Hugging Face's logo

jh-yi
/
Video-Panda-7B

Video-Text-to-Text
Transformers
Inference Endpoints
Model card Files Community
1
jh-yi commited on
Commit
6e2080c
·
verified ·
1 Parent(s): a193587

Upload 20 files

Browse files
Files changed (20) hide show
  1. videopanda_fitu/config.json +46 -0
  2. videopanda_fitu/generation_config.json +9 -0
  3. videopanda_fitu/pytorch_model-00001-of-00002.bin +3 -0
  4. videopanda_fitu/pytorch_model-00002-of-00002.bin +3 -0
  5. videopanda_fitu/pytorch_model.bin.index.json +0 -0
  6. videopanda_fitu/special_tokens_map.json +24 -0
  7. videopanda_fitu/tokenizer.model +3 -0
  8. videopanda_fitu/tokenizer_config.json +35 -0
  9. videopanda_fitu/trainer_state.json +655 -0
  10. videopanda_fitu/training_args.bin +3 -0
  11. videopanda_prtr1/config.json +46 -0
  12. videopanda_prtr1/generation_config.json +9 -0
  13. videopanda_prtr1/pytorch_model-00001-of-00002.bin +3 -0
  14. videopanda_prtr1/pytorch_model-00002-of-00002.bin +3 -0
  15. videopanda_prtr1/pytorch_model.bin.index.json +0 -0
  16. videopanda_prtr1/special_tokens_map.json +24 -0
  17. videopanda_prtr1/tokenizer.model +3 -0
  18. videopanda_prtr1/tokenizer_config.json +35 -0
  19. videopanda_prtr1/trainer_state.json +2227 -0
  20. videopanda_prtr1/training_args.bin +3 -0
videopanda_fitu/config.json ADDED
@@ -0,0 +1,46 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "checkpoints/Video-Panda-7B/videopanda_prtr1",
3
+ "architectures": ["VideoPandaLlamaForCausalLM"],
4
+ "bos_token_id": 1,
5
+ "eos_token_id": 2,
6
+ "hidden_act": "silu",
7
+ "hidden_size": 4096,
8
+ "image_aspect_ratio": "square",
9
+ "image_grid_pinpoints": null,
10
+ "initializer_range": 0.02,
11
+ "intermediate_size": 11008,
12
+ "max_position_embeddings": 4096,
13
+ "mm_hidden_size": 1024,
14
+ "mm_projector_lr": 2e-05,
15
+ "mm_projector_type": "mlp2x_gelu",
16
+ "mm_use_im_patch_token": false,
17
+ "mm_use_im_start_end": false,
18
+ "mm_video_tower": "videopanda/config/videopanda-video-LB-patch14-anypixel-448",
19
+ "mm_video_tower_teacher": "LanguageBind/LanguageBind_Video_merge",
20
+ "mm_vision_select_feature": "patch",
21
+ "mm_vision_select_layer": -2,
22
+ "mm_vision_tower": "videopanda/config/videopanda-vision-LB-patch14-anypixel-448",
23
+ "mm_vision_tower_teacher": "LanguageBind/LanguageBind_Image",
24
+ "model_type": "videopanda",
25
+ "num_attention_heads": 32,
26
+ "num_hidden_layers": 32,
27
+ "num_key_value_heads": 32,
28
+ "pad_token_id": 0,
29
+ "pretraining_tp": 1,
30
+ "requires_video_distill": true,
31
+ "requires_image_distill": false,
32
+ "rms_norm_eps": 1e-05,
33
+ "rope_scaling": null,
34
+ "tie_word_embeddings": false,
35
+ "torch_dtype": "bfloat16",
36
+ "transformers_version": "4.31.0",
37
+ "tune_mm_mlp_adapter": false,
38
+ "use_cache": true,
39
+ "use_mm_proj": true,
40
+ "video_hidden_size": 1024,
41
+ "video_projector_lr": 2e-05,
42
+ "video_tower_lr": 2e-05,
43
+ "video_tower_teacher": "LanguageBind/LanguageBind_Video_merge",
44
+ "vision_tower_lr": 2e-05,
45
+ "vocab_size": 32000
46
+ }
videopanda_fitu/generation_config.json ADDED
@@ -0,0 +1,9 @@
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "bos_token_id": 1,
3
+ "eos_token_id": 2,
4
+ "max_length": 4096,
5
+ "pad_token_id": 0,
6
+ "temperature": 0.9,
7
+ "top_p": 0.6,
8
+ "transformers_version": "4.31.0"
9
+ }
videopanda_fitu/pytorch_model-00001-of-00002.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:295d0055a90975f83e76a76479aa460c562d1e2e36fca47b95dc5a629ffe3dc8
3
+ size 9976631973
videopanda_fitu/pytorch_model-00002-of-00002.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:839292103f724a4067a4cdc76674654d641659b37fc4ca5c30db78a7d6525abc
3
+ size 6688819122
videopanda_fitu/pytorch_model.bin.index.json ADDED
The diff for this file is too large to render. See raw diff
 
videopanda_fitu/special_tokens_map.json ADDED
@@ -0,0 +1,24 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "bos_token": {
3
+ "content": "<s>",
4
+ "lstrip": false,
5
+ "normalized": false,
6
+ "rstrip": false,
7
+ "single_word": false
8
+ },
9
+ "eos_token": {
10
+ "content": "</s>",
11
+ "lstrip": false,
12
+ "normalized": false,
13
+ "rstrip": false,
14
+ "single_word": false
15
+ },
16
+ "pad_token": "<unk>",
17
+ "unk_token": {
18
+ "content": "<unk>",
19
+ "lstrip": false,
20
+ "normalized": false,
21
+ "rstrip": false,
22
+ "single_word": false
23
+ }
24
+ }
videopanda_fitu/tokenizer.model ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9e556afd44213b6bd1be2b850ebbbd98f5481437a8021afaf58ee7fb1818d347
3
+ size 499723
videopanda_fitu/tokenizer_config.json ADDED
@@ -0,0 +1,35 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "add_bos_token": true,
3
+ "add_eos_token": false,
4
+ "bos_token": {
5
+ "__type": "AddedToken",
6
+ "content": "<s>",
7
+ "lstrip": false,
8
+ "normalized": false,
9
+ "rstrip": false,
10
+ "single_word": false
11
+ },
12
+ "clean_up_tokenization_spaces": false,
13
+ "eos_token": {
14
+ "__type": "AddedToken",
15
+ "content": "</s>",
16
+ "lstrip": false,
17
+ "normalized": false,
18
+ "rstrip": false,
19
+ "single_word": false
20
+ },
21
+ "legacy": false,
22
+ "model_max_length": 4096,
23
+ "pad_token": null,
24
+ "padding_side": "right",
25
+ "sp_model_kwargs": {},
26
+ "tokenizer_class": "LlamaTokenizer",
27
+ "unk_token": {
28
+ "__type": "AddedToken",
29
+ "content": "<unk>",
30
+ "lstrip": false,
31
+ "normalized": false,
32
+ "rstrip": false,
33
+ "single_word": false
34
+ }
35
+ }
videopanda_fitu/trainer_state.json ADDED
@@ -0,0 +1,655 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": null,
3
+ "best_model_checkpoint": null,
4
+ "epoch": 1.0,
5
+ "global_step": 105,
6
+ "is_hyper_param_search": false,
7
+ "is_local_process_zero": true,
8
+ "is_world_process_zero": true,
9
+ "log_history": [
10
+ {
11
+ "epoch": 0.01,
12
+ "learning_rate": 1e-05,
13
+ "loss": 3.0768,
14
+ "step": 1
15
+ },
16
+ {
17
+ "epoch": 0.02,
18
+ "learning_rate": 2e-05,
19
+ "loss": 3.076,
20
+ "step": 2
21
+ },
22
+ {
23
+ "epoch": 0.03,
24
+ "learning_rate": 1.9995348836233517e-05,
25
+ "loss": 2.7773,
26
+ "step": 3
27
+ },
28
+ {
29
+ "epoch": 0.04,
30
+ "learning_rate": 1.998139967159894e-05,
31
+ "loss": 2.4612,
32
+ "step": 4
33
+ },
34
+ {
35
+ "epoch": 0.05,
36
+ "learning_rate": 1.9958165482066094e-05,
37
+ "loss": 2.3567,
38
+ "step": 5
39
+ },
40
+ {
41
+ "epoch": 0.06,
42
+ "learning_rate": 1.992566788083908e-05,
43
+ "loss": 2.2825,
44
+ "step": 6
45
+ },
46
+ {
47
+ "epoch": 0.07,
48
+ "learning_rate": 1.9883937098250962e-05,
49
+ "loss": 2.2337,
50
+ "step": 7
51
+ },
52
+ {
53
+ "epoch": 0.08,
54
+ "learning_rate": 1.9833011953642525e-05,
55
+ "loss": 2.174,
56
+ "step": 8
57
+ },
58
+ {
59
+ "epoch": 0.09,
60
+ "learning_rate": 1.9772939819251247e-05,
61
+ "loss": 2.1664,
62
+ "step": 9
63
+ },
64
+ {
65
+ "epoch": 0.1,
66
+ "learning_rate": 1.9703776576144106e-05,
67
+ "loss": 2.1585,
68
+ "step": 10
69
+ },
70
+ {
71
+ "epoch": 0.1,
72
+ "learning_rate": 1.962558656223516e-05,
73
+ "loss": 2.1401,
74
+ "step": 11
75
+ },
76
+ {
77
+ "epoch": 0.11,
78
+ "learning_rate": 1.953844251243633e-05,
79
+ "loss": 2.1218,
80
+ "step": 12
81
+ },
82
+ {
83
+ "epoch": 0.12,
84
+ "learning_rate": 1.9442425490996987e-05,
85
+ "loss": 2.1043,
86
+ "step": 13
87
+ },
88
+ {
89
+ "epoch": 0.13,
90
+ "learning_rate": 1.933762481609536e-05,
91
+ "loss": 2.1185,
92
+ "step": 14
93
+ },
94
+ {
95
+ "epoch": 0.14,
96
+ "learning_rate": 1.9224137976751797e-05,
97
+ "loss": 2.1042,
98
+ "step": 15
99
+ },
100
+ {
101
+ "epoch": 0.15,
102
+ "learning_rate": 1.910207054214133e-05,
103
+ "loss": 2.0782,
104
+ "step": 16
105
+ },
106
+ {
107
+ "epoch": 0.16,
108
+ "learning_rate": 1.8971536063389745e-05,
109
+ "loss": 2.0411,
110
+ "step": 17
111
+ },
112
+ {
113
+ "epoch": 0.17,
114
+ "learning_rate": 1.8832655967944607e-05,
115
+ "loss": 2.0752,
116
+ "step": 18
117
+ },
118
+ {
119
+ "epoch": 0.18,
120
+ "learning_rate": 1.868555944661949e-05,
121
+ "loss": 2.0272,
122
+ "step": 19
123
+ },
124
+ {
125
+ "epoch": 0.19,
126
+ "learning_rate": 1.853038333341642e-05,
127
+ "loss": 2.0785,
128
+ "step": 20
129
+ },
130
+ {
131
+ "epoch": 0.2,
132
+ "learning_rate": 1.8367271978238422e-05,
133
+ "loss": 2.0635,
134
+ "step": 21
135
+ },
136
+ {
137
+ "epoch": 0.21,
138
+ "learning_rate": 1.8196377112610524e-05,
139
+ "loss": 2.0478,
140
+ "step": 22
141
+ },
142
+ {
143
+ "epoch": 0.22,
144
+ "learning_rate": 1.8017857708534107e-05,
145
+ "loss": 2.0435,
146
+ "step": 23
147
+ },
148
+ {
149
+ "epoch": 0.23,
150
+ "learning_rate": 1.783187983060594e-05,
151
+ "loss": 2.0279,
152
+ "step": 24
153
+ },
154
+ {
155
+ "epoch": 0.24,
156
+ "learning_rate": 1.763861648153945e-05,
157
+ "loss": 2.0068,
158
+ "step": 25
159
+ },
160
+ {
161
+ "epoch": 0.25,
162
+ "learning_rate": 1.743824744123196e-05,
163
+ "loss": 2.0045,
164
+ "step": 26
165
+ },
166
+ {
167
+ "epoch": 0.26,
168
+ "learning_rate": 1.7230959099527512e-05,
169
+ "loss": 1.9743,
170
+ "step": 27
171
+ },
172
+ {
173
+ "epoch": 0.27,
174
+ "learning_rate": 1.7016944282830935e-05,
175
+ "loss": 2.0074,
176
+ "step": 28
177
+ },
178
+ {
179
+ "epoch": 0.28,
180
+ "learning_rate": 1.6796402074734404e-05,
181
+ "loss": 2.0146,
182
+ "step": 29
183
+ },
184
+ {
185
+ "epoch": 0.29,
186
+ "learning_rate": 1.6569537630823385e-05,
187
+ "loss": 2.0052,
188
+ "step": 30
189
+ },
190
+ {
191
+ "epoch": 0.3,
192
+ "learning_rate": 1.6336561987834155e-05,
193
+ "loss": 2.0006,
194
+ "step": 31
195
+ },
196
+ {
197
+ "epoch": 0.3,
198
+ "learning_rate": 1.6097691867340547e-05,
199
+ "loss": 1.986,
200
+ "step": 32
201
+ },
202
+ {
203
+ "epoch": 0.31,
204
+ "learning_rate": 1.585314947415242e-05,
205
+ "loss": 1.9663,
206
+ "step": 33
207
+ },
208
+ {
209
+ "epoch": 0.32,
210
+ "learning_rate": 1.5603162289613503e-05,
211
+ "loss": 1.9941,
212
+ "step": 34
213
+ },
214
+ {
215
+ "epoch": 0.33,
216
+ "learning_rate": 1.5347962859990744e-05,
217
+ "loss": 1.9706,
218
+ "step": 35
219
+ },
220
+ {
221
+ "epoch": 0.34,
222
+ "learning_rate": 1.5087788580152207e-05,
223
+ "loss": 1.9814,
224
+ "step": 36
225
+ },
226
+ {
227
+ "epoch": 0.35,
228
+ "learning_rate": 1.4822881472734563e-05,
229
+ "loss": 1.9713,
230
+ "step": 37
231
+ },
232
+ {
233
+ "epoch": 0.36,
234
+ "learning_rate": 1.4553487963005712e-05,
235
+ "loss": 1.9883,
236
+ "step": 38
237
+ },
238
+ {
239
+ "epoch": 0.37,
240
+ "learning_rate": 1.427985864963193e-05,
241
+ "loss": 1.9605,
242
+ "step": 39
243
+ },
244
+ {
245
+ "epoch": 0.38,
246
+ "learning_rate": 1.400224807156278e-05,
247
+ "loss": 1.9615,
248
+ "step": 40
249
+ },
250
+ {
251
+ "epoch": 0.39,
252
+ "learning_rate": 1.3720914471250644e-05,
253
+ "loss": 1.9548,
254
+ "step": 41
255
+ },
256
+ {
257
+ "epoch": 0.4,
258
+ "learning_rate": 1.3436119554425133e-05,
259
+ "loss": 1.9549,
260
+ "step": 42
261
+ },
262
+ {
263
+ "epoch": 0.41,
264
+ "learning_rate": 1.314812824664585e-05,
265
+ "loss": 1.9659,
266
+ "step": 43
267
+ },
268
+ {
269
+ "epoch": 0.42,
270
+ "learning_rate": 1.285720844685996e-05,
271
+ "loss": 1.9628,
272
+ "step": 44
273
+ },
274
+ {
275
+ "epoch": 0.43,
276
+ "learning_rate": 1.2563630778193805e-05,
277
+ "loss": 1.9431,
278
+ "step": 45
279
+ },
280
+ {
281
+ "epoch": 0.44,
282
+ "learning_rate": 1.2267668336210411e-05,
283
+ "loss": 1.9727,
284
+ "step": 46
285
+ },
286
+ {
287
+ "epoch": 0.45,
288
+ "learning_rate": 1.1969596434867063e-05,
289
+ "loss": 1.9604,
290
+ "step": 47
291
+ },
292
+ {
293
+ "epoch": 0.46,
294
+ "learning_rate": 1.1669692350409223e-05,
295
+ "loss": 1.939,
296
+ "step": 48
297
+ },
298
+ {
299
+ "epoch": 0.47,
300
+ "learning_rate": 1.1368235063439103e-05,
301
+ "loss": 1.9414,
302
+ "step": 49
303
+ },
304
+ {
305
+ "epoch": 0.48,
306
+ "learning_rate": 1.1065504999398762e-05,
307
+ "loss": 1.9298,
308
+ "step": 50
309
+ },
310
+ {
311
+ "epoch": 0.49,
312
+ "learning_rate": 1.0761783767709182e-05,
313
+ "loss": 1.9248,
314
+ "step": 51
315
+ },
316
+ {
317
+ "epoch": 0.5,
318
+ "learning_rate": 1.0457353899807947e-05,
319
+ "loss": 1.9556,
320
+ "step": 52
321
+ },
322
+ {
323
+ "epoch": 0.5,
324
+ "learning_rate": 1.015249858632926e-05,
325
+ "loss": 1.9553,
326
+ "step": 53
327
+ },
328
+ {
329
+ "epoch": 0.51,
330
+ "learning_rate": 9.847501413670742e-06,
331
+ "loss": 1.9302,
332
+ "step": 54
333
+ },
334
+ {
335
+ "epoch": 0.52,
336
+ "learning_rate": 9.542646100192056e-06,
337
+ "loss": 1.9398,
338
+ "step": 55
339
+ },
340
+ {
341
+ "epoch": 0.53,
342
+ "learning_rate": 9.238216232290821e-06,
343
+ "loss": 1.9346,
344
+ "step": 56
345
+ },
346
+ {
347
+ "epoch": 0.54,
348
+ "learning_rate": 8.934495000601241e-06,
349
+ "loss": 1.9306,
350
+ "step": 57
351
+ },
352
+ {
353
+ "epoch": 0.55,
354
+ "learning_rate": 8.6317649365609e-06,
355
+ "loss": 1.9383,
356
+ "step": 58
357
+ },
358
+ {
359
+ "epoch": 0.56,
360
+ "learning_rate": 8.330307649590782e-06,
361
+ "loss": 1.9141,
362
+ "step": 59
363
+ },
364
+ {
365
+ "epoch": 0.57,
366
+ "learning_rate": 8.030403565132942e-06,
367
+ "loss": 1.9167,
368
+ "step": 60
369
+ },
370
+ {
371
+ "epoch": 0.58,
372
+ "learning_rate": 7.732331663789592e-06,
373
+ "loss": 1.933,
374
+ "step": 61
375
+ },
376
+ {
377
+ "epoch": 0.59,
378
+ "learning_rate": 7.436369221806201e-06,
379
+ "loss": 1.9143,
380
+ "step": 62
381
+ },
382
+ {
383
+ "epoch": 0.6,
384
+ "learning_rate": 7.142791553140045e-06,
385
+ "loss": 1.9258,
386
+ "step": 63
387
+ },
388
+ {
389
+ "epoch": 0.61,
390
+ "learning_rate": 6.851871753354154e-06,
391
+ "loss": 1.9299,
392
+ "step": 64
393
+ },
394
+ {
395
+ "epoch": 0.62,
396
+ "learning_rate": 6.563880445574873e-06,
397
+ "loss": 1.9293,
398
+ "step": 65
399
+ },
400
+ {
401
+ "epoch": 0.63,
402
+ "learning_rate": 6.2790855287493605e-06,
403
+ "loss": 1.9081,
404
+ "step": 66
405
+ },
406
+ {
407
+ "epoch": 0.64,
408
+ "learning_rate": 5.99775192843722e-06,
409
+ "loss": 1.915,
410
+ "step": 67
411
+ },
412
+ {
413
+ "epoch": 0.65,
414
+ "learning_rate": 5.720141350368072e-06,
415
+ "loss": 1.9079,
416
+ "step": 68
417
+ },
418
+ {
419
+ "epoch": 0.66,
420
+ "learning_rate": 5.446512036994287e-06,
421
+ "loss": 1.9374,
422
+ "step": 69
423
+ },
424
+ {
425
+ "epoch": 0.67,
426
+ "learning_rate": 5.177118527265438e-06,
427
+ "loss": 1.91,
428
+ "step": 70
429
+ },
430
+ {
431
+ "epoch": 0.68,
432
+ "learning_rate": 4.912211419847795e-06,
433
+ "loss": 1.9288,
434
+ "step": 71
435
+ },
436
+ {
437
+ "epoch": 0.69,
438
+ "learning_rate": 4.652037140009259e-06,
439
+ "loss": 1.9157,
440
+ "step": 72
441
+ },
442
+ {
443
+ "epoch": 0.7,
444
+ "learning_rate": 4.396837710386503e-06,
445
+ "loss": 1.9313,
446
+ "step": 73
447
+ },
448
+ {
449
+ "epoch": 0.7,
450
+ "learning_rate": 4.1468505258475785e-06,
451
+ "loss": 1.9036,
452
+ "step": 74
453
+ },
454
+ {
455
+ "epoch": 0.71,
456
+ "learning_rate": 3.902308132659457e-06,
457
+ "loss": 1.9472,
458
+ "step": 75
459
+ },
460
+ {
461
+ "epoch": 0.72,
462
+ "learning_rate": 3.6634380121658484e-06,
463
+ "loss": 1.9138,
464
+ "step": 76
465
+ },
466
+ {
467
+ "epoch": 0.73,
468
+ "learning_rate": 3.4304623691766193e-06,
469
+ "loss": 1.939,
470
+ "step": 77
471
+ },
472
+ {
473
+ "epoch": 0.74,
474
+ "learning_rate": 3.203597925265598e-06,
475
+ "loss": 1.9081,
476
+ "step": 78
477
+ },
478
+ {
479
+ "epoch": 0.75,
480
+ "learning_rate": 2.98305571716907e-06,
481
+ "loss": 1.8986,
482
+ "step": 79
483
+ },
484
+ {
485
+ "epoch": 0.76,
486
+ "learning_rate": 2.7690409004724883e-06,
487
+ "loss": 1.893,
488
+ "step": 80
489
+ },
490
+ {
491
+ "epoch": 0.77,
492
+ "learning_rate": 2.56175255876804e-06,
493
+ "loss": 1.8973,
494
+ "step": 81
495
+ },
496
+ {
497
+ "epoch": 0.78,
498
+ "learning_rate": 2.3613835184605527e-06,
499
+ "loss": 1.9217,
500
+ "step": 82
501
+ },
502
+ {
503
+ "epoch": 0.79,
504
+ "learning_rate": 2.1681201693940667e-06,
505
+ "loss": 1.9373,
506
+ "step": 83
507
+ },
508
+ {
509
+ "epoch": 0.8,
510
+ "learning_rate": 1.982142291465896e-06,
511
+ "loss": 1.8995,
512
+ "step": 84
513
+ },
514
+ {
515
+ "epoch": 0.81,
516
+ "learning_rate": 1.8036228873894745e-06,
517
+ "loss": 1.9153,
518
+ "step": 85
519
+ },
520
+ {
521
+ "epoch": 0.82,
522
+ "learning_rate": 1.6327280217615793e-06,
523
+ "loss": 1.8962,
524
+ "step": 86
525
+ },
526
+ {
527
+ "epoch": 0.83,
528
+ "learning_rate": 1.4696166665835853e-06,
529
+ "loss": 1.9066,
530
+ "step": 87
531
+ },
532
+ {
533
+ "epoch": 0.84,
534
+ "learning_rate": 1.3144405533805138e-06,
535
+ "loss": 1.9025,
536
+ "step": 88
537
+ },
538
+ {
539
+ "epoch": 0.85,
540
+ "learning_rate": 1.1673440320553941e-06,
541
+ "loss": 1.893,
542
+ "step": 89
543
+ },
544
+ {
545
+ "epoch": 0.86,
546
+ "learning_rate": 1.02846393661026e-06,
547
+ "loss": 1.9014,
548
+ "step": 90
549
+ },
550
+ {
551
+ "epoch": 0.87,
552
+ "learning_rate": 8.979294578586739e-07,
553
+ "loss": 1.8953,
554
+ "step": 91
555
+ },
556
+ {
557
+ "epoch": 0.88,
558
+ "learning_rate": 7.758620232482083e-07,
559
+ "loss": 1.9089,
560
+ "step": 92
561
+ },
562
+ {
563
+ "epoch": 0.89,
564
+ "learning_rate": 6.623751839046455e-07,
565
+ "loss": 1.8885,
566
+ "step": 93
567
+ },
568
+ {
569
+ "epoch": 0.9,
570
+ "learning_rate": 5.575745090030138e-07,
571
+ "loss": 1.909,
572
+ "step": 94
573
+ },
574
+ {
575
+ "epoch": 0.9,
576
+ "learning_rate": 4.61557487563673e-07,
577
+ "loss": 1.9084,
578
+ "step": 95
579
+ },
580
+ {
581
+ "epoch": 0.91,
582
+ "learning_rate": 3.7441343776484116e-07,
583
+ "loss": 1.9123,
584
+ "step": 96
585
+ },
586
+ {
587
+ "epoch": 0.92,
588
+ "learning_rate": 2.9622342385589256e-07,
589
+ "loss": 1.9122,
590
+ "step": 97
591
+ },
592
+ {
593
+ "epoch": 0.93,
594
+ "learning_rate": 2.2706018074875046e-07,
595
+ "loss": 1.9041,
596
+ "step": 98
597
+ },
598
+ {
599
+ "epoch": 0.94,
600
+ "learning_rate": 1.669880463574758e-07,
601
+ "loss": 1.8988,
602
+ "step": 99
603
+ },
604
+ {
605
+ "epoch": 0.95,
606
+ "learning_rate": 1.160629017490389e-07,
607
+ "loss": 1.9071,
608
+ "step": 100
609
+ },
610
+ {
611
+ "epoch": 0.96,
612
+ "learning_rate": 7.433211916092143e-08,
613
+ "loss": 1.9057,
614
+ "step": 101
615
+ },
616
+ {
617
+ "epoch": 0.97,
618
+ "learning_rate": 4.183451793390747e-08,
619
+ "loss": 1.9146,
620
+ "step": 102
621
+ },
622
+ {
623
+ "epoch": 0.98,
624
+ "learning_rate": 1.860032840106163e-08,
625
+ "loss": 1.9051,
626
+ "step": 103
627
+ },
628
+ {
629
+ "epoch": 0.99,
630
+ "learning_rate": 4.651163766484779e-09,
631
+ "loss": 1.9185,
632
+ "step": 104
633
+ },
634
+ {
635
+ "epoch": 1.0,
636
+ "learning_rate": 0.0,
637
+ "loss": 1.8839,
638
+ "step": 105
639
+ },
640
+ {
641
+ "epoch": 1.0,
642
+ "step": 105,
643
+ "total_flos": 6.475878851632169e+17,
644
+ "train_loss": 2.0050100269771756,
645
+ "train_runtime": 1593.9957,
646
+ "train_samples_per_second": 62.742,
647
+ "train_steps_per_second": 0.066
648
+ }
649
+ ],
650
+ "max_steps": 105,
651
+ "num_train_epochs": 1,
652
+ "total_flos": 6.475878851632169e+17,
653
+ "trial_name": null,
654
+ "trial_params": null
655
+ }
videopanda_fitu/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b1868e0ab49445abd12c4c99f0d2d4a6fb3c371752ac59dc3193cbadb9aa49a8
3
+ size 6200
videopanda_prtr1/config.json ADDED
@@ -0,0 +1,46 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "checkpoints/Video-Panda-7B/veve_prtr0",
3
+ "architectures": ["VideoPandaLlamaForCausalLM"],
4
+ "bos_token_id": 1,
5
+ "eos_token_id": 2,
6
+ "hidden_act": "silu",
7
+ "hidden_size": 4096,
8
+ "image_aspect_ratio": "square",
9
+ "image_grid_pinpoints": null,
10
+ "initializer_range": 0.02,
11
+ "intermediate_size": 11008,
12
+ "max_position_embeddings": 4096,
13
+ "mm_hidden_size": 1024,
14
+ "mm_projector_lr": null,
15
+ "mm_projector_type": "mlp2x_gelu",
16
+ "mm_use_im_patch_token": false,
17
+ "mm_use_im_start_end": false,
18
+ "mm_video_tower": "videopanda/config/videopanda-video-LB-patch14-anypixel-448",
19
+ "mm_video_tower_teacher": "LanguageBind/LanguageBind_Video_merge",
20
+ "mm_vision_select_feature": "patch",
21
+ "mm_vision_select_layer": -2,
22
+ "mm_vision_tower": "videopanda/config/videopanda-vision-LB-patch14-anypixel-448",
23
+ "mm_vision_tower_teacher": "LanguageBind/LanguageBind_Image",
24
+ "model_type": "videopanda",
25
+ "num_attention_heads": 32,
26
+ "num_hidden_layers": 32,
27
+ "num_key_value_heads": 32,
28
+ "pad_token_id": 0,
29
+ "pretraining_tp": 1,
30
+ "requires_video_distill": true,
31
+ "requires_image_distill": false,
32
+ "rms_norm_eps": 1e-05,
33
+ "rope_scaling": null,
34
+ "tie_word_embeddings": false,
35
+ "torch_dtype": "bfloat16",
36
+ "transformers_version": "4.31.0",
37
+ "tune_mm_mlp_adapter": false,
38
+ "use_cache": true,
39
+ "use_mm_proj": true,
40
+ "video_hidden_size": 1024,
41
+ "video_projector_lr": null,
42
+ "video_tower_lr": null,
43
+ "video_tower_teacher": "LanguageBind/LanguageBind_Video_merge",
44
+ "vision_tower_lr": null,
45
+ "vocab_size": 32000
46
+ }
videopanda_prtr1/generation_config.json ADDED
@@ -0,0 +1,9 @@
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "bos_token_id": 1,
3
+ "eos_token_id": 2,
4
+ "max_length": 4096,
5
+ "pad_token_id": 0,
6
+ "temperature": 0.9,
7
+ "top_p": 0.6,
8
+ "transformers_version": "4.31.0"
9
+ }
videopanda_prtr1/pytorch_model-00001-of-00002.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:2b32082b68e63a90f4635ff1be5272d2290cc665a91ccb7acffc36d8ee5de69e
3
+ size 9976631973
videopanda_prtr1/pytorch_model-00002-of-00002.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d85923a8e0f595ae07df022d1ef0ad225f9794327703ffc356a6416880326790
3
+ size 6688819122
videopanda_prtr1/pytorch_model.bin.index.json ADDED
The diff for this file is too large to render. See raw diff
 
videopanda_prtr1/special_tokens_map.json ADDED
@@ -0,0 +1,24 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "bos_token": {
3
+ "content": "<s>",
4
+ "lstrip": false,
5
+ "normalized": false,
6
+ "rstrip": false,
7
+ "single_word": false
8
+ },
9
+ "eos_token": {
10
+ "content": "</s>",
11
+ "lstrip": false,
12
+ "normalized": false,
13
+ "rstrip": false,
14
+ "single_word": false
15
+ },
16
+ "pad_token": "<unk>",
17
+ "unk_token": {
18
+ "content": "<unk>",
19
+ "lstrip": false,
20
+ "normalized": false,
21
+ "rstrip": false,
22
+ "single_word": false
23
+ }
24
+ }
videopanda_prtr1/tokenizer.model ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9e556afd44213b6bd1be2b850ebbbd98f5481437a8021afaf58ee7fb1818d347
3
+ size 499723
videopanda_prtr1/tokenizer_config.json ADDED
@@ -0,0 +1,35 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "add_bos_token": true,
3
+ "add_eos_token": false,
4
+ "bos_token": {
5
+ "__type": "AddedToken",
6
+ "content": "<s>",
7
+ "lstrip": false,
8
+ "normalized": false,
9
+ "rstrip": false,
10
+ "single_word": false
11
+ },
12
+ "clean_up_tokenization_spaces": false,
13
+ "eos_token": {
14
+ "__type": "AddedToken",
15
+ "content": "</s>",
16
+ "lstrip": false,
17
+ "normalized": false,
18
+ "rstrip": false,
19
+ "single_word": false
20
+ },
21
+ "legacy": false,
22
+ "model_max_length": 4096,
23
+ "pad_token": null,
24
+ "padding_side": "right",
25
+ "sp_model_kwargs": {},
26
+ "tokenizer_class": "LlamaTokenizer",
27
+ "unk_token": {
28
+ "__type": "AddedToken",
29
+ "content": "<unk>",
30
+ "lstrip": false,
31
+ "normalized": false,
32
+ "rstrip": false,
33
+ "single_word": false
34
+ }
35
+ }
videopanda_prtr1/trainer_state.json ADDED
@@ -0,0 +1,2227 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": null,
3
+ "best_model_checkpoint": null,
4
+ "epoch": 1.0,
5
+ "global_step": 367,
6
+ "is_hyper_param_search": false,
7
+ "is_local_process_zero": true,
8
+ "is_world_process_zero": true,
9
+ "log_history": [
10
+ {
11
+ "epoch": 0.0,
12
+ "learning_rate": 1e-05,
13
+ "loss": 3.2173,
14
+ "step": 1
15
+ },
16
+ {
17
+ "epoch": 0.01,
18
+ "learning_rate": 2e-05,
19
+ "loss": 3.2031,
20
+ "step": 2
21
+ },
22
+ {
23
+ "epoch": 0.01,
24
+ "learning_rate": 3.0000000000000004e-05,
25
+ "loss": 3.1896,
26
+ "step": 3
27
+ },
28
+ {
29
+ "epoch": 0.01,
30
+ "learning_rate": 4e-05,
31
+ "loss": 3.5399,
32
+ "step": 4
33
+ },
34
+ {
35
+ "epoch": 0.01,
36
+ "learning_rate": 3.999925099660787e-05,
37
+ "loss": 3.249,
38
+ "step": 5
39
+ },
40
+ {
41
+ "epoch": 0.02,
42
+ "learning_rate": 3.999700404253208e-05,
43
+ "loss": 4.346,
44
+ "step": 6
45
+ },
46
+ {
47
+ "epoch": 0.02,
48
+ "learning_rate": 3.9993259306070256e-05,
49
+ "loss": 4.1291,
50
+ "step": 7
51
+ },
52
+ {
53
+ "epoch": 0.02,
54
+ "learning_rate": 3.998801706770442e-05,
55
+ "loss": 3.7064,
56
+ "step": 8
57
+ },
58
+ {
59
+ "epoch": 0.02,
60
+ "learning_rate": 3.9981277720080015e-05,
61
+ "loss": 3.5643,
62
+ "step": 9
63
+ },
64
+ {
65
+ "epoch": 0.03,
66
+ "learning_rate": 3.9973041767976466e-05,
67
+ "loss": 3.5031,
68
+ "step": 10
69
+ },
70
+ {
71
+ "epoch": 0.03,
72
+ "learning_rate": 3.996330982826937e-05,
73
+ "loss": 3.4195,
74
+ "step": 11
75
+ },
76
+ {
77
+ "epoch": 0.03,
78
+ "learning_rate": 3.995208262988431e-05,
79
+ "loss": 3.3598,
80
+ "step": 12
81
+ },
82
+ {
83
+ "epoch": 0.04,
84
+ "learning_rate": 3.9939361013742275e-05,
85
+ "loss": 3.3177,
86
+ "step": 13
87
+ },
88
+ {
89
+ "epoch": 0.04,
90
+ "learning_rate": 3.99251459326966e-05,
91
+ "loss": 3.2443,
92
+ "step": 14
93
+ },
94
+ {
95
+ "epoch": 0.04,
96
+ "learning_rate": 3.9909438451461695e-05,
97
+ "loss": 3.2215,
98
+ "step": 15
99
+ },
100
+ {
101
+ "epoch": 0.04,
102
+ "learning_rate": 3.989223974653323e-05,
103
+ "loss": 3.149,
104
+ "step": 16
105
+ },
106
+ {
107
+ "epoch": 0.05,
108
+ "learning_rate": 3.9873551106100035e-05,
109
+ "loss": 3.1167,
110
+ "step": 17
111
+ },
112
+ {
113
+ "epoch": 0.05,
114
+ "learning_rate": 3.985337392994763e-05,
115
+ "loss": 3.0538,
116
+ "step": 18
117
+ },
118
+ {
119
+ "epoch": 0.05,
120
+ "learning_rate": 3.983170972935333e-05,
121
+ "loss": 3.029,
122
+ "step": 19
123
+ },
124
+ {
125
+ "epoch": 0.05,
126
+ "learning_rate": 3.9808560126973126e-05,
127
+ "loss": 3.0055,
128
+ "step": 20
129
+ },
130
+ {
131
+ "epoch": 0.06,
132
+ "learning_rate": 3.9783926856720085e-05,
133
+ "loss": 2.9598,
134
+ "step": 21
135
+ },
136
+ {
137
+ "epoch": 0.06,
138
+ "learning_rate": 3.975781176363451e-05,
139
+ "loss": 2.9572,
140
+ "step": 22
141
+ },
142
+ {
143
+ "epoch": 0.06,
144
+ "learning_rate": 3.973021680374571e-05,
145
+ "loss": 2.9004,
146
+ "step": 23
147
+ },
148
+ {
149
+ "epoch": 0.07,
150
+ "learning_rate": 3.9701144043925576e-05,
151
+ "loss": 2.8984,
152
+ "step": 24
153
+ },
154
+ {
155
+ "epoch": 0.07,
156
+ "learning_rate": 3.9670595661733654e-05,
157
+ "loss": 2.874,
158
+ "step": 25
159
+ },
160
+ {
161
+ "epoch": 0.07,
162
+ "learning_rate": 3.963857394525413e-05,
163
+ "loss": 2.8689,
164
+ "step": 26
165
+ },
166
+ {
167
+ "epoch": 0.07,
168
+ "learning_rate": 3.960508129292446e-05,
169
+ "loss": 2.8885,
170
+ "step": 27
171
+ },
172
+ {
173
+ "epoch": 0.08,
174
+ "learning_rate": 3.9570120213355636e-05,
175
+ "loss": 2.8331,
176
+ "step": 28
177
+ },
178
+ {
179
+ "epoch": 0.08,
180
+ "learning_rate": 3.953369332514438e-05,
181
+ "loss": 2.8093,
182
+ "step": 29
183
+ },
184
+ {
185
+ "epoch": 0.08,
186
+ "learning_rate": 3.949580335667699e-05,
187
+ "loss": 2.7919,
188
+ "step": 30
189
+ },
190
+ {
191
+ "epoch": 0.08,
192
+ "learning_rate": 3.945645314592495e-05,
193
+ "loss": 2.7876,
194
+ "step": 31
195
+ },
196
+ {
197
+ "epoch": 0.09,
198
+ "learning_rate": 3.9415645640232386e-05,
199
+ "loss": 2.791,
200
+ "step": 32
201
+ },
202
+ {
203
+ "epoch": 0.09,
204
+ "learning_rate": 3.937338389609533e-05,
205
+ "loss": 2.7751,
206
+ "step": 33
207
+ },
208
+ {
209
+ "epoch": 0.09,
210
+ "learning_rate": 3.932967107893274e-05,
211
+ "loss": 2.7673,
212
+ "step": 34
213
+ },
214
+ {
215
+ "epoch": 0.1,
216
+ "learning_rate": 3.928451046284946e-05,
217
+ "loss": 2.74,
218
+ "step": 35
219
+ },
220
+ {
221
+ "epoch": 0.1,
222
+ "learning_rate": 3.923790543039095e-05,
223
+ "loss": 2.7354,
224
+ "step": 36
225
+ },
226
+ {
227
+ "epoch": 0.1,
228
+ "learning_rate": 3.9189859472289956e-05,
229
+ "loss": 2.699,
230
+ "step": 37
231
+ },
232
+ {
233
+ "epoch": 0.1,
234
+ "learning_rate": 3.9140376187205025e-05,
235
+ "loss": 2.7079,
236
+ "step": 38
237
+ },
238
+ {
239
+ "epoch": 0.11,
240
+ "learning_rate": 3.9089459281451e-05,
241
+ "loss": 2.6673,
242
+ "step": 39
243
+ },
244
+ {
245
+ "epoch": 0.11,
246
+ "learning_rate": 3.903711256872139e-05,
247
+ "loss": 2.6733,
248
+ "step": 40
249
+ },
250
+ {
251
+ "epoch": 0.11,
252
+ "learning_rate": 3.898333996980275e-05,
253
+ "loss": 2.659,
254
+ "step": 41
255
+ },
256
+ {
257
+ "epoch": 0.11,
258
+ "learning_rate": 3.8928145512280973e-05,
259
+ "loss": 2.6657,
260
+ "step": 42
261
+ },
262
+ {
263
+ "epoch": 0.12,
264
+ "learning_rate": 3.8871533330239646e-05,
265
+ "loss": 2.6348,
266
+ "step": 43
267
+ },
268
+ {
269
+ "epoch": 0.12,
270
+ "learning_rate": 3.8813507663950404e-05,
271
+ "loss": 2.6497,
272
+ "step": 44
273
+ },
274
+ {
275
+ "epoch": 0.12,
276
+ "learning_rate": 3.8754072859555346e-05,
277
+ "loss": 2.5951,
278
+ "step": 45
279
+ },
280
+ {
281
+ "epoch": 0.13,
282
+ "learning_rate": 3.869323336874146e-05,
283
+ "loss": 2.6479,
284
+ "step": 46
285
+ },
286
+ {
287
+ "epoch": 0.13,
288
+ "learning_rate": 3.8630993748407274e-05,
289
+ "loss": 2.6018,
290
+ "step": 47
291
+ },
292
+ {
293
+ "epoch": 0.13,
294
+ "learning_rate": 3.856735866032145e-05,
295
+ "loss": 2.5817,
296
+ "step": 48
297
+ },
298
+ {
299
+ "epoch": 0.13,
300
+ "learning_rate": 3.8502332870773675e-05,
301
+ "loss": 2.5903,
302
+ "step": 49
303
+ },
304
+ {
305
+ "epoch": 0.14,
306
+ "learning_rate": 3.843592125021764e-05,
307
+ "loss": 2.5757,
308
+ "step": 50
309
+ },
310
+ {
311
+ "epoch": 0.14,
312
+ "learning_rate": 3.8368128772906254e-05,
313
+ "loss": 2.5705,
314
+ "step": 51
315
+ },
316
+ {
317
+ "epoch": 0.14,
318
+ "learning_rate": 3.829896051651907e-05,
319
+ "loss": 2.5914,
320
+ "step": 52
321
+ },
322
+ {
323
+ "epoch": 0.14,
324
+ "learning_rate": 3.822842166178194e-05,
325
+ "loss": 2.5694,
326
+ "step": 53
327
+ },
328
+ {
329
+ "epoch": 0.15,
330
+ "learning_rate": 3.815651749207902e-05,
331
+ "loss": 2.5826,
332
+ "step": 54
333
+ },
334
+ {
335
+ "epoch": 0.15,
336
+ "learning_rate": 3.8083253393057006e-05,
337
+ "loss": 2.5435,
338
+ "step": 55
339
+ },
340
+ {
341
+ "epoch": 0.15,
342
+ "learning_rate": 3.8008634852221777e-05,
343
+ "loss": 2.5286,
344
+ "step": 56
345
+ },
346
+ {
347
+ "epoch": 0.16,
348
+ "learning_rate": 3.793266745852735e-05,
349
+ "loss": 2.5152,
350
+ "step": 57
351
+ },
352
+ {
353
+ "epoch": 0.16,
354
+ "learning_rate": 3.785535690195728e-05,
355
+ "loss": 2.4879,
356
+ "step": 58
357
+ },
358
+ {
359
+ "epoch": 0.16,
360
+ "learning_rate": 3.7776708973098476e-05,
361
+ "loss": 2.5058,
362
+ "step": 59
363
+ },
364
+ {
365
+ "epoch": 0.16,
366
+ "learning_rate": 3.769672956270749e-05,
367
+ "loss": 2.5437,
368
+ "step": 60
369
+ },
370
+ {
371
+ "epoch": 0.17,
372
+ "learning_rate": 3.761542466126929e-05,
373
+ "loss": 2.4668,
374
+ "step": 61
375
+ },
376
+ {
377
+ "epoch": 0.17,
378
+ "learning_rate": 3.753280035854857e-05,
379
+ "loss": 2.501,
380
+ "step": 62
381
+ },
382
+ {
383
+ "epoch": 0.17,
384
+ "learning_rate": 3.7448862843133644e-05,
385
+ "loss": 2.4697,
386
+ "step": 63
387
+ },
388
+ {
389
+ "epoch": 0.17,
390
+ "learning_rate": 3.736361840197288e-05,
391
+ "loss": 2.4651,
392
+ "step": 64
393
+ },
394
+ {
395
+ "epoch": 0.18,
396
+ "learning_rate": 3.727707341990383e-05,
397
+ "loss": 2.4394,
398
+ "step": 65
399
+ },
400
+ {
401
+ "epoch": 0.18,
402
+ "learning_rate": 3.718923437917503e-05,
403
+ "loss": 2.4239,
404
+ "step": 66
405
+ },
406
+ {
407
+ "epoch": 0.18,
408
+ "learning_rate": 3.7100107858960404e-05,
409
+ "loss": 2.4541,
410
+ "step": 67
411
+ },
412
+ {
413
+ "epoch": 0.19,
414
+ "learning_rate": 3.7009700534866557e-05,
415
+ "loss": 2.445,
416
+ "step": 68
417
+ },
418
+ {
419
+ "epoch": 0.19,
420
+ "learning_rate": 3.691801917843273e-05,
421
+ "loss": 2.422,
422
+ "step": 69
423
+ },
424
+ {
425
+ "epoch": 0.19,
426
+ "learning_rate": 3.6825070656623626e-05,
427
+ "loss": 2.4274,
428
+ "step": 70
429
+ },
430
+ {
431
+ "epoch": 0.19,
432
+ "learning_rate": 3.6730861931315054e-05,
433
+ "loss": 2.4062,
434
+ "step": 71
435
+ },
436
+ {
437
+ "epoch": 0.2,
438
+ "learning_rate": 3.663540005877249e-05,
439
+ "loss": 2.4,
440
+ "step": 72
441
+ },
442
+ {
443
+ "epoch": 0.2,
444
+ "learning_rate": 3.653869218912258e-05,
445
+ "loss": 2.4229,
446
+ "step": 73
447
+ },
448
+ {
449
+ "epoch": 0.2,
450
+ "learning_rate": 3.6440745565817556e-05,
451
+ "loss": 2.3909,
452
+ "step": 74
453
+ },
454
+ {
455
+ "epoch": 0.2,
456
+ "learning_rate": 3.6341567525092727e-05,
457
+ "loss": 2.3607,
458
+ "step": 75
459
+ },
460
+ {
461
+ "epoch": 0.21,
462
+ "learning_rate": 3.6241165495417006e-05,
463
+ "loss": 2.399,
464
+ "step": 76
465
+ },
466
+ {
467
+ "epoch": 0.21,
468
+ "learning_rate": 3.613954699693645e-05,
469
+ "loss": 2.3902,
470
+ "step": 77
471
+ },
472
+ {
473
+ "epoch": 0.21,
474
+ "learning_rate": 3.603671964091107e-05,
475
+ "loss": 2.3692,
476
+ "step": 78
477
+ },
478
+ {
479
+ "epoch": 0.22,
480
+ "learning_rate": 3.593269112914472e-05,
481
+ "loss": 2.3436,
482
+ "step": 79
483
+ },
484
+ {
485
+ "epoch": 0.22,
486
+ "learning_rate": 3.582746925340822e-05,
487
+ "loss": 2.3629,
488
+ "step": 80
489
+ },
490
+ {
491
+ "epoch": 0.22,
492
+ "learning_rate": 3.5721061894855756e-05,
493
+ "loss": 2.3287,
494
+ "step": 81
495
+ },
496
+ {
497
+ "epoch": 0.22,
498
+ "learning_rate": 3.561347702343456e-05,
499
+ "loss": 2.3376,
500
+ "step": 82
501
+ },
502
+ {
503
+ "epoch": 0.23,
504
+ "learning_rate": 3.5504722697288025e-05,
505
+ "loss": 2.3418,
506
+ "step": 83
507
+ },
508
+ {
509
+ "epoch": 0.23,
510
+ "learning_rate": 3.539480706215204e-05,
511
+ "loss": 2.3182,
512
+ "step": 84
513
+ },
514
+ {
515
+ "epoch": 0.23,
516
+ "learning_rate": 3.5283738350744986e-05,
517
+ "loss": 2.3287,
518
+ "step": 85
519
+ },
520
+ {
521
+ "epoch": 0.23,
522
+ "learning_rate": 3.517152488215101e-05,
523
+ "loss": 2.3222,
524
+ "step": 86
525
+ },
526
+ {
527
+ "epoch": 0.24,
528
+ "learning_rate": 3.505817506119698e-05,
529
+ "loss": 2.3393,
530
+ "step": 87
531
+ },
532
+ {
533
+ "epoch": 0.24,
534
+ "learning_rate": 3.494369737782293e-05,
535
+ "loss": 2.322,
536
+ "step": 88
537
+ },
538
+ {
539
+ "epoch": 0.24,
540
+ "learning_rate": 3.4828100406446184e-05,
541
+ "loss": 2.2907,
542
+ "step": 89
543
+ },
544
+ {
545
+ "epoch": 0.25,
546
+ "learning_rate": 3.47113928053191e-05,
547
+ "loss": 2.3132,
548
+ "step": 90
549
+ },
550
+ {
551
+ "epoch": 0.25,
552
+ "learning_rate": 3.45935833158806e-05,
553
+ "loss": 2.2972,
554
+ "step": 91
555
+ },
556
+ {
557
+ "epoch": 0.25,
558
+ "learning_rate": 3.44746807621014e-05,
559
+ "loss": 2.2643,
560
+ "step": 92
561
+ },
562
+ {
563
+ "epoch": 0.25,
564
+ "learning_rate": 3.4354694049823124e-05,
565
+ "loss": 2.2527,
566
+ "step": 93
567
+ },
568
+ {
569
+ "epoch": 0.26,
570
+ "learning_rate": 3.4233632166091205e-05,
571
+ "loss": 2.2746,
572
+ "step": 94
573
+ },
574
+ {
575
+ "epoch": 0.26,
576
+ "learning_rate": 3.4111504178481813e-05,
577
+ "loss": 2.2479,
578
+ "step": 95
579
+ },
580
+ {
581
+ "epoch": 0.26,
582
+ "learning_rate": 3.3988319234422636e-05,
583
+ "loss": 2.2954,
584
+ "step": 96
585
+ },
586
+ {
587
+ "epoch": 0.26,
588
+ "learning_rate": 3.3864086560507785e-05,
589
+ "loss": 2.2455,
590
+ "step": 97
591
+ },
592
+ {
593
+ "epoch": 0.27,
594
+ "learning_rate": 3.373881546180666e-05,
595
+ "loss": 2.2442,
596
+ "step": 98
597
+ },
598
+ {
599
+ "epoch": 0.27,
600
+ "learning_rate": 3.361251532116707e-05,
601
+ "loss": 2.2035,
602
+ "step": 99
603
+ },
604
+ {
605
+ "epoch": 0.27,
606
+ "learning_rate": 3.3485195598512365e-05,
607
+ "loss": 2.2234,
608
+ "step": 100
609
+ },
610
+ {
611
+ "epoch": 0.28,
612
+ "learning_rate": 3.3356865830132976e-05,
613
+ "loss": 2.2226,
614
+ "step": 101
615
+ },
616
+ {
617
+ "epoch": 0.28,
618
+ "learning_rate": 3.322753562797209e-05,
619
+ "loss": 2.2167,
620
+ "step": 102
621
+ },
622
+ {
623
+ "epoch": 0.28,
624
+ "learning_rate": 3.309721467890571e-05,
625
+ "loss": 2.2481,
626
+ "step": 103
627
+ },
628
+ {
629
+ "epoch": 0.28,
630
+ "learning_rate": 3.296591274401712e-05,
631
+ "loss": 2.1965,
632
+ "step": 104
633
+ },
634
+ {
635
+ "epoch": 0.29,
636
+ "learning_rate": 3.28336396578658e-05,
637
+ "loss": 2.1935,
638
+ "step": 105
639
+ },
640
+ {
641
+ "epoch": 0.29,
642
+ "learning_rate": 3.270040532775077e-05,
643
+ "loss": 2.196,
644
+ "step": 106
645
+ },
646
+ {
647
+ "epoch": 0.29,
648
+ "learning_rate": 3.256621973296854e-05,
649
+ "loss": 2.1733,
650
+ "step": 107
651
+ },
652
+ {
653
+ "epoch": 0.29,
654
+ "learning_rate": 3.243109292406568e-05,
655
+ "loss": 2.1735,
656
+ "step": 108
657
+ },
658
+ {
659
+ "epoch": 0.3,
660
+ "learning_rate": 3.229503502208602e-05,
661
+ "loss": 2.1915,
662
+ "step": 109
663
+ },
664
+ {
665
+ "epoch": 0.3,
666
+ "learning_rate": 3.215805621781256e-05,
667
+ "loss": 2.1795,
668
+ "step": 110
669
+ },
670
+ {
671
+ "epoch": 0.3,
672
+ "learning_rate": 3.202016677100422e-05,
673
+ "loss": 2.2021,
674
+ "step": 111
675
+ },
676
+ {
677
+ "epoch": 0.31,
678
+ "learning_rate": 3.188137700962733e-05,
679
+ "loss": 2.201,
680
+ "step": 112
681
+ },
682
+ {
683
+ "epoch": 0.31,
684
+ "learning_rate": 3.174169732908209e-05,
685
+ "loss": 2.1857,
686
+ "step": 113
687
+ },
688
+ {
689
+ "epoch": 0.31,
690
+ "learning_rate": 3.1601138191423966e-05,
691
+ "loss": 2.1818,
692
+ "step": 114
693
+ },
694
+ {
695
+ "epoch": 0.31,
696
+ "learning_rate": 3.145971012458005e-05,
697
+ "loss": 2.1438,
698
+ "step": 115
699
+ },
700
+ {
701
+ "epoch": 0.32,
702
+ "learning_rate": 3.13174237215605e-05,
703
+ "loss": 2.1359,
704
+ "step": 116
705
+ },
706
+ {
707
+ "epoch": 0.32,
708
+ "learning_rate": 3.11742896396652e-05,
709
+ "loss": 2.1668,
710
+ "step": 117
711
+ },
712
+ {
713
+ "epoch": 0.32,
714
+ "learning_rate": 3.103031859968542e-05,
715
+ "loss": 2.1769,
716
+ "step": 118
717
+ },
718
+ {
719
+ "epoch": 0.32,
720
+ "learning_rate": 3.0885521385100885e-05,
721
+ "loss": 2.1445,
722
+ "step": 119
723
+ },
724
+ {
725
+ "epoch": 0.33,
726
+ "learning_rate": 3.0739908841272095e-05,
727
+ "loss": 2.1193,
728
+ "step": 120
729
+ },
730
+ {
731
+ "epoch": 0.33,
732
+ "learning_rate": 3.059349187462798e-05,
733
+ "loss": 2.1285,
734
+ "step": 121
735
+ },
736
+ {
737
+ "epoch": 0.33,
738
+ "learning_rate": 3.044628145184899e-05,
739
+ "loss": 2.1411,
740
+ "step": 122
741
+ },
742
+ {
743
+ "epoch": 0.34,
744
+ "learning_rate": 3.0298288599045747e-05,
745
+ "loss": 2.1321,
746
+ "step": 123
747
+ },
748
+ {
749
+ "epoch": 0.34,
750
+ "learning_rate": 3.0149524400933114e-05,
751
+ "loss": 2.1008,
752
+ "step": 124
753
+ },
754
+ {
755
+ "epoch": 0.34,
756
+ "learning_rate": 3.0000000000000004e-05,
757
+ "loss": 2.1437,
758
+ "step": 125
759
+ },
760
+ {
761
+ "epoch": 0.34,
762
+ "learning_rate": 2.9849726595674756e-05,
763
+ "loss": 2.1224,
764
+ "step": 126
765
+ },
766
+ {
767
+ "epoch": 0.35,
768
+ "learning_rate": 2.9698715443486338e-05,
769
+ "loss": 2.083,
770
+ "step": 127
771
+ },
772
+ {
773
+ "epoch": 0.35,
774
+ "learning_rate": 2.9546977854221266e-05,
775
+ "loss": 2.1156,
776
+ "step": 128
777
+ },
778
+ {
779
+ "epoch": 0.35,
780
+ "learning_rate": 2.9394525193076454e-05,
781
+ "loss": 2.1127,
782
+ "step": 129
783
+ },
784
+ {
785
+ "epoch": 0.35,
786
+ "learning_rate": 2.9241368878807925e-05,
787
+ "loss": 2.0949,
788
+ "step": 130
789
+ },
790
+ {
791
+ "epoch": 0.36,
792
+ "learning_rate": 2.908752038287558e-05,
793
+ "loss": 2.0821,
794
+ "step": 131
795
+ },
796
+ {
797
+ "epoch": 0.36,
798
+ "learning_rate": 2.8932991228583954e-05,
799
+ "loss": 2.0735,
800
+ "step": 132
801
+ },
802
+ {
803
+ "epoch": 0.36,
804
+ "learning_rate": 2.877779299021912e-05,
805
+ "loss": 2.0996,
806
+ "step": 133
807
+ },
808
+ {
809
+ "epoch": 0.37,
810
+ "learning_rate": 2.8621937292181768e-05,
811
+ "loss": 2.0967,
812
+ "step": 134
813
+ },
814
+ {
815
+ "epoch": 0.37,
816
+ "learning_rate": 2.846543580811656e-05,
817
+ "loss": 2.0313,
818
+ "step": 135
819
+ },
820
+ {
821
+ "epoch": 0.37,
822
+ "learning_rate": 2.8308300260037734e-05,
823
+ "loss": 2.0891,
824
+ "step": 136
825
+ },
826
+ {
827
+ "epoch": 0.37,
828
+ "learning_rate": 2.8150542417451144e-05,
829
+ "loss": 2.0816,
830
+ "step": 137
831
+ },
832
+ {
833
+ "epoch": 0.38,
834
+ "learning_rate": 2.7992174096472714e-05,
835
+ "loss": 2.0765,
836
+ "step": 138
837
+ },
838
+ {
839
+ "epoch": 0.38,
840
+ "learning_rate": 2.783320715894341e-05,
841
+ "loss": 2.026,
842
+ "step": 139
843
+ },
844
+ {
845
+ "epoch": 0.38,
846
+ "learning_rate": 2.767365351154077e-05,
847
+ "loss": 2.0424,
848
+ "step": 140
849
+ },
850
+ {
851
+ "epoch": 0.38,
852
+ "learning_rate": 2.751352510488711e-05,
853
+ "loss": 2.074,
854
+ "step": 141
855
+ },
856
+ {
857
+ "epoch": 0.39,
858
+ "learning_rate": 2.7352833932654402e-05,
859
+ "loss": 2.0189,
860
+ "step": 142
861
+ },
862
+ {
863
+ "epoch": 0.39,
864
+ "learning_rate": 2.719159203066597e-05,
865
+ "loss": 2.0283,
866
+ "step": 143
867
+ },
868
+ {
869
+ "epoch": 0.39,
870
+ "learning_rate": 2.702981147599495e-05,
871
+ "loss": 2.0373,
872
+ "step": 144
873
+ },
874
+ {
875
+ "epoch": 0.4,
876
+ "learning_rate": 2.6867504386059776e-05,
877
+ "loss": 2.0141,
878
+ "step": 145
879
+ },
880
+ {
881
+ "epoch": 0.4,
882
+ "learning_rate": 2.6704682917716528e-05,
883
+ "loss": 2.0197,
884
+ "step": 146
885
+ },
886
+ {
887
+ "epoch": 0.4,
888
+ "learning_rate": 2.6541359266348437e-05,
889
+ "loss": 2.0168,
890
+ "step": 147
891
+ },
892
+ {
893
+ "epoch": 0.4,
894
+ "learning_rate": 2.637754566495238e-05,
895
+ "loss": 2.0032,
896
+ "step": 148
897
+ },
898
+ {
899
+ "epoch": 0.41,
900
+ "learning_rate": 2.6213254383222665e-05,
901
+ "loss": 2.0038,
902
+ "step": 149
903
+ },
904
+ {
905
+ "epoch": 0.41,
906
+ "learning_rate": 2.6048497726632023e-05,
907
+ "loss": 1.9901,
908
+ "step": 150
909
+ },
910
+ {
911
+ "epoch": 0.41,
912
+ "learning_rate": 2.588328803550993e-05,
913
+ "loss": 1.9917,
914
+ "step": 151
915
+ },
916
+ {
917
+ "epoch": 0.41,
918
+ "learning_rate": 2.571763768411829e-05,
919
+ "loss": 1.9718,
920
+ "step": 152
921
+ },
922
+ {
923
+ "epoch": 0.42,
924
+ "learning_rate": 2.555155907972461e-05,
925
+ "loss": 2.0155,
926
+ "step": 153
927
+ },
928
+ {
929
+ "epoch": 0.42,
930
+ "learning_rate": 2.5385064661672692e-05,
931
+ "loss": 2.0072,
932
+ "step": 154
933
+ },
934
+ {
935
+ "epoch": 0.42,
936
+ "learning_rate": 2.5218166900450937e-05,
937
+ "loss": 1.9935,
938
+ "step": 155
939
+ },
940
+ {
941
+ "epoch": 0.43,
942
+ "learning_rate": 2.5050878296758255e-05,
943
+ "loss": 1.9879,
944
+ "step": 156
945
+ },
946
+ {
947
+ "epoch": 0.43,
948
+ "learning_rate": 2.488321138056783e-05,
949
+ "loss": 2.0028,
950
+ "step": 157
951
+ },
952
+ {
953
+ "epoch": 0.43,
954
+ "learning_rate": 2.471517871018855e-05,
955
+ "loss": 1.9669,
956
+ "step": 158
957
+ },
958
+ {
959
+ "epoch": 0.43,
960
+ "learning_rate": 2.4546792871324424e-05,
961
+ "loss": 1.9854,
962
+ "step": 159
963
+ },
964
+ {
965
+ "epoch": 0.44,
966
+ "learning_rate": 2.43780664761319e-05,
967
+ "loss": 1.9613,
968
+ "step": 160
969
+ },
970
+ {
971
+ "epoch": 0.44,
972
+ "learning_rate": 2.4209012162275217e-05,
973
+ "loss": 1.9474,
974
+ "step": 161
975
+ },
976
+ {
977
+ "epoch": 0.44,
978
+ "learning_rate": 2.4039642591979825e-05,
979
+ "loss": 1.9424,
980
+ "step": 162
981
+ },
982
+ {
983
+ "epoch": 0.44,
984
+ "learning_rate": 2.3869970451083996e-05,
985
+ "loss": 1.9656,
986
+ "step": 163
987
+ },
988
+ {
989
+ "epoch": 0.45,
990
+ "learning_rate": 2.370000844808863e-05,
991
+ "loss": 1.9686,
992
+ "step": 164
993
+ },
994
+ {
995
+ "epoch": 0.45,
996
+ "learning_rate": 2.3529769313205423e-05,
997
+ "loss": 1.9655,
998
+ "step": 165
999
+ },
1000
+ {
1001
+ "epoch": 0.45,
1002
+ "learning_rate": 2.3359265797403297e-05,
1003
+ "loss": 1.97,
1004
+ "step": 166
1005
+ },
1006
+ {
1007
+ "epoch": 0.46,
1008
+ "learning_rate": 2.318851067145345e-05,
1009
+ "loss": 1.9435,
1010
+ "step": 167
1011
+ },
1012
+ {
1013
+ "epoch": 0.46,
1014
+ "learning_rate": 2.3017516724972716e-05,
1015
+ "loss": 1.972,
1016
+ "step": 168
1017
+ },
1018
+ {
1019
+ "epoch": 0.46,
1020
+ "learning_rate": 2.2846296765465708e-05,
1021
+ "loss": 1.9789,
1022
+ "step": 169
1023
+ },
1024
+ {
1025
+ "epoch": 0.46,
1026
+ "learning_rate": 2.267486361736546e-05,
1027
+ "loss": 1.9466,
1028
+ "step": 170
1029
+ },
1030
+ {
1031
+ "epoch": 0.47,
1032
+ "learning_rate": 2.250323012107292e-05,
1033
+ "loss": 1.9231,
1034
+ "step": 171
1035
+ },
1036
+ {
1037
+ "epoch": 0.47,
1038
+ "learning_rate": 2.2331409131995186e-05,
1039
+ "loss": 1.9672,
1040
+ "step": 172
1041
+ },
1042
+ {
1043
+ "epoch": 0.47,
1044
+ "learning_rate": 2.2159413519582623e-05,
1045
+ "loss": 1.9265,
1046
+ "step": 173
1047
+ },
1048
+ {
1049
+ "epoch": 0.47,
1050
+ "learning_rate": 2.1987256166364937e-05,
1051
+ "loss": 1.911,
1052
+ "step": 174
1053
+ },
1054
+ {
1055
+ "epoch": 0.48,
1056
+ "learning_rate": 2.1814949966986288e-05,
1057
+ "loss": 1.9236,
1058
+ "step": 175
1059
+ },
1060
+ {
1061
+ "epoch": 0.48,
1062
+ "learning_rate": 2.1642507827239455e-05,
1063
+ "loss": 1.9543,
1064
+ "step": 176
1065
+ },
1066
+ {
1067
+ "epoch": 0.48,
1068
+ "learning_rate": 2.1469942663099208e-05,
1069
+ "loss": 1.9296,
1070
+ "step": 177
1071
+ },
1072
+ {
1073
+ "epoch": 0.49,
1074
+ "learning_rate": 2.129726739975486e-05,
1075
+ "loss": 1.9292,
1076
+ "step": 178
1077
+ },
1078
+ {
1079
+ "epoch": 0.49,
1080
+ "learning_rate": 2.112449497064223e-05,
1081
+ "loss": 1.9132,
1082
+ "step": 179
1083
+ },
1084
+ {
1085
+ "epoch": 0.49,
1086
+ "learning_rate": 2.095163831647485e-05,
1087
+ "loss": 1.9087,
1088
+ "step": 180
1089
+ },
1090
+ {
1091
+ "epoch": 0.49,
1092
+ "learning_rate": 2.0778710384274757e-05,
1093
+ "loss": 1.9137,
1094
+ "step": 181
1095
+ },
1096
+ {
1097
+ "epoch": 0.5,
1098
+ "learning_rate": 2.0605724126402734e-05,
1099
+ "loss": 1.8797,
1100
+ "step": 182
1101
+ },
1102
+ {
1103
+ "epoch": 0.5,
1104
+ "learning_rate": 2.0432692499588174e-05,
1105
+ "loss": 1.8937,
1106
+ "step": 183
1107
+ },
1108
+ {
1109
+ "epoch": 0.5,
1110
+ "learning_rate": 2.025962846395862e-05,
1111
+ "loss": 1.9124,
1112
+ "step": 184
1113
+ },
1114
+ {
1115
+ "epoch": 0.5,
1116
+ "learning_rate": 2.0086544982069046e-05,
1117
+ "loss": 1.8763,
1118
+ "step": 185
1119
+ },
1120
+ {
1121
+ "epoch": 0.51,
1122
+ "learning_rate": 1.9913455017930964e-05,
1123
+ "loss": 1.8699,
1124
+ "step": 186
1125
+ },
1126
+ {
1127
+ "epoch": 0.51,
1128
+ "learning_rate": 1.9740371536041388e-05,
1129
+ "loss": 1.8841,
1130
+ "step": 187
1131
+ },
1132
+ {
1133
+ "epoch": 0.51,
1134
+ "learning_rate": 1.9567307500411833e-05,
1135
+ "loss": 1.9116,
1136
+ "step": 188
1137
+ },
1138
+ {
1139
+ "epoch": 0.51,
1140
+ "learning_rate": 1.939427587359727e-05,
1141
+ "loss": 1.8997,
1142
+ "step": 189
1143
+ },
1144
+ {
1145
+ "epoch": 0.52,
1146
+ "learning_rate": 1.922128961572525e-05,
1147
+ "loss": 1.8716,
1148
+ "step": 190
1149
+ },
1150
+ {
1151
+ "epoch": 0.52,
1152
+ "learning_rate": 1.9048361683525155e-05,
1153
+ "loss": 1.8651,
1154
+ "step": 191
1155
+ },
1156
+ {
1157
+ "epoch": 0.52,
1158
+ "learning_rate": 1.8875505029357775e-05,
1159
+ "loss": 1.8827,
1160
+ "step": 192
1161
+ },
1162
+ {
1163
+ "epoch": 0.53,
1164
+ "learning_rate": 1.8702732600245138e-05,
1165
+ "loss": 1.871,
1166
+ "step": 193
1167
+ },
1168
+ {
1169
+ "epoch": 0.53,
1170
+ "learning_rate": 1.8530057336900805e-05,
1171
+ "loss": 1.8788,
1172
+ "step": 194
1173
+ },
1174
+ {
1175
+ "epoch": 0.53,
1176
+ "learning_rate": 1.835749217276055e-05,
1177
+ "loss": 1.8641,
1178
+ "step": 195
1179
+ },
1180
+ {
1181
+ "epoch": 0.53,
1182
+ "learning_rate": 1.8185050033013715e-05,
1183
+ "loss": 1.8856,
1184
+ "step": 196
1185
+ },
1186
+ {
1187
+ "epoch": 0.54,
1188
+ "learning_rate": 1.8012743833635067e-05,
1189
+ "loss": 1.8524,
1190
+ "step": 197
1191
+ },
1192
+ {
1193
+ "epoch": 0.54,
1194
+ "learning_rate": 1.7840586480417387e-05,
1195
+ "loss": 1.8142,
1196
+ "step": 198
1197
+ },
1198
+ {
1199
+ "epoch": 0.54,
1200
+ "learning_rate": 1.766859086800482e-05,
1201
+ "loss": 1.8361,
1202
+ "step": 199
1203
+ },
1204
+ {
1205
+ "epoch": 0.54,
1206
+ "learning_rate": 1.7496769878927085e-05,
1207
+ "loss": 1.8197,
1208
+ "step": 200
1209
+ },
1210
+ {
1211
+ "epoch": 0.55,
1212
+ "learning_rate": 1.7325136382634547e-05,
1213
+ "loss": 1.8723,
1214
+ "step": 201
1215
+ },
1216
+ {
1217
+ "epoch": 0.55,
1218
+ "learning_rate": 1.7153703234534302e-05,
1219
+ "loss": 1.8088,
1220
+ "step": 202
1221
+ },
1222
+ {
1223
+ "epoch": 0.55,
1224
+ "learning_rate": 1.6982483275027287e-05,
1225
+ "loss": 1.8264,
1226
+ "step": 203
1227
+ },
1228
+ {
1229
+ "epoch": 0.56,
1230
+ "learning_rate": 1.6811489328546557e-05,
1231
+ "loss": 1.7955,
1232
+ "step": 204
1233
+ },
1234
+ {
1235
+ "epoch": 0.56,
1236
+ "learning_rate": 1.6640734202596702e-05,
1237
+ "loss": 1.8337,
1238
+ "step": 205
1239
+ },
1240
+ {
1241
+ "epoch": 0.56,
1242
+ "learning_rate": 1.647023068679459e-05,
1243
+ "loss": 1.8571,
1244
+ "step": 206
1245
+ },
1246
+ {
1247
+ "epoch": 0.56,
1248
+ "learning_rate": 1.6299991551911373e-05,
1249
+ "loss": 1.8662,
1250
+ "step": 207
1251
+ },
1252
+ {
1253
+ "epoch": 0.57,
1254
+ "learning_rate": 1.6130029548916007e-05,
1255
+ "loss": 1.8593,
1256
+ "step": 208
1257
+ },
1258
+ {
1259
+ "epoch": 0.57,
1260
+ "learning_rate": 1.5960357408020178e-05,
1261
+ "loss": 1.8387,
1262
+ "step": 209
1263
+ },
1264
+ {
1265
+ "epoch": 0.57,
1266
+ "learning_rate": 1.579098783772479e-05,
1267
+ "loss": 1.8395,
1268
+ "step": 210
1269
+ },
1270
+ {
1271
+ "epoch": 0.57,
1272
+ "learning_rate": 1.5621933523868106e-05,
1273
+ "loss": 1.8271,
1274
+ "step": 211
1275
+ },
1276
+ {
1277
+ "epoch": 0.58,
1278
+ "learning_rate": 1.5453207128675583e-05,
1279
+ "loss": 1.8738,
1280
+ "step": 212
1281
+ },
1282
+ {
1283
+ "epoch": 0.58,
1284
+ "learning_rate": 1.5284821289811453e-05,
1285
+ "loss": 1.8598,
1286
+ "step": 213
1287
+ },
1288
+ {
1289
+ "epoch": 0.58,
1290
+ "learning_rate": 1.5116788619432177e-05,
1291
+ "loss": 1.832,
1292
+ "step": 214
1293
+ },
1294
+ {
1295
+ "epoch": 0.59,
1296
+ "learning_rate": 1.4949121703241747e-05,
1297
+ "loss": 1.8176,
1298
+ "step": 215
1299
+ },
1300
+ {
1301
+ "epoch": 0.59,
1302
+ "learning_rate": 1.4781833099549072e-05,
1303
+ "loss": 1.8302,
1304
+ "step": 216
1305
+ },
1306
+ {
1307
+ "epoch": 0.59,
1308
+ "learning_rate": 1.461493533832731e-05,
1309
+ "loss": 1.848,
1310
+ "step": 217
1311
+ },
1312
+ {
1313
+ "epoch": 0.59,
1314
+ "learning_rate": 1.4448440920275402e-05,
1315
+ "loss": 1.7919,
1316
+ "step": 218
1317
+ },
1318
+ {
1319
+ "epoch": 0.6,
1320
+ "learning_rate": 1.4282362315881719e-05,
1321
+ "loss": 1.8084,
1322
+ "step": 219
1323
+ },
1324
+ {
1325
+ "epoch": 0.6,
1326
+ "learning_rate": 1.4116711964490076e-05,
1327
+ "loss": 1.7979,
1328
+ "step": 220
1329
+ },
1330
+ {
1331
+ "epoch": 0.6,
1332
+ "learning_rate": 1.395150227336798e-05,
1333
+ "loss": 1.8167,
1334
+ "step": 221
1335
+ },
1336
+ {
1337
+ "epoch": 0.6,
1338
+ "learning_rate": 1.3786745616777348e-05,
1339
+ "loss": 1.8015,
1340
+ "step": 222
1341
+ },
1342
+ {
1343
+ "epoch": 0.61,
1344
+ "learning_rate": 1.3622454335047631e-05,
1345
+ "loss": 1.8359,
1346
+ "step": 223
1347
+ },
1348
+ {
1349
+ "epoch": 0.61,
1350
+ "learning_rate": 1.345864073365157e-05,
1351
+ "loss": 1.8111,
1352
+ "step": 224
1353
+ },
1354
+ {
1355
+ "epoch": 0.61,
1356
+ "learning_rate": 1.329531708228347e-05,
1357
+ "loss": 1.814,
1358
+ "step": 225
1359
+ },
1360
+ {
1361
+ "epoch": 0.62,
1362
+ "learning_rate": 1.3132495613940237e-05,
1363
+ "loss": 1.7876,
1364
+ "step": 226
1365
+ },
1366
+ {
1367
+ "epoch": 0.62,
1368
+ "learning_rate": 1.2970188524005058e-05,
1369
+ "loss": 1.7877,
1370
+ "step": 227
1371
+ },
1372
+ {
1373
+ "epoch": 0.62,
1374
+ "learning_rate": 1.2808407969334037e-05,
1375
+ "loss": 1.7683,
1376
+ "step": 228
1377
+ },
1378
+ {
1379
+ "epoch": 0.62,
1380
+ "learning_rate": 1.2647166067345598e-05,
1381
+ "loss": 1.8017,
1382
+ "step": 229
1383
+ },
1384
+ {
1385
+ "epoch": 0.63,
1386
+ "learning_rate": 1.24864748951129e-05,
1387
+ "loss": 1.7989,
1388
+ "step": 230
1389
+ },
1390
+ {
1391
+ "epoch": 0.63,
1392
+ "learning_rate": 1.2326346488459237e-05,
1393
+ "loss": 1.752,
1394
+ "step": 231
1395
+ },
1396
+ {
1397
+ "epoch": 0.63,
1398
+ "learning_rate": 1.2166792841056596e-05,
1399
+ "loss": 1.7771,
1400
+ "step": 232
1401
+ },
1402
+ {
1403
+ "epoch": 0.63,
1404
+ "learning_rate": 1.2007825903527287e-05,
1405
+ "loss": 1.7739,
1406
+ "step": 233
1407
+ },
1408
+ {
1409
+ "epoch": 0.64,
1410
+ "learning_rate": 1.1849457582548864e-05,
1411
+ "loss": 1.7821,
1412
+ "step": 234
1413
+ },
1414
+ {
1415
+ "epoch": 0.64,
1416
+ "learning_rate": 1.1691699739962275e-05,
1417
+ "loss": 1.7745,
1418
+ "step": 235
1419
+ },
1420
+ {
1421
+ "epoch": 0.64,
1422
+ "learning_rate": 1.153456419188345e-05,
1423
+ "loss": 1.7759,
1424
+ "step": 236
1425
+ },
1426
+ {
1427
+ "epoch": 0.65,
1428
+ "learning_rate": 1.137806270781824e-05,
1429
+ "loss": 1.795,
1430
+ "step": 237
1431
+ },
1432
+ {
1433
+ "epoch": 0.65,
1434
+ "learning_rate": 1.1222207009780888e-05,
1435
+ "loss": 1.7348,
1436
+ "step": 238
1437
+ },
1438
+ {
1439
+ "epoch": 0.65,
1440
+ "learning_rate": 1.1067008771416047e-05,
1441
+ "loss": 1.7385,
1442
+ "step": 239
1443
+ },
1444
+ {
1445
+ "epoch": 0.65,
1446
+ "learning_rate": 1.091247961712442e-05,
1447
+ "loss": 1.798,
1448
+ "step": 240
1449
+ },
1450
+ {
1451
+ "epoch": 0.66,
1452
+ "learning_rate": 1.0758631121192075e-05,
1453
+ "loss": 1.7615,
1454
+ "step": 241
1455
+ },
1456
+ {
1457
+ "epoch": 0.66,
1458
+ "learning_rate": 1.0605474806923556e-05,
1459
+ "loss": 1.7768,
1460
+ "step": 242
1461
+ },
1462
+ {
1463
+ "epoch": 0.66,
1464
+ "learning_rate": 1.0453022145778742e-05,
1465
+ "loss": 1.735,
1466
+ "step": 243
1467
+ },
1468
+ {
1469
+ "epoch": 0.66,
1470
+ "learning_rate": 1.0301284556513669e-05,
1471
+ "loss": 1.7635,
1472
+ "step": 244
1473
+ },
1474
+ {
1475
+ "epoch": 0.67,
1476
+ "learning_rate": 1.0150273404325244e-05,
1477
+ "loss": 1.7395,
1478
+ "step": 245
1479
+ },
1480
+ {
1481
+ "epoch": 0.67,
1482
+ "learning_rate": 1.0000000000000006e-05,
1483
+ "loss": 1.7238,
1484
+ "step": 246
1485
+ },
1486
+ {
1487
+ "epoch": 0.67,
1488
+ "learning_rate": 9.85047559906689e-06,
1489
+ "loss": 1.7633,
1490
+ "step": 247
1491
+ },
1492
+ {
1493
+ "epoch": 0.68,
1494
+ "learning_rate": 9.70171140095426e-06,
1495
+ "loss": 1.7585,
1496
+ "step": 248
1497
+ },
1498
+ {
1499
+ "epoch": 0.68,
1500
+ "learning_rate": 9.553718548151011e-06,
1501
+ "loss": 1.7379,
1502
+ "step": 249
1503
+ },
1504
+ {
1505
+ "epoch": 0.68,
1506
+ "learning_rate": 9.406508125372034e-06,
1507
+ "loss": 1.7427,
1508
+ "step": 250
1509
+ },
1510
+ {
1511
+ "epoch": 0.68,
1512
+ "learning_rate": 9.260091158727913e-06,
1513
+ "loss": 1.7205,
1514
+ "step": 251
1515
+ },
1516
+ {
1517
+ "epoch": 0.69,
1518
+ "learning_rate": 9.114478614899123e-06,
1519
+ "loss": 1.7576,
1520
+ "step": 252
1521
+ },
1522
+ {
1523
+ "epoch": 0.69,
1524
+ "learning_rate": 8.969681400314589e-06,
1525
+ "loss": 1.7559,
1526
+ "step": 253
1527
+ },
1528
+ {
1529
+ "epoch": 0.69,
1530
+ "learning_rate": 8.825710360334812e-06,
1531
+ "loss": 1.7465,
1532
+ "step": 254
1533
+ },
1534
+ {
1535
+ "epoch": 0.69,
1536
+ "learning_rate": 8.682576278439504e-06,
1537
+ "loss": 1.7571,
1538
+ "step": 255
1539
+ },
1540
+ {
1541
+ "epoch": 0.7,
1542
+ "learning_rate": 8.540289875419962e-06,
1543
+ "loss": 1.7239,
1544
+ "step": 256
1545
+ },
1546
+ {
1547
+ "epoch": 0.7,
1548
+ "learning_rate": 8.39886180857604e-06,
1549
+ "loss": 1.737,
1550
+ "step": 257
1551
+ },
1552
+ {
1553
+ "epoch": 0.7,
1554
+ "learning_rate": 8.258302670917915e-06,
1555
+ "loss": 1.7449,
1556
+ "step": 258
1557
+ },
1558
+ {
1559
+ "epoch": 0.71,
1560
+ "learning_rate": 8.118622990372676e-06,
1561
+ "loss": 1.7399,
1562
+ "step": 259
1563
+ },
1564
+ {
1565
+ "epoch": 0.71,
1566
+ "learning_rate": 7.979833228995782e-06,
1567
+ "loss": 1.7411,
1568
+ "step": 260
1569
+ },
1570
+ {
1571
+ "epoch": 0.71,
1572
+ "learning_rate": 7.841943782187435e-06,
1573
+ "loss": 1.7333,
1574
+ "step": 261
1575
+ },
1576
+ {
1577
+ "epoch": 0.71,
1578
+ "learning_rate": 7.704964977913984e-06,
1579
+ "loss": 1.7641,
1580
+ "step": 262
1581
+ },
1582
+ {
1583
+ "epoch": 0.72,
1584
+ "learning_rate": 7.568907075934322e-06,
1585
+ "loss": 1.7311,
1586
+ "step": 263
1587
+ },
1588
+ {
1589
+ "epoch": 0.72,
1590
+ "learning_rate": 7.433780267031463e-06,
1591
+ "loss": 1.7022,
1592
+ "step": 264
1593
+ },
1594
+ {
1595
+ "epoch": 0.72,
1596
+ "learning_rate": 7.299594672249231e-06,
1597
+ "loss": 1.7178,
1598
+ "step": 265
1599
+ },
1600
+ {
1601
+ "epoch": 0.72,
1602
+ "learning_rate": 7.166360342134202e-06,
1603
+ "loss": 1.7497,
1604
+ "step": 266
1605
+ },
1606
+ {
1607
+ "epoch": 0.73,
1608
+ "learning_rate": 7.034087255982882e-06,
1609
+ "loss": 1.7196,
1610
+ "step": 267
1611
+ },
1612
+ {
1613
+ "epoch": 0.73,
1614
+ "learning_rate": 6.902785321094301e-06,
1615
+ "loss": 1.707,
1616
+ "step": 268
1617
+ },
1618
+ {
1619
+ "epoch": 0.73,
1620
+ "learning_rate": 6.7724643720279156e-06,
1621
+ "loss": 1.7525,
1622
+ "step": 269
1623
+ },
1624
+ {
1625
+ "epoch": 0.74,
1626
+ "learning_rate": 6.643134169867031e-06,
1627
+ "loss": 1.6995,
1628
+ "step": 270
1629
+ },
1630
+ {
1631
+ "epoch": 0.74,
1632
+ "learning_rate": 6.514804401487642e-06,
1633
+ "loss": 1.7229,
1634
+ "step": 271
1635
+ },
1636
+ {
1637
+ "epoch": 0.74,
1638
+ "learning_rate": 6.38748467883294e-06,
1639
+ "loss": 1.7021,
1640
+ "step": 272
1641
+ },
1642
+ {
1643
+ "epoch": 0.74,
1644
+ "learning_rate": 6.261184538193341e-06,
1645
+ "loss": 1.6753,
1646
+ "step": 273
1647
+ },
1648
+ {
1649
+ "epoch": 0.75,
1650
+ "learning_rate": 6.135913439492227e-06,
1651
+ "loss": 1.7111,
1652
+ "step": 274
1653
+ },
1654
+ {
1655
+ "epoch": 0.75,
1656
+ "learning_rate": 6.01168076557737e-06,
1657
+ "loss": 1.7046,
1658
+ "step": 275
1659
+ },
1660
+ {
1661
+ "epoch": 0.75,
1662
+ "learning_rate": 5.888495821518194e-06,
1663
+ "loss": 1.7362,
1664
+ "step": 276
1665
+ },
1666
+ {
1667
+ "epoch": 0.75,
1668
+ "learning_rate": 5.7663678339087995e-06,
1669
+ "loss": 1.7811,
1670
+ "step": 277
1671
+ },
1672
+ {
1673
+ "epoch": 0.76,
1674
+ "learning_rate": 5.6453059501768806e-06,
1675
+ "loss": 1.6819,
1676
+ "step": 278
1677
+ },
1678
+ {
1679
+ "epoch": 0.76,
1680
+ "learning_rate": 5.5253192378985966e-06,
1681
+ "loss": 1.7154,
1682
+ "step": 279
1683
+ },
1684
+ {
1685
+ "epoch": 0.76,
1686
+ "learning_rate": 5.4064166841194e-06,
1687
+ "loss": 1.7069,
1688
+ "step": 280
1689
+ },
1690
+ {
1691
+ "epoch": 0.77,
1692
+ "learning_rate": 5.288607194680899e-06,
1693
+ "loss": 1.715,
1694
+ "step": 281
1695
+ },
1696
+ {
1697
+ "epoch": 0.77,
1698
+ "learning_rate": 5.171899593553824e-06,
1699
+ "loss": 1.7173,
1700
+ "step": 282
1701
+ },
1702
+ {
1703
+ "epoch": 0.77,
1704
+ "learning_rate": 5.056302622177074e-06,
1705
+ "loss": 1.6873,
1706
+ "step": 283
1707
+ },
1708
+ {
1709
+ "epoch": 0.77,
1710
+ "learning_rate": 4.941824938803024e-06,
1711
+ "loss": 1.7291,
1712
+ "step": 284
1713
+ },
1714
+ {
1715
+ "epoch": 0.78,
1716
+ "learning_rate": 4.828475117848992e-06,
1717
+ "loss": 1.6928,
1718
+ "step": 285
1719
+ },
1720
+ {
1721
+ "epoch": 0.78,
1722
+ "learning_rate": 4.716261649255021e-06,
1723
+ "loss": 1.6815,
1724
+ "step": 286
1725
+ },
1726
+ {
1727
+ "epoch": 0.78,
1728
+ "learning_rate": 4.605192937847962e-06,
1729
+ "loss": 1.6701,
1730
+ "step": 287
1731
+ },
1732
+ {
1733
+ "epoch": 0.78,
1734
+ "learning_rate": 4.495277302711982e-06,
1735
+ "loss": 1.7119,
1736
+ "step": 288
1737
+ },
1738
+ {
1739
+ "epoch": 0.79,
1740
+ "learning_rate": 4.386522976565439e-06,
1741
+ "loss": 1.6813,
1742
+ "step": 289
1743
+ },
1744
+ {
1745
+ "epoch": 0.79,
1746
+ "learning_rate": 4.278938105144255e-06,
1747
+ "loss": 1.6945,
1748
+ "step": 290
1749
+ },
1750
+ {
1751
+ "epoch": 0.79,
1752
+ "learning_rate": 4.172530746591783e-06,
1753
+ "loss": 1.7221,
1754
+ "step": 291
1755
+ },
1756
+ {
1757
+ "epoch": 0.8,
1758
+ "learning_rate": 4.06730887085528e-06,
1759
+ "loss": 1.7254,
1760
+ "step": 292
1761
+ },
1762
+ {
1763
+ "epoch": 0.8,
1764
+ "learning_rate": 3.963280359088933e-06,
1765
+ "loss": 1.6873,
1766
+ "step": 293
1767
+ },
1768
+ {
1769
+ "epoch": 0.8,
1770
+ "learning_rate": 3.86045300306356e-06,
1771
+ "loss": 1.6773,
1772
+ "step": 294
1773
+ },
1774
+ {
1775
+ "epoch": 0.8,
1776
+ "learning_rate": 3.7588345045830044e-06,
1777
+ "loss": 1.7121,
1778
+ "step": 295
1779
+ },
1780
+ {
1781
+ "epoch": 0.81,
1782
+ "learning_rate": 3.658432474907274e-06,
1783
+ "loss": 1.6798,
1784
+ "step": 296
1785
+ },
1786
+ {
1787
+ "epoch": 0.81,
1788
+ "learning_rate": 3.559254434182451e-06,
1789
+ "loss": 1.6762,
1790
+ "step": 297
1791
+ },
1792
+ {
1793
+ "epoch": 0.81,
1794
+ "learning_rate": 3.461307810877428e-06,
1795
+ "loss": 1.6934,
1796
+ "step": 298
1797
+ },
1798
+ {
1799
+ "epoch": 0.81,
1800
+ "learning_rate": 3.364599941227513e-06,
1801
+ "loss": 1.6862,
1802
+ "step": 299
1803
+ },
1804
+ {
1805
+ "epoch": 0.82,
1806
+ "learning_rate": 3.2691380686849517e-06,
1807
+ "loss": 1.6708,
1808
+ "step": 300
1809
+ },
1810
+ {
1811
+ "epoch": 0.82,
1812
+ "learning_rate": 3.174929343376374e-06,
1813
+ "loss": 1.6764,
1814
+ "step": 301
1815
+ },
1816
+ {
1817
+ "epoch": 0.82,
1818
+ "learning_rate": 3.081980821567272e-06,
1819
+ "loss": 1.6886,
1820
+ "step": 302
1821
+ },
1822
+ {
1823
+ "epoch": 0.83,
1824
+ "learning_rate": 2.990299465133446e-06,
1825
+ "loss": 1.6679,
1826
+ "step": 303
1827
+ },
1828
+ {
1829
+ "epoch": 0.83,
1830
+ "learning_rate": 2.8998921410396e-06,
1831
+ "loss": 1.6854,
1832
+ "step": 304
1833
+ },
1834
+ {
1835
+ "epoch": 0.83,
1836
+ "learning_rate": 2.8107656208249733e-06,
1837
+ "loss": 1.681,
1838
+ "step": 305
1839
+ },
1840
+ {
1841
+ "epoch": 0.83,
1842
+ "learning_rate": 2.72292658009617e-06,
1843
+ "loss": 1.6987,
1844
+ "step": 306
1845
+ },
1846
+ {
1847
+ "epoch": 0.84,
1848
+ "learning_rate": 2.6363815980271248e-06,
1849
+ "loss": 1.6574,
1850
+ "step": 307
1851
+ },
1852
+ {
1853
+ "epoch": 0.84,
1854
+ "learning_rate": 2.551137156866357e-06,
1855
+ "loss": 1.7037,
1856
+ "step": 308
1857
+ },
1858
+ {
1859
+ "epoch": 0.84,
1860
+ "learning_rate": 2.4671996414514276e-06,
1861
+ "loss": 1.702,
1862
+ "step": 309
1863
+ },
1864
+ {
1865
+ "epoch": 0.84,
1866
+ "learning_rate": 2.384575338730717e-06,
1867
+ "loss": 1.6788,
1868
+ "step": 310
1869
+ },
1870
+ {
1871
+ "epoch": 0.85,
1872
+ "learning_rate": 2.3032704372925176e-06,
1873
+ "loss": 1.6624,
1874
+ "step": 311
1875
+ },
1876
+ {
1877
+ "epoch": 0.85,
1878
+ "learning_rate": 2.223291026901533e-06,
1879
+ "loss": 1.693,
1880
+ "step": 312
1881
+ },
1882
+ {
1883
+ "epoch": 0.85,
1884
+ "learning_rate": 2.144643098042727e-06,
1885
+ "loss": 1.6687,
1886
+ "step": 313
1887
+ },
1888
+ {
1889
+ "epoch": 0.86,
1890
+ "learning_rate": 2.0673325414726574e-06,
1891
+ "loss": 1.6914,
1892
+ "step": 314
1893
+ },
1894
+ {
1895
+ "epoch": 0.86,
1896
+ "learning_rate": 1.991365147778228e-06,
1897
+ "loss": 1.6708,
1898
+ "step": 315
1899
+ },
1900
+ {
1901
+ "epoch": 0.86,
1902
+ "learning_rate": 1.9167466069429964e-06,
1903
+ "loss": 1.6752,
1904
+ "step": 316
1905
+ },
1906
+ {
1907
+ "epoch": 0.86,
1908
+ "learning_rate": 1.8434825079209884e-06,
1909
+ "loss": 1.6893,
1910
+ "step": 317
1911
+ },
1912
+ {
1913
+ "epoch": 0.87,
1914
+ "learning_rate": 1.7715783382180672e-06,
1915
+ "loss": 1.6681,
1916
+ "step": 318
1917
+ },
1918
+ {
1919
+ "epoch": 0.87,
1920
+ "learning_rate": 1.7010394834809373e-06,
1921
+ "loss": 1.6982,
1922
+ "step": 319
1923
+ },
1924
+ {
1925
+ "epoch": 0.87,
1926
+ "learning_rate": 1.6318712270937442e-06,
1927
+ "loss": 1.6421,
1928
+ "step": 320
1929
+ },
1930
+ {
1931
+ "epoch": 0.87,
1932
+ "learning_rate": 1.5640787497823585e-06,
1933
+ "loss": 1.7007,
1934
+ "step": 321
1935
+ },
1936
+ {
1937
+ "epoch": 0.88,
1938
+ "learning_rate": 1.4976671292263257e-06,
1939
+ "loss": 1.6832,
1940
+ "step": 322
1941
+ },
1942
+ {
1943
+ "epoch": 0.88,
1944
+ "learning_rate": 1.4326413396785488e-06,
1945
+ "loss": 1.6584,
1946
+ "step": 323
1947
+ },
1948
+ {
1949
+ "epoch": 0.88,
1950
+ "learning_rate": 1.3690062515927239e-06,
1951
+ "loss": 1.7009,
1952
+ "step": 324
1953
+ },
1954
+ {
1955
+ "epoch": 0.89,
1956
+ "learning_rate": 1.306766631258536e-06,
1957
+ "loss": 1.6755,
1958
+ "step": 325
1959
+ },
1960
+ {
1961
+ "epoch": 0.89,
1962
+ "learning_rate": 1.245927140444665e-06,
1963
+ "loss": 1.686,
1964
+ "step": 326
1965
+ },
1966
+ {
1967
+ "epoch": 0.89,
1968
+ "learning_rate": 1.1864923360496028e-06,
1969
+ "loss": 1.6916,
1970
+ "step": 327
1971
+ },
1972
+ {
1973
+ "epoch": 0.89,
1974
+ "learning_rate": 1.128466669760362e-06,
1975
+ "loss": 1.6923,
1976
+ "step": 328
1977
+ },
1978
+ {
1979
+ "epoch": 0.9,
1980
+ "learning_rate": 1.0718544877190306e-06,
1981
+ "loss": 1.7079,
1982
+ "step": 329
1983
+ },
1984
+ {
1985
+ "epoch": 0.9,
1986
+ "learning_rate": 1.0166600301972517e-06,
1987
+ "loss": 1.6865,
1988
+ "step": 330
1989
+ },
1990
+ {
1991
+ "epoch": 0.9,
1992
+ "learning_rate": 9.628874312786096e-07,
1993
+ "loss": 1.6624,
1994
+ "step": 331
1995
+ },
1996
+ {
1997
+ "epoch": 0.9,
1998
+ "learning_rate": 9.105407185490067e-07,
1999
+ "loss": 1.6489,
2000
+ "step": 332
2001
+ },
2002
+ {
2003
+ "epoch": 0.91,
2004
+ "learning_rate": 8.59623812794983e-07,
2005
+ "loss": 1.6923,
2006
+ "step": 333
2007
+ },
2008
+ {
2009
+ "epoch": 0.91,
2010
+ "learning_rate": 8.101405277100549e-07,
2011
+ "loss": 1.678,
2012
+ "step": 334
2013
+ },
2014
+ {
2015
+ "epoch": 0.91,
2016
+ "learning_rate": 7.620945696090532e-07,
2017
+ "loss": 1.6302,
2018
+ "step": 335
2019
+ },
2020
+ {
2021
+ "epoch": 0.92,
2022
+ "learning_rate": 7.154895371505421e-07,
2023
+ "loss": 1.6555,
2024
+ "step": 336
2025
+ },
2026
+ {
2027
+ "epoch": 0.92,
2028
+ "learning_rate": 6.703289210672603e-07,
2029
+ "loss": 1.6805,
2030
+ "step": 337
2031
+ },
2032
+ {
2033
+ "epoch": 0.92,
2034
+ "learning_rate": 6.266161039046737e-07,
2035
+ "loss": 1.6744,
2036
+ "step": 338
2037
+ },
2038
+ {
2039
+ "epoch": 0.92,
2040
+ "learning_rate": 5.843543597676138e-07,
2041
+ "loss": 1.6585,
2042
+ "step": 339
2043
+ },
2044
+ {
2045
+ "epoch": 0.93,
2046
+ "learning_rate": 5.435468540750544e-07,
2047
+ "loss": 1.6951,
2048
+ "step": 340
2049
+ },
2050
+ {
2051
+ "epoch": 0.93,
2052
+ "learning_rate": 5.041966433230094e-07,
2053
+ "loss": 1.6666,
2054
+ "step": 341
2055
+ },
2056
+ {
2057
+ "epoch": 0.93,
2058
+ "learning_rate": 4.6630667485561885e-07,
2059
+ "loss": 1.6757,
2060
+ "step": 342
2061
+ },
2062
+ {
2063
+ "epoch": 0.93,
2064
+ "learning_rate": 4.2987978664436936e-07,
2065
+ "loss": 1.6719,
2066
+ "step": 343
2067
+ },
2068
+ {
2069
+ "epoch": 0.94,
2070
+ "learning_rate": 3.9491870707554445e-07,
2071
+ "loss": 1.6575,
2072
+ "step": 344
2073
+ },
2074
+ {
2075
+ "epoch": 0.94,
2076
+ "learning_rate": 3.614260547458659e-07,
2077
+ "loss": 1.6398,
2078
+ "step": 345
2079
+ },
2080
+ {
2081
+ "epoch": 0.94,
2082
+ "learning_rate": 3.2940433826635257e-07,
2083
+ "loss": 1.6523,
2084
+ "step": 346
2085
+ },
2086
+ {
2087
+ "epoch": 0.95,
2088
+ "learning_rate": 2.9885595607443086e-07,
2089
+ "loss": 1.6521,
2090
+ "step": 347
2091
+ },
2092
+ {
2093
+ "epoch": 0.95,
2094
+ "learning_rate": 2.697831962542874e-07,
2095
+ "loss": 1.6598,
2096
+ "step": 348
2097
+ },
2098
+ {
2099
+ "epoch": 0.95,
2100
+ "learning_rate": 2.4218823636549703e-07,
2101
+ "loss": 1.6721,
2102
+ "step": 349
2103
+ },
2104
+ {
2105
+ "epoch": 0.95,
2106
+ "learning_rate": 2.1607314327991791e-07,
2107
+ "loss": 1.6975,
2108
+ "step": 350
2109
+ },
2110
+ {
2111
+ "epoch": 0.96,
2112
+ "learning_rate": 1.9143987302687738e-07,
2113
+ "loss": 1.6652,
2114
+ "step": 351
2115
+ },
2116
+ {
2117
+ "epoch": 0.96,
2118
+ "learning_rate": 1.682902706466738e-07,
2119
+ "loss": 1.6554,
2120
+ "step": 352
2121
+ },
2122
+ {
2123
+ "epoch": 0.96,
2124
+ "learning_rate": 1.4662607005237805e-07,
2125
+ "loss": 1.6603,
2126
+ "step": 353
2127
+ },
2128
+ {
2129
+ "epoch": 0.96,
2130
+ "learning_rate": 1.264488938999664e-07,
2131
+ "loss": 1.6843,
2132
+ "step": 354
2133
+ },
2134
+ {
2135
+ "epoch": 0.97,
2136
+ "learning_rate": 1.0776025346677321e-07,
2137
+ "loss": 1.69,
2138
+ "step": 355
2139
+ },
2140
+ {
2141
+ "epoch": 0.97,
2142
+ "learning_rate": 9.056154853830823e-08,
2143
+ "loss": 1.6755,
2144
+ "step": 356
2145
+ },
2146
+ {
2147
+ "epoch": 0.97,
2148
+ "learning_rate": 7.485406730340483e-08,
2149
+ "loss": 1.6577,
2150
+ "step": 357
2151
+ },
2152
+ {
2153
+ "epoch": 0.98,
2154
+ "learning_rate": 6.06389862577328e-08,
2155
+ "loss": 1.7059,
2156
+ "step": 358
2157
+ },
2158
+ {
2159
+ "epoch": 0.98,
2160
+ "learning_rate": 4.7917370115688756e-08,
2161
+ "loss": 1.7009,
2162
+ "step": 359
2163
+ },
2164
+ {
2165
+ "epoch": 0.98,
2166
+ "learning_rate": 3.66901717306356e-08,
2167
+ "loss": 1.6863,
2168
+ "step": 360
2169
+ },
2170
+ {
2171
+ "epoch": 0.98,
2172
+ "learning_rate": 2.6958232023539532e-08,
2173
+ "loss": 1.6557,
2174
+ "step": 361
2175
+ },
2176
+ {
2177
+ "epoch": 0.99,
2178
+ "learning_rate": 1.8722279919987098e-08,
2179
+ "loss": 1.6717,
2180
+ "step": 362
2181
+ },
2182
+ {
2183
+ "epoch": 0.99,
2184
+ "learning_rate": 1.1982932295582227e-08,
2185
+ "loss": 1.6401,
2186
+ "step": 363
2187
+ },
2188
+ {
2189
+ "epoch": 0.99,
2190
+ "learning_rate": 6.7406939297520734e-09,
2191
+ "loss": 1.6846,
2192
+ "step": 364
2193
+ },
2194
+ {
2195
+ "epoch": 0.99,
2196
+ "learning_rate": 2.995957467923916e-09,
2197
+ "loss": 1.6775,
2198
+ "step": 365
2199
+ },
2200
+ {
2201
+ "epoch": 1.0,
2202
+ "learning_rate": 7.490033921331296e-10,
2203
+ "loss": 1.675,
2204
+ "step": 366
2205
+ },
2206
+ {
2207
+ "epoch": 1.0,
2208
+ "learning_rate": 0.0,
2209
+ "loss": 1.6419,
2210
+ "step": 367
2211
+ },
2212
+ {
2213
+ "epoch": 1.0,
2214
+ "step": 367,
2215
+ "total_flos": 3.839475069608788e+18,
2216
+ "train_loss": 2.069092570598509,
2217
+ "train_runtime": 4018.2223,
2218
+ "train_samples_per_second": 174.946,
2219
+ "train_steps_per_second": 0.091
2220
+ }
2221
+ ],
2222
+ "max_steps": 367,
2223
+ "num_train_epochs": 1,
2224
+ "total_flos": 3.839475069608788e+18,
2225
+ "trial_name": null,
2226
+ "trial_params": null
2227
+ }
videopanda_prtr1/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ead5735ee7e225df797e6b040526b57e8b2693e87913d98773d1ee5b02173da8
3
+ size 6136