+ },
761
+ {
762
+ "epoch": 0.43,
763
+ "learning_rate": 5e-05,
764
+ "loss": 0.526,
765
+ "step": 1260
766
+ },
767
+ {
768
+ "epoch": 0.43,
769
+ "learning_rate": 5e-05,
770
+ "loss": 0.5308,
771
+ "step": 1270
772
+ },
773
+ {
774
+ "epoch": 0.43,
775
+ "learning_rate": 5e-05,
776
+ "loss": 0.5226,
777
+ "step": 1280
778
+ },
779
+ {
780
+ "epoch": 0.44,
781
+ "learning_rate": 5e-05,
782
+ "loss": 0.5295,
783
+ "step": 1290
784
+ },
785
+ {
786
+ "epoch": 0.44,
787
+ "learning_rate": 5e-05,
788
+ "loss": 0.5297,
789
+ "step": 1300
790
+ },
791
+ {
792
+ "epoch": 0.44,
793
+ "learning_rate": 5e-05,
794
+ "loss": 0.5293,
795
+ "step": 1310
796
+ },
797
+ {
798
+ "epoch": 0.45,
799
+ "learning_rate": 5e-05,
800
+ "loss": 0.5387,
801
+ "step": 1320
802
+ },
803
+ {
804
+ "epoch": 0.45,
805
+ "learning_rate": 5e-05,
806
+ "loss": 0.5252,
807
+ "step": 1330
808
+ },
809
+ {
810
+ "epoch": 0.45,
811
+ "learning_rate": 5e-05,
812
+ "loss": 0.524,
813
+ "step": 1340
814
+ },
815
+ {
816
+ "epoch": 0.46,
817
+ "learning_rate": 5e-05,
818
+ "loss": 0.5231,
819
+ "step": 1350
820
+ },
821
+ {
822
+ "epoch": 0.46,
823
+ "learning_rate": 5e-05,
824
+ "loss": 0.5248,
825
+ "step": 1360
826
+ },
827
+ {
828
+ "epoch": 0.46,
829
+ "learning_rate": 5e-05,
830
+ "loss": 0.5332,
831
+ "step": 1370
832
+ },
833
+ {
834
+ "epoch": 0.47,
835
+ "learning_rate": 5e-05,
836
+ "loss": 0.5366,
837
+ "step": 1380
838
+ },
839
+ {
840
+ "epoch": 0.47,
841
+ "learning_rate": 5e-05,
842
+ "loss": 0.5231,
843
+ "step": 1390
844
+ },
845
+ {
846
+ "epoch": 0.47,
847
+ "learning_rate": 5e-05,
848
+ "loss": 0.5206,
849
+ "step": 1400
850
+ },
851
+ {
852
+ "epoch": 0.48,
853
+ "learning_rate": 5e-05,
854
+ "loss": 0.5237,
855
+ "step": 1410
856
+ },
857
+ {
858
+ "epoch": 0.48,
859
+ "learning_rate": 5e-05,
860
+ "loss": 0.5201,
861
+ "step": 1420
862
+ },
863
+ {
864
+ "epoch": 0.48,
865
+ "learning_rate": 5e-05,
866
+ "loss": 0.5366,
867
+ "step": 1430
868
+ },
869
+ {
870
+ "epoch": 0.49,
871
+ "learning_rate": 5e-05,
872
+ "loss": 0.5336,
873
+ "step": 1440
874
+ },
875
+ {
876
+ "epoch": 0.49,
877
+ "learning_rate": 5e-05,
878
+ "loss": 0.5282,
879
+ "step": 1450
880
+ },
881
+ {
882
+ "epoch": 0.49,
883
+ "learning_rate": 5e-05,
884
+ "loss": 0.5269,
885
+ "step": 1460
886
+ },
887
+ {
888
+ "epoch": 0.5,
889
+ "learning_rate": 5e-05,
890
+ "loss": 0.5245,
891
+ "step": 1470
892
+ },
893
+ {
894
+ "epoch": 0.5,
895
+ "learning_rate": 5e-05,
896
+ "loss": 0.5203,
897
+ "step": 1480
898
+ },
899
+ {
900
+ "epoch": 0.5,
901
+ "learning_rate": 5e-05,
902
+ "loss": 0.5225,
903
+ "step": 1490
904
+ },
905
+ {
906
+ "epoch": 0.51,
907
+ "learning_rate": 5e-05,
908
+ "loss": 0.5311,
909
+ "step": 1500
910
+ },
911
+ {
912
+ "epoch": 0.51,
913
+ "learning_rate": 5e-05,
914
+ "loss": 0.5229,
915
+ "step": 1510
916
+ },
917
+ {
918
+ "epoch": 0.51,
919
+ "learning_rate": 5e-05,
920
+ "loss": 0.5266,
921
+ "step": 1520
922
+ },
923
+ {
924
+ "epoch": 0.52,
925
+ "learning_rate": 5e-05,
926
+ "loss": 0.5283,
927
+ "step": 1530
928
+ },
929
+ {
930
+ "epoch": 0.52,
931
+ "learning_rate": 5e-05,
932
+ "loss": 0.5279,
933
+ "step": 1540
934
+ },
935
+ {
936
+ "epoch": 0.52,
937
+ "learning_rate": 5e-05,
938
+ "loss": 0.5299,
939
+ "step": 1550
940
+ },
941
+ {
942
+ "epoch": 0.53,
943
+ "learning_rate": 5e-05,
944
+ "loss": 0.5267,
945
+ "step": 1560
946
+ },
947
+ {
948
+ "epoch": 0.53,
949
+ "learning_rate": 5e-05,
950
+ "loss": 0.5186,
951
+ "step": 1570
952
+ },
953
+ {
954
+ "epoch": 0.53,
955
+ "learning_rate": 5e-05,
956
+ "loss": 0.5238,
957
+ "step": 1580
958
+ },
959
+ {
960
+ "epoch": 0.54,
961
+ "learning_rate": 5e-05,
962
+ "loss": 0.5249,
963
+ "step": 1590
964
+ },
965
+ {
966
+ "epoch": 0.54,
967
+ "learning_rate": 5e-05,
968
+ "loss": 0.5225,
969
+ "step": 1600
970
+ },
971
+ {
972
+ "epoch": 0.54,
973
+ "learning_rate": 5e-05,
974
+ "loss": 0.5226,
975
+ "step": 1610
976
+ },
977
+ {
978
+ "epoch": 0.55,
979
+ "learning_rate": 5e-05,
980
+ "loss": 0.5188,
981
+ "step": 1620
982
+ },
983
+ {
984
+ "epoch": 0.55,
985
+ "learning_rate": 5e-05,
986
+ "loss": 0.525,
987
+ "step": 1630
988
+ },
989
+ {
990
+ "epoch": 0.55,
991
+ "learning_rate": 5e-05,
992
+ "loss": 0.5249,
993
+ "step": 1640
994
+ },
995
+ {
996
+ "epoch": 0.56,
997
+ "learning_rate": 5e-05,
998
+ "loss": 0.5225,
999
+ "step": 1650
1000
+ },
1001
+ {
1002
+ "epoch": 0.56,
1003
+ "learning_rate": 5e-05,
1004
+ "loss": 0.5281,
1005
+ "step": 1660
1006
+ },
1007
+ {
1008
+ "epoch": 0.56,
1009
+ "learning_rate": 5e-05,
1010
+ "loss": 0.5204,
1011
+ "step": 1670
1012
+ },
1013
+ {
1014
+ "epoch": 0.57,
1015
+ "learning_rate": 5e-05,
1016
+ "loss": 0.5193,
1017
+ "step": 1680
1018
+ },
1019
+ {
1020
+ "epoch": 0.57,
1021
+ "learning_rate": 5e-05,
1022
+ "loss": 0.5278,
1023
+ "step": 1690
1024
+ },
1025
+ {
1026
+ "epoch": 0.57,
1027
+ "learning_rate": 5e-05,
1028
+ "loss": 0.5224,
1029
+ "step": 1700
1030
+ },
1031
+ {
1032
+ "epoch": 0.58,
1033
+ "learning_rate": 5e-05,
1034
+ "loss": 0.5218,
1035
+ "step": 1710
1036
+ },
1037
+ {
1038
+ "epoch": 0.58,
1039
+ "learning_rate": 5e-05,
1040
+ "loss": 0.5171,
1041
+ "step": 1720
1042
+ },
1043
+ {
1044
+ "epoch": 0.58,
1045
+ "learning_rate": 5e-05,
1046
+ "loss": 0.5239,
1047
+ "step": 1730
1048
+ },
1049
+ {
1050
+ "epoch": 0.59,
1051
+ "learning_rate": 5e-05,
1052
+ "loss": 0.5231,
1053
+ "step": 1740
1054
+ },
1055
+ {
1056
+ "epoch": 0.59,
1057
+ "learning_rate": 5e-05,
1058
+ "loss": 0.5164,
1059
+ "step": 1750
1060
+ },
1061
+ {
1062
+ "epoch": 0.59,
1063
+ "learning_rate": 5e-05,
1064
+ "loss": 0.5181,
1065
+ "step": 1760
1066
+ },
1067
+ {
1068
+ "epoch": 0.6,
1069
+ "learning_rate": 5e-05,
1070
+ "loss": 0.5166,
1071
+ "step": 1770
1072
+ },
1073
+ {
1074
+ "epoch": 0.6,
1075
+ "learning_rate": 5e-05,
1076
+ "loss": 0.5179,
1077
+ "step": 1780
1078
+ },
1079
+ {
1080
+ "epoch": 0.6,
1081
+ "learning_rate": 5e-05,
1082
+ "loss": 0.5169,
1083
+ "step": 1790
1084
+ },
1085
+ {
1086
+ "epoch": 0.61,
1087
+ "learning_rate": 5e-05,
1088
+ "loss": 0.5275,
1089
+ "step": 1800
1090
+ },
1091
+ {
1092
+ "epoch": 0.61,
1093
+ "learning_rate": 5e-05,
1094
+ "loss": 0.5179,
1095
+ "step": 1810
1096
+ },
1097
+ {
1098
+ "epoch": 0.61,
1099
+ "learning_rate": 5e-05,
1100
+ "loss": 0.5216,
1101
+ "step": 1820
1102
+ },
1103
+ {
1104
+ "epoch": 0.62,
1105
+ "learning_rate": 5e-05,
1106
+ "loss": 0.5236,
1107
+ "step": 1830
1108
+ },
1109
+ {
1110
+ "epoch": 0.62,
1111
+ "learning_rate": 5e-05,
1112
+ "loss": 0.5125,
1113
+ "step": 1840
1114
+ },
1115
+ {
1116
+ "epoch": 0.62,
1117
+ "learning_rate": 5e-05,
1118
+ "loss": 0.527,
1119
+ "step": 1850
1120
+ },
1121
+ {
1122
+ "epoch": 0.63,
1123
+ "learning_rate": 5e-05,
1124
+ "loss": 0.5224,
1125
+ "step": 1860
1126
+ },
1127
+ {
1128
+ "epoch": 0.63,
1129
+ "learning_rate": 5e-05,
1130
+ "loss": 0.5304,
1131
+ "step": 1870
1132
+ },
1133
+ {
1134
+ "epoch": 0.63,
1135
+ "learning_rate": 5e-05,
1136
+ "loss": 0.5242,
1137
+ "step": 1880
1138
+ },
1139
+ {
1140
+ "epoch": 0.64,
1141
+ "learning_rate": 5e-05,
1142
+ "loss": 0.5351,
1143
+ "step": 1890
1144
+ },
1145
+ {
1146
+ "epoch": 0.64,
1147
+ "learning_rate": 5e-05,
1148
+ "loss": 0.5262,
1149
+ "step": 1900
1150
+ },
1151
+ {
1152
+ "epoch": 0.64,
1153
+ "learning_rate": 5e-05,
1154
+ "loss": 0.5178,
1155
+ "step": 1910
1156
+ },
1157
+ {
1158
+ "epoch": 0.65,
1159
+ "learning_rate": 5e-05,
1160
+ "loss": 0.5104,
1161
+ "step": 1920
1162
+ },
1163
+ {
1164
+ "epoch": 0.65,
1165
+ "learning_rate": 5e-05,
1166
+ "loss": 0.5172,
1167
+ "step": 1930
1168
+ },
1169
+ {
1170
+ "epoch": 0.65,
1171
+ "learning_rate": 5e-05,
1172
+ "loss": 0.517,
1173
+ "step": 1940
1174
+ },
1175
+ {
1176
+ "epoch": 0.66,
1177
+ "learning_rate": 5e-05,
1178
+ "loss": 0.5154,
1179
+ "step": 1950
1180
+ },
1181
+ {
1182
+ "epoch": 0.66,
1183
+ "learning_rate": 5e-05,
1184
+ "loss": 0.5285,
1185
+ "step": 1960
1186
+ },
1187
+ {
1188
+ "epoch": 0.66,
1189
+ "learning_rate": 5e-05,
1190
+ "loss": 0.5163,
1191
+ "step": 1970
1192
+ },
1193
+ {
1194
+ "epoch": 0.67,
1195
+ "learning_rate": 5e-05,
1196
+ "loss": 0.5104,
1197
+ "step": 1980
1198
+ },
1199
+ {
1200
+ "epoch": 0.67,
1201
+ "learning_rate": 5e-05,
1202
+ "loss": 0.5228,
1203
+ "step": 1990
1204
+ },
1205
+ {
1206
+ "epoch": 0.67,
1207
+ "learning_rate": 5e-05,
1208
+ "loss": 0.5131,
1209
+ "step": 2000
1210
+ },
1211
+ {
1212
+ "epoch": 0.68,
1213
+ "learning_rate": 5e-05,
1214
+ "loss": 0.5173,
1215
+ "step": 2010
1216
+ },
1217
+ {
1218
+ "epoch": 0.68,
1219
+ "learning_rate": 5e-05,
1220
+ "loss": 0.5137,
1221
+ "step": 2020
1222
+ },
1223
+ {
1224
+ "epoch": 0.68,
1225
+ "learning_rate": 5e-05,
1226
+ "loss": 0.5132,
1227
+ "step": 2030
1228
+ },
1229
+ {
1230
+ "epoch": 0.69,
1231
+ "learning_rate": 5e-05,
1232
+ "loss": 0.5154,
1233
+ "step": 2040
1234
+ },
1235
+ {
1236
+ "epoch": 0.69,
1237
+ "learning_rate": 5e-05,
1238
+ "loss": 0.5194,
1239
+ "step": 2050
1240
+ },
1241
+ {
1242
+ "epoch": 0.69,
1243
+ "learning_rate": 5e-05,
1244
+ "loss": 0.521,
1245
+ "step": 2060
1246
+ },
1247
+ {
1248
+ "epoch": 0.7,
1249
+ "learning_rate": 5e-05,
1250
+ "loss": 0.5166,
1251
+ "step": 2070
1252
+ },
1253
+ {
1254
+ "epoch": 0.7,
1255
+ "learning_rate": 5e-05,
1256
+ "loss": 0.5161,
1257
+ "step": 2080
1258
+ },
1259
+ {
1260
+ "epoch": 0.71,
1261
+ "learning_rate": 5e-05,
1262
+ "loss": 0.5143,
1263
+ "step": 2090
1264
+ },
1265
+ {
1266
+ "epoch": 0.71,
1267
+ "learning_rate": 5e-05,
1268
+ "loss": 0.5133,
1269
+ "step": 2100
1270
+ },
1271
+ {
1272
+ "epoch": 0.71,
1273
+ "learning_rate": 5e-05,
1274
+ "loss": 0.5188,
1275
+ "step": 2110
1276
+ },
1277
+ {
1278
+ "epoch": 0.72,
1279
+ "learning_rate": 5e-05,
1280
+ "loss": 0.5099,
1281
+ "step": 2120
1282
+ },
1283
+ {
1284
+ "epoch": 0.72,
1285
+ "learning_rate": 5e-05,
1286
+ "loss": 0.5181,
1287
+ "step": 2130
1288
+ },
1289
+ {
1290
+ "epoch": 0.72,
1291
+ "learning_rate": 5e-05,
1292
+ "loss": 0.5148,
1293
+ "step": 2140
1294
+ },
1295
+ {
1296
+ "epoch": 0.73,
1297
+ "learning_rate": 5e-05,
1298
+ "loss": 0.5155,
1299
+ "step": 2150
1300
+ },
1301
+ {
1302
+ "epoch": 0.73,
1303
+ "learning_rate": 5e-05,
1304
+ "loss": 0.5108,
1305
+ "step": 2160
1306
+ },
1307
+ {
1308
+ "epoch": 0.73,
1309
+ "learning_rate": 5e-05,
1310
+ "loss": 0.5026,
1311
+ "step": 2170
1312
+ },
1313
+ {
1314
+ "epoch": 0.74,
1315
+ "learning_rate": 5e-05,
1316
+ "loss": 0.5187,
1317
+ "step": 2180
1318
+ },
1319
+ {
1320
+ "epoch": 0.74,
1321
+ "learning_rate": 5e-05,
1322
+ "loss": 0.5156,
1323
+ "step": 2190
1324
+ },
1325
+ {
1326
+ "epoch": 0.74,
1327
+ "learning_rate": 5e-05,
1328
+ "loss": 0.5164,
1329
+ "step": 2200
1330
+ },
1331
+ {
1332
+ "epoch": 0.75,
1333
+ "learning_rate": 5e-05,
1334
+ "loss": 0.5129,
1335
+ "step": 2210
1336
+ },
1337
+ {
1338
+ "epoch": 0.75,
1339
+ "learning_rate": 5e-05,
1340
+ "loss": 0.5143,
1341
+ "step": 2220
1342
+ },
1343
+ {
1344
+ "epoch": 0.75,
1345
+ "learning_rate": 5e-05,
1346
+ "loss": 0.5206,
1347
+ "step": 2230
1348
+ },
1349
+ {
1350
+ "epoch": 0.76,
1351
+ "learning_rate": 5e-05,
1352
+ "loss": 0.508,
1353
+ "step": 2240
1354
+ },
1355
+ {
1356
+ "epoch": 0.76,
1357
+ "learning_rate": 5e-05,
1358
+ "loss": 0.5101,
1359
+ "step": 2250
1360
+ },
1361
+ {
1362
+ "epoch": 0.76,
1363
+ "learning_rate": 5e-05,
1364
+ "loss": 0.5069,
1365
+ "step": 2260
1366
+ },
1367
+ {
1368
+ "epoch": 0.77,
1369
+ "learning_rate": 5e-05,
1370
+ "loss": 0.5082,
1371
+ "step": 2270
1372
+ },
1373
+ {
1374
+ "epoch": 0.77,
1375
+ "learning_rate": 5e-05,
1376
+ "loss": 0.5206,
1377
+ "step": 2280
1378
+ },
1379
+ {
1380
+ "epoch": 0.77,
1381
+ "learning_rate": 5e-05,
1382
+ "loss": 0.5134,
1383
+ "step": 2290
1384
+ },
1385
+ {
1386
+ "epoch": 0.78,
1387
+ "learning_rate": 5e-05,
1388
+ "loss": 0.5099,
1389
+ "step": 2300
1390
+ },
1391
+ {
1392
+ "epoch": 0.78,
1393
+ "learning_rate": 5e-05,
1394
+ "loss": 0.5114,
1395
+ "step": 2310
1396
+ },
1397
+ {
1398
+ "epoch": 0.78,
1399
+ "learning_rate": 5e-05,
1400
+ "loss": 0.5132,
1401
+ "step": 2320
1402
+ },
1403
+ {
1404
+ "epoch": 0.79,
1405
+ "learning_rate": 5e-05,
1406
+ "loss": 0.5125,
1407
+ "step": 2330
1408
+ },
1409
+ {
1410
+ "epoch": 0.79,
1411
+ "learning_rate": 5e-05,
1412
+ "loss": 0.5055,
1413
+ "step": 2340
1414
+ },
1415
+ {
1416
+ "epoch": 0.79,
1417
+ "learning_rate": 5e-05,
1418
+ "loss": 0.514,
1419
+ "step": 2350
1420
+ },
1421
+ {
1422
+ "epoch": 0.8,
1423
+ "learning_rate": 5e-05,
1424
+ "loss": 0.5093,
1425
+ "step": 2360
1426
+ },
1427
+ {
1428
+ "epoch": 0.8,
1429
+ "learning_rate": 5e-05,
1430
+ "loss": 0.506,
1431
+ "step": 2370
1432
+ },
1433
+ {
1434
+ "epoch": 0.8,
1435
+ "learning_rate": 5e-05,
1436
+ "loss": 0.5198,
1437
+ "step": 2380
1438
+ },
1439
+ {
1440
+ "epoch": 0.81,
1441
+ "learning_rate": 5e-05,
1442
+ "loss": 0.5187,
1443
+ "step": 2390
1444
+ },
1445
+ {
1446
+ "epoch": 0.81,
1447
+ "learning_rate": 5e-05,
1448
+ "loss": 0.5112,
1449
+ "step": 2400
1450
+ },
1451
+ {
1452
+ "epoch": 0.81,
1453
+ "learning_rate": 5e-05,
1454
+ "loss": 0.5189,
1455
+ "step": 2410
1456
+ },
1457
+ {
1458
+ "epoch": 0.82,
1459
+ "learning_rate": 5e-05,
1460
+ "loss": 0.5104,
1461
+ "step": 2420
1462
+ },
1463
+ {
1464
+ "epoch": 0.82,
1465
+ "learning_rate": 5e-05,
1466
+ "loss": 0.5105,
1467
+ "step": 2430
1468
+ },
1469
+ {
1470
+ "epoch": 0.82,
1471
+ "learning_rate": 5e-05,
1472
+ "loss": 0.5165,
1473
+ "step": 2440
1474
+ },
1475
+ {
1476
+ "epoch": 0.83,
1477
+ "learning_rate": 5e-05,
1478
+ "loss": 0.5088,
1479
+ "step": 2450
1480
+ },
1481
+ {
1482
+ "epoch": 0.83,
1483
+ "learning_rate": 5e-05,
1484
+ "loss": 0.504,
1485
+ "step": 2460
1486
+ },
1487
+ {
1488
+ "epoch": 0.83,
1489
+ "learning_rate": 5e-05,
1490
+ "loss": 0.5065,
1491
+ "step": 2470
1492
+ },
1493
+ {
1494
+ "epoch": 0.84,
1495
+ "learning_rate": 5e-05,
1496
+ "loss": 0.5053,
1497
+ "step": 2480
1498
+ },
1499
+ {
1500
+ "epoch": 0.84,
1501
+ "learning_rate": 5e-05,
1502
+ "loss": 0.5135,
1503
+ "step": 2490
1504
+ },
1505
+ {
1506
+ "epoch": 0.84,
1507
+ "learning_rate": 5e-05,
1508
+ "loss": 0.5078,
1509
+ "step": 2500
1510
+ },
1511
+ {
1512
+ "epoch": 0.85,
1513
+ "learning_rate": 5e-05,
1514
+ "loss": 0.508,
1515
+ "step": 2510
1516
+ },
1517
+ {
1518
+ "epoch": 0.85,
1519
+ "learning_rate": 5e-05,
1520
+ "loss": 0.5038,
1521
+ "step": 2520
1522
+ },
1523
+ {
1524
+ "epoch": 0.85,
1525
+ "learning_rate": 5e-05,
1526
+ "loss": 0.5114,
1527
+ "step": 2530
1528
+ },
1529
+ {
1530
+ "epoch": 0.86,
1531
+ "learning_rate": 5e-05,
1532
+ "loss": 0.5055,
1533
+ "step": 2540
1534
+ },
1535
+ {
1536
+ "epoch": 0.86,
1537
+ "learning_rate": 5e-05,
1538
+ "loss": 0.5011,
1539
+ "step": 2550
1540
+ },
1541
+ {
1542
+ "epoch": 0.86,
1543
+ "learning_rate": 5e-05,
1544
+ "loss": 0.5025,
1545
+ "step": 2560
1546
+ },
1547
+ {
1548
+ "epoch": 0.87,
1549
+ "learning_rate": 5e-05,
1550
+ "loss": 0.5113,
1551
+ "step": 2570
1552
+ },
1553
+ {
1554
+ "epoch": 0.87,
1555
+ "learning_rate": 5e-05,
1556
+ "loss": 0.506,
1557
+ "step": 2580
1558
+ },
1559
+ {
1560
+ "epoch": 0.87,
1561
+ "learning_rate": 5e-05,
1562
+ "loss": 0.5031,
1563
+ "step": 2590
1564
+ },
1565
+ {
1566
+ "epoch": 0.88,
1567
+ "learning_rate": 5e-05,
1568
+ "loss": 0.505,
1569
+ "step": 2600
1570
+ },
1571
+ {
1572
+ "epoch": 0.88,
1573
+ "learning_rate": 5e-05,
1574
+ "loss": 0.4966,
1575
+ "step": 2610
1576
+ },
1577
+ {
1578
+ "epoch": 0.88,
1579
+ "learning_rate": 5e-05,
1580
+ "loss": 0.5132,
1581
+ "step": 2620
1582
+ },
1583
+ {
1584
+ "epoch": 0.89,
1585
+ "learning_rate": 5e-05,
1586
+ "loss": 0.4975,
1587
+ "step": 2630
1588
+ },
1589
+ {
1590
+ "epoch": 0.89,
1591
+ "learning_rate": 5e-05,
1592
+ "loss": 0.5108,
1593
+ "step": 2640
1594
+ },
1595
+ {
1596
+ "epoch": 0.89,
1597
+ "learning_rate": 5e-05,
1598
+ "loss": 0.508,
1599
+ "step": 2650
1600
+ },
1601
+ {
1602
+ "epoch": 0.9,
1603
+ "learning_rate": 5e-05,
1604
+ "loss": 0.514,
1605
+ "step": 2660
1606
+ },
1607
+ {
1608
+ "epoch": 0.9,
1609
+ "learning_rate": 5e-05,
1610
+ "loss": 0.5071,
1611
+ "step": 2670
1612
+ },
1613
+ {
1614
+ "epoch": 0.9,
1615
+ "learning_rate": 5e-05,
1616
+ "loss": 0.4952,
1617
+ "step": 2680
1618
+ },
1619
+ {
1620
+ "epoch": 0.91,
1621
+ "learning_rate": 5e-05,
1622
+ "loss": 0.5021,
1623
+ "step": 2690
1624
+ },
1625
+ {
1626
+ "epoch": 0.91,
1627
+ "learning_rate": 5e-05,
1628
+ "loss": 0.5112,
1629
+ "step": 2700
1630
+ },
1631
+ {
1632
+ "epoch": 0.91,
1633
+ "learning_rate": 5e-05,
1634
+ "loss": 0.5037,
1635
+ "step": 2710
1636
+ },
1637
+ {
1638
+ "epoch": 0.92,
1639
+ "learning_rate": 5e-05,
1640
+ "loss": 0.5106,
1641
+ "step": 2720
1642
+ },
1643
+ {
1644
+ "epoch": 0.92,
1645
+ "learning_rate": 5e-05,
1646
+ "loss": 0.5011,
1647
+ "step": 2730
1648
+ },
1649
+ {
1650
+ "epoch": 0.92,
1651
+ "learning_rate": 5e-05,
1652
+ "loss": 0.5103,
1653
+ "step": 2740
1654
+ },
1655
+ {
1656
+ "epoch": 0.93,
1657
+ "learning_rate": 5e-05,
1658
+ "loss": 0.5004,
1659
+ "step": 2750
1660
+ },
1661
+ {
1662
+ "epoch": 0.93,
1663
+ "learning_rate": 5e-05,
1664
+ "loss": 0.5045,
1665
+ "step": 2760
1666
+ },
1667
+ {
1668
+ "epoch": 0.93,
1669
+ "learning_rate": 5e-05,
1670
+ "loss": 0.5019,
1671
+ "step": 2770
1672
+ },
1673
+ {
1674
+ "epoch": 0.94,
1675
+ "learning_rate": 5e-05,
1676
+ "loss": 0.4928,
1677
+ "step": 2780
1678
+ },
1679
+ {
1680
+ "epoch": 0.94,
1681
+ "learning_rate": 5e-05,
1682
+ "loss": 0.5029,
1683
+ "step": 2790
1684
+ },
1685
+ {
1686
+ "epoch": 0.94,
1687
+ "learning_rate": 5e-05,
1688
+ "loss": 0.5175,
1689
+ "step": 2800
1690
+ },
1691
+ {
1692
+ "epoch": 0.95,
1693
+ "learning_rate": 5e-05,
1694
+ "loss": 0.5114,
1695
+ "step": 2810
1696
+ },
1697
+ {
1698
+ "epoch": 0.95,
1699
+ "learning_rate": 5e-05,
1700
+ "loss": 0.5021,
1701
+ "step": 2820
1702
+ },
1703
+ {
1704
+ "epoch": 0.95,
1705
+ "learning_rate": 5e-05,
1706
+ "loss": 0.5084,
1707
+ "step": 2830
1708
+ },
1709
+ {
1710
+ "epoch": 0.96,
1711
+ "learning_rate": 5e-05,
1712
+ "loss": 0.5008,
1713
+ "step": 2840
1714
+ },
1715
+ {
1716
+ "epoch": 0.96,
1717
+ "learning_rate": 5e-05,
1718
+ "loss": 0.5102,
1719
+ "step": 2850
1720
+ },
1721
+ {
1722
+ "epoch": 0.96,
1723
+ "learning_rate": 5e-05,
1724
+ "loss": 0.5133,
1725
+ "step": 2860
1726
+ },
1727
+ {
1728
+ "epoch": 0.97,
1729
+ "learning_rate": 5e-05,
1730
+ "loss": 0.508,
1731
+ "step": 2870
1732
+ },
1733
+ {
1734
+ "epoch": 0.97,
1735
+ "learning_rate": 5e-05,
1736
+ "loss": 0.5054,
1737
+ "step": 2880
1738
+ },
1739
+ {
1740
+ "epoch": 0.97,
1741
+ "learning_rate": 5e-05,
1742
+ "loss": 0.5191,
1743
+ "step": 2890
1744
+ },
1745
+ {
1746
+ "epoch": 0.98,
1747
+ "learning_rate": 5e-05,
1748
+ "loss": 0.5051,
1749
+ "step": 2900
1750
+ },
1751
+ {
1752
+ "epoch": 0.98,
1753
+ "learning_rate": 5e-05,
1754
+ "loss": 0.4943,
1755
+ "step": 2910
1756
+ },
1757
+ {
1758
+ "epoch": 0.99,
1759
+ "learning_rate": 5e-05,
1760
+ "loss": 0.5044,
1761
+ "step": 2920
1762
+ },
1763
+ {
1764
+ "epoch": 0.99,
1765
+ "learning_rate": 5e-05,
1766
+ "loss": 0.4943,
1767
+ "step": 2930
1768
+ },
1769
+ {
1770
+ "epoch": 0.99,
1771
+ "learning_rate": 5e-05,
1772
+ "loss": 0.504,
1773
+ "step": 2940
1774
+ },
1775
+ {
1776
+ "epoch": 1.0,
1777
+ "learning_rate": 5e-05,
1778
+ "loss": 0.4978,
1779
+ "step": 2950
1780
+ },
1781
+ {
1782
+ "epoch": 1.0,
1783
+ "learning_rate": 5e-05,
1784
+ "loss": 0.492,
1785
+ "step": 2960
1786
+ },
1787
+ {
1788
+ "epoch": 1.0,
1789
+ "learning_rate": 5e-05,
1790
+ "loss": 0.4938,
1791
+ "step": 2970
1792
+ },
1793
+ {
1794
+ "epoch": 1.01,
1795
+ "learning_rate": 5e-05,
1796
+ "loss": 0.5005,
1797
+ "step": 2980
1798
+ },
1799
+ {
1800
+ "epoch": 1.01,
1801
+ "learning_rate": 5e-05,
1802
+ "loss": 0.4997,
1803
+ "step": 2990
1804
+ },
1805
+ {
1806
+ "epoch": 1.01,
1807
+ "learning_rate": 5e-05,
1808
+ "loss": 0.4968,
1809
+ "step": 3000
1810
+ }
1811
+ ],
1812
+ "logging_steps": 10,
1813
+ "max_steps": 8892,
1814
+ "num_input_tokens_seen": 0,
1815
+ "num_train_epochs": 3,
1816
+ "save_steps": 1000,
1817
+ "total_flos": 0.0,
1818
+ "train_batch_size": 8,
1819
+ "trial_name": null,
1820
+ "trial_params": null
1821
+ }
cvnet_small_v1_global/checkpoint-3000/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:dddfb786197274aad152ee4c2cc9787279a04f3c148146f1b57e85d4fb19f6fc
3
+ size 4856
cvnet_small_v1_global/checkpoint-4000/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d25bb4e965b6951e2ee3ab893a0d0ce4998c4876a8469be3e67a9e30cbf24767
3
+ size 237860264
cvnet_small_v1_global/checkpoint-4000/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:eb4094205d4f0d797c40d85e82a83e95c93d5a364494deadaafd79984544eee9
3
+ size 475804282
cvnet_small_v1_global/checkpoint-4000/rng_state_0.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:30e9d80f2c7b48e823660edbe971433389c5987502ff5b8904c6fa06a4b8fac4
3
+ size 15920
cvnet_small_v1_global/checkpoint-4000/rng_state_1.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d6dc64ef283107cc20517b99d8a4260d5dc35fe0739df81e0627d71f753f071e
3
+ size 15920
cvnet_small_v1_global/checkpoint-4000/rng_state_2.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3984ffaba3ef211236ed7149a8baca5ccfdc9f355d0c4a2f94ee586ef97610e1
3
+ size 15920
cvnet_small_v1_global/checkpoint-4000/rng_state_3.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:aef672fc70b0c0d7f41d3aa0c8a5fc28c8b3a305a6b083477713cb302d8d72a0
3
+ size 15920
cvnet_small_v1_global/checkpoint-4000/rng_state_4.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7a0341c1736e01084f618482fe049b36f4ff83d3c592a9632213c226953a9084
3
+ size 15920
cvnet_small_v1_global/checkpoint-4000/rng_state_5.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:89b1c55b3a76909bc8b26c01f698db08cb90488d67ec9d13d047aea4788fd653
3
+ size 15920
cvnet_small_v1_global/checkpoint-4000/rng_state_6.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:73db3b651070332fced69b13785993fdefdaac849cbcb23e06c08cde7fdc5d55
3
+ size 15920
cvnet_small_v1_global/checkpoint-4000/rng_state_7.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8f21a5acb7274574d65d51c6e1a0d8b70345cf7b9c9dd99e4b85e293546a6e40
3
+ size 15920
cvnet_small_v1_global/checkpoint-4000/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b1f5b5ac4a53f35c9923ba2e114d395c8f0ca2ff8283a0adc3fb6db30ab7f070
3
+ size 1064