Hugging Face's logo

336labs
/
VisionEncoder-to-MLLM-ModelZoo

Safetensors
vision-encoder-evaluation
multimodal
visual-tokenizer-benchmark
checkpoints
Model card Files
xet
 Community
lyunee commited on
Commit
d8744c3
·
verified ·
1 Parent(s): 2b34c41

Upload continuous evaluation checkpoints, batch 23

Browse files
Files changed (50) hide show
  1. .gitattributes +3 -0
  2. continuous/qwen3/siglip2_l16_512/pretrain/config.json +60 -0
  3. continuous/qwen3/siglip2_l16_512/pretrain/mm_projector.bin +3 -0
  4. continuous/qwen3/siglip2_l16_512/pretrain/trainer_state.json +421 -0
  5. continuous/qwen3/siglip2_sm14_224/finetune/added_tokens.json +28 -0
  6. continuous/qwen3/siglip2_sm14_224/finetune/config.json +59 -0
  7. continuous/qwen3/siglip2_sm14_224/finetune/generation_config.json +13 -0
  8. continuous/qwen3/siglip2_sm14_224/finetune/merges.txt +0 -0
  9. continuous/qwen3/siglip2_sm14_224/finetune/model.safetensors +3 -0
  10. continuous/qwen3/siglip2_sm14_224/finetune/special_tokens_map.json +31 -0
  11. continuous/qwen3/siglip2_sm14_224/finetune/tokenizer.json +3 -0
  12. continuous/qwen3/siglip2_sm14_224/finetune/tokenizer_config.json +241 -0
  13. continuous/qwen3/siglip2_sm14_224/finetune/trainer_state.json +1856 -0
  14. continuous/qwen3/siglip2_sm14_224/finetune/training_args.bin +3 -0
  15. continuous/qwen3/siglip2_sm14_224/finetune/vocab.json +0 -0
  16. continuous/qwen3/siglip2_sm14_224/pretrain/config.json +60 -0
  17. continuous/qwen3/siglip2_sm14_224/pretrain/mm_projector.bin +3 -0
  18. continuous/qwen3/siglip2_sm14_224/pretrain/trainer_state.json +421 -0
  19. continuous/qwen3/siglip2_sm14_384/finetune/added_tokens.json +28 -0
  20. continuous/qwen3/siglip2_sm14_384/finetune/config.json +59 -0
  21. continuous/qwen3/siglip2_sm14_384/finetune/generation_config.json +13 -0
  22. continuous/qwen3/siglip2_sm14_384/finetune/merges.txt +0 -0
  23. continuous/qwen3/siglip2_sm14_384/finetune/model.safetensors +3 -0
  24. continuous/qwen3/siglip2_sm14_384/finetune/special_tokens_map.json +31 -0
  25. continuous/qwen3/siglip2_sm14_384/finetune/tokenizer.json +3 -0
  26. continuous/qwen3/siglip2_sm14_384/finetune/tokenizer_config.json +241 -0
  27. continuous/qwen3/siglip2_sm14_384/finetune/trainer_state.json +1856 -0
  28. continuous/qwen3/siglip2_sm14_384/finetune/training_args.bin +3 -0
  29. continuous/qwen3/siglip2_sm14_384/finetune/vocab.json +0 -0
  30. continuous/qwen3/siglip2_sm14_384/pretrain/config.json +60 -0
  31. continuous/qwen3/siglip2_sm14_384/pretrain/mm_projector.bin +3 -0
  32. continuous/qwen3/siglip2_sm14_384/pretrain/trainer_state.json +421 -0
  33. continuous/qwen3/siglip2_sm16_256/finetune/added_tokens.json +28 -0
  34. continuous/qwen3/siglip2_sm16_256/finetune/config.json +59 -0
  35. continuous/qwen3/siglip2_sm16_256/finetune/generation_config.json +13 -0
  36. continuous/qwen3/siglip2_sm16_256/finetune/merges.txt +0 -0
  37. continuous/qwen3/siglip2_sm16_256/finetune/model.safetensors +3 -0
  38. continuous/qwen3/siglip2_sm16_256/finetune/special_tokens_map.json +31 -0
  39. continuous/qwen3/siglip2_sm16_256/finetune/tokenizer.json +3 -0
  40. continuous/qwen3/siglip2_sm16_256/finetune/tokenizer_config.json +241 -0
  41. continuous/qwen3/siglip2_sm16_256/finetune/trainer_state.json +1856 -0
  42. continuous/qwen3/siglip2_sm16_256/finetune/training_args.bin +3 -0
  43. continuous/qwen3/siglip2_sm16_256/finetune/vocab.json +0 -0
  44. continuous/qwen3/siglip2_sm16_256/pretrain/config.json +60 -0
  45. continuous/qwen3/siglip2_sm16_256/pretrain/mm_projector.bin +3 -0
  46. continuous/qwen3/siglip2_sm16_256/pretrain/trainer_state.json +421 -0
  47. continuous/qwen3/siglip2_sm16_384/finetune/added_tokens.json +28 -0
  48. continuous/qwen3/siglip2_sm16_384/finetune/config.json +59 -0
  49. continuous/qwen3/siglip2_sm16_384/finetune/generation_config.json +13 -0
  50. continuous/qwen3/siglip2_sm16_384/finetune/merges.txt +0 -0
.gitattributes CHANGED
@@ -167,3 +167,6 @@ continuous/qwen3/siglip2_g16_384/finetune/tokenizer.json filter=lfs diff=lfs mer
167
  continuous/qwen3/siglip2_l16_256/finetune/tokenizer.json filter=lfs diff=lfs merge=lfs -text
168
  continuous/qwen3/siglip2_l16_384/finetune/tokenizer.json filter=lfs diff=lfs merge=lfs -text
169
  continuous/qwen3/siglip2_l16_512/finetune/tokenizer.json filter=lfs diff=lfs merge=lfs -text
 
 
 
 
167
  continuous/qwen3/siglip2_l16_256/finetune/tokenizer.json filter=lfs diff=lfs merge=lfs -text
168
  continuous/qwen3/siglip2_l16_384/finetune/tokenizer.json filter=lfs diff=lfs merge=lfs -text
169
  continuous/qwen3/siglip2_l16_512/finetune/tokenizer.json filter=lfs diff=lfs merge=lfs -text
170
+ continuous/qwen3/siglip2_sm14_224/finetune/tokenizer.json filter=lfs diff=lfs merge=lfs -text
171
+ continuous/qwen3/siglip2_sm14_384/finetune/tokenizer.json filter=lfs diff=lfs merge=lfs -text
172
+ continuous/qwen3/siglip2_sm16_256/finetune/tokenizer.json filter=lfs diff=lfs merge=lfs -text
continuous/qwen3/siglip2_l16_512/pretrain/config.json ADDED
@@ -0,0 +1,60 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_attn_implementation_autoset": true,
3
+ "add_faster_video": false,
4
+ "add_time_instruction": false,
5
+ "architectures": [
6
+ "Qwen3ForCausalLM"
7
+ ],
8
+ "attention_bias": false,
9
+ "attention_dropout": 0.0,
10
+ "bos_token_id": 151643,
11
+ "eos_token_id": 151645,
12
+ "faster_token_stride": 10,
13
+ "force_sample": false,
14
+ "head_dim": 128,
15
+ "hidden_act": "silu",
16
+ "hidden_size": 2048,
17
+ "image_aspect_ratio": "square",
18
+ "image_crop_resolution": null,
19
+ "image_grid_pinpoints": null,
20
+ "image_split_resolution": null,
21
+ "initializer_range": 0.02,
22
+ "intermediate_size": 6144,
23
+ "max_position_embeddings": 40960,
24
+ "max_window_layers": 28,
25
+ "mm_hidden_size": 1024,
26
+ "mm_newline_position": "grid",
27
+ "mm_patch_merge_type": "flat",
28
+ "mm_projector_lr": null,
29
+ "mm_projector_type": "mlp2x_gelu",
30
+ "mm_resampler_type": null,
31
+ "mm_spatial_pool_mode": "bilinear",
32
+ "mm_spatial_pool_stride": null,
33
+ "mm_tunable_parts": "mm_mlp_adapter",
34
+ "mm_use_im_patch_token": false,
35
+ "mm_use_im_start_end": false,
36
+ "mm_vision_select_feature": "patch",
37
+ "mm_vision_select_layer": -2,
38
+ "mm_vision_tower": "open_clip_hub:ViT-L-16-SigLIP2-512",
39
+ "mm_vision_tower_lr": null,
40
+ "model_type": "qwen3",
41
+ "num_attention_heads": 16,
42
+ "num_hidden_layers": 28,
43
+ "num_key_value_heads": 8,
44
+ "pos_skipping_range": 4096,
45
+ "rms_norm_eps": 1e-06,
46
+ "rope_scaling": null,
47
+ "rope_theta": 1000000,
48
+ "sliding_window": null,
49
+ "tie_word_embeddings": true,
50
+ "tokenizer_model_max_length": 8192,
51
+ "tokenizer_padding_side": "right",
52
+ "torch_dtype": "bfloat16",
53
+ "transformers_version": "4.51.3",
54
+ "use_cache": true,
55
+ "use_mm_proj": true,
56
+ "use_pos_skipping": false,
57
+ "use_sliding_window": false,
58
+ "vision_tower_pretrained": "/cache/ckpt/download/tokenizer/continuous/siglip2/siglip2_l16_512.npz",
59
+ "vocab_size": 151936
60
+ }
continuous/qwen3/siglip2_l16_512/pretrain/mm_projector.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5475fcbc967fae04b03eb7f1a30775c98ec580fb28e6da262a2f84dcd24ce8e2
3
+ size 25184312
continuous/qwen3/siglip2_l16_512/pretrain/trainer_state.json ADDED
@@ -0,0 +1,421 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_global_step": null,
3
+ "best_metric": null,
4
+ "best_model_checkpoint": null,
5
+ "epoch": 1.0,
6
+ "eval_steps": 500,
7
+ "global_step": 545,
8
+ "is_hyper_param_search": false,
9
+ "is_local_process_zero": true,
10
+ "is_world_process_zero": true,
11
+ "log_history": [
12
+ {
13
+ "epoch": 0.01834862385321101,
14
+ "grad_norm": 1.1055939197540283,
15
+ "learning_rate": 0.00032142857142857147,
16
+ "loss": 8.8225,
17
+ "step": 10
18
+ },
19
+ {
20
+ "epoch": 0.03669724770642202,
21
+ "grad_norm": 0.08038423210382462,
22
+ "learning_rate": 0.0006785714285714287,
23
+ "loss": 5.38,
24
+ "step": 20
25
+ },
26
+ {
27
+ "epoch": 0.05504587155963303,
28
+ "grad_norm": 0.022893423214554787,
29
+ "learning_rate": 0.0009999907688176173,
30
+ "loss": 4.8449,
31
+ "step": 30
32
+ },
33
+ {
34
+ "epoch": 0.07339449541284404,
35
+ "grad_norm": 0.02245972864329815,
36
+ "learning_rate": 0.0009988834393115766,
37
+ "loss": 4.7261,
38
+ "step": 40
39
+ },
40
+ {
41
+ "epoch": 0.09174311926605505,
42
+ "grad_norm": 0.011426088400185108,
43
+ "learning_rate": 0.0009959345572940772,
44
+ "loss": 4.6807,
45
+ "step": 50
46
+ },
47
+ {
48
+ "epoch": 0.11009174311926606,
49
+ "grad_norm": 0.009295499883592129,
50
+ "learning_rate": 0.0009911550081155983,
51
+ "loss": 4.658,
52
+ "step": 60
53
+ },
54
+ {
55
+ "epoch": 0.12844036697247707,
56
+ "grad_norm": 0.012599905952811241,
57
+ "learning_rate": 0.0009845624347566433,
58
+ "loss": 4.6271,
59
+ "step": 70
60
+ },
61
+ {
62
+ "epoch": 0.14678899082568808,
63
+ "grad_norm": 0.025406504049897194,
64
+ "learning_rate": 0.0009761811727013547,
65
+ "loss": 4.6098,
66
+ "step": 80
67
+ },
68
+ {
69
+ "epoch": 0.1651376146788991,
70
+ "grad_norm": 0.03194766864180565,
71
+ "learning_rate": 0.0009660421601067666,
72
+ "loss": 4.5708,
73
+ "step": 90
74
+ },
75
+ {
76
+ "epoch": 0.1834862385321101,
77
+ "grad_norm": 0.04352506250143051,
78
+ "learning_rate": 0.0009541828235992926,
79
+ "loss": 4.5294,
80
+ "step": 100
81
+ },
82
+ {
83
+ "epoch": 0.2018348623853211,
84
+ "grad_norm": 0.04785310849547386,
85
+ "learning_rate": 0.0009406469401200151,
86
+ "loss": 4.4809,
87
+ "step": 110
88
+ },
89
+ {
90
+ "epoch": 0.22018348623853212,
91
+ "grad_norm": 0.03392815217375755,
92
+ "learning_rate": 0.0009254844753287492,
93
+ "loss": 4.456,
94
+ "step": 120
95
+ },
96
+ {
97
+ "epoch": 0.23853211009174313,
98
+ "grad_norm": 0.0567086860537529,
99
+ "learning_rate": 0.0009087513991633924,
100
+ "loss": 4.4118,
101
+ "step": 130
102
+ },
103
+ {
104
+ "epoch": 0.25688073394495414,
105
+ "grad_norm": 0.04148750752210617,
106
+ "learning_rate": 0.0008905094792353917,
107
+ "loss": 4.3904,
108
+ "step": 140
109
+ },
110
+ {
111
+ "epoch": 0.27522935779816515,
112
+ "grad_norm": 0.04047899693250656,
113
+ "learning_rate": 0.0008708260528239789,
114
+ "loss": 4.3764,
115
+ "step": 150
116
+ },
117
+ {
118
+ "epoch": 0.29357798165137616,
119
+ "grad_norm": 0.03482604771852493,
120
+ "learning_rate": 0.0008497737783108237,
121
+ "loss": 4.3517,
122
+ "step": 160
123
+ },
124
+ {
125
+ "epoch": 0.3119266055045872,
126
+ "grad_norm": 0.05196757614612579,
127
+ "learning_rate": 0.0008274303669726426,
128
+ "loss": 4.3408,
129
+ "step": 170
130
+ },
131
+ {
132
+ "epoch": 0.3302752293577982,
133
+ "grad_norm": 0.034356147050857544,
134
+ "learning_rate": 0.0008038782961218137,
135
+ "loss": 4.3215,
136
+ "step": 180
137
+ },
138
+ {
139
+ "epoch": 0.3486238532110092,
140
+ "grad_norm": 0.043870747089385986,
141
+ "learning_rate": 0.000779204504653894,
142
+ "loss": 4.3144,
143
+ "step": 190
144
+ },
145
+ {
146
+ "epoch": 0.3669724770642202,
147
+ "grad_norm": 0.04068443179130554,
148
+ "learning_rate": 0.0007535000721258791,
149
+ "loss": 4.288,
150
+ "step": 200
151
+ },
152
+ {
153
+ "epoch": 0.3853211009174312,
154
+ "grad_norm": 0.04307029768824577,
155
+ "learning_rate": 0.00072685988254984,
156
+ "loss": 4.2809,
157
+ "step": 210
158
+ },
159
+ {
160
+ "epoch": 0.4036697247706422,
161
+ "grad_norm": 0.0323168970644474,
162
+ "learning_rate": 0.0006993822741429906,
163
+ "loss": 4.2892,
164
+ "step": 220
165
+ },
166
+ {
167
+ "epoch": 0.42201834862385323,
168
+ "grad_norm": 0.040014076977968216,
169
+ "learning_rate": 0.0006711686763270818,
170
+ "loss": 4.2831,
171
+ "step": 230
172
+ },
173
+ {
174
+ "epoch": 0.44036697247706424,
175
+ "grad_norm": 0.04625088348984718,
176
+ "learning_rate": 0.0006423232353170831,
177
+ "loss": 4.2644,
178
+ "step": 240
179
+ },
180
+ {
181
+ "epoch": 0.45871559633027525,
182
+ "grad_norm": 0.03950924798846245,
183
+ "learning_rate": 0.0006129524296812335,
184
+ "loss": 4.246,
185
+ "step": 250
186
+ },
187
+ {
188
+ "epoch": 0.47706422018348627,
189
+ "grad_norm": 0.034068699926137924,
190
+ "learning_rate": 0.0005831646772915651,
191
+ "loss": 4.2566,
192
+ "step": 260
193
+ },
194
+ {
195
+ "epoch": 0.4954128440366973,
196
+ "grad_norm": 0.03259763494133949,
197
+ "learning_rate": 0.0005530699351157809,
198
+ "loss": 4.2333,
199
+ "step": 270
200
+ },
201
+ {
202
+ "epoch": 0.5137614678899083,
203
+ "grad_norm": 0.03122481144964695,
204
+ "learning_rate": 0.0005227792933277943,
205
+ "loss": 4.2277,
206
+ "step": 280
207
+ },
208
+ {
209
+ "epoch": 0.5321100917431193,
210
+ "grad_norm": 0.03465070202946663,
211
+ "learning_rate": 0.0004924045652352092,
212
+ "loss": 4.212,
213
+ "step": 290
214
+ },
215
+ {
216
+ "epoch": 0.5504587155963303,
217
+ "grad_norm": 0.029754551127552986,
218
+ "learning_rate": 0.0004620578745374633,
219
+ "loss": 4.2123,
220
+ "step": 300
221
+ },
222
+ {
223
+ "epoch": 0.5688073394495413,
224
+ "grad_norm": 0.03287261724472046,
225
+ "learning_rate": 0.0004318512414382059,
226
+ "loss": 4.221,
227
+ "step": 310
228
+ },
229
+ {
230
+ "epoch": 0.5871559633027523,
231
+ "grad_norm": 0.02530483901500702,
232
+ "learning_rate": 0.0004018961691397148,
233
+ "loss": 4.2115,
234
+ "step": 320
235
+ },
236
+ {
237
+ "epoch": 0.6055045871559633,
238
+ "grad_norm": 0.03148903697729111,
239
+ "learning_rate": 0.0003723032322457458,
240
+ "loss": 4.192,
241
+ "step": 330
242
+ },
243
+ {
244
+ "epoch": 0.6238532110091743,
245
+ "grad_norm": 0.02618285082280636,
246
+ "learning_rate": 0.0003431816685921625,
247
+ "loss": 4.1935,
248
+ "step": 340
249
+ },
250
+ {
251
+ "epoch": 0.6422018348623854,
252
+ "grad_norm": 0.026181507855653763,
253
+ "learning_rate": 0.0003146389760120416,
254
+ "loss": 4.1944,
255
+ "step": 350
256
+ },
257
+ {
258
+ "epoch": 0.6605504587155964,
259
+ "grad_norm": 0.03454752266407013,
260
+ "learning_rate": 0.00028678051552373485,
261
+ "loss": 4.2031,
262
+ "step": 360
263
+ },
264
+ {
265
+ "epoch": 0.6788990825688074,
266
+ "grad_norm": 0.028200663626194,
267
+ "learning_rate": 0.0002597091224066581,
268
+ "loss": 4.1842,
269
+ "step": 370
270
+ },
271
+ {
272
+ "epoch": 0.6972477064220184,
273
+ "grad_norm": 0.02730667218565941,
274
+ "learning_rate": 0.0002335247266004629,
275
+ "loss": 4.1965,
276
+ "step": 380
277
+ },
278
+ {
279
+ "epoch": 0.7155963302752294,
280
+ "grad_norm": 0.028234241530299187,
281
+ "learning_rate": 0.000208323983828828,
282
+ "loss": 4.1779,
283
+ "step": 390
284
+ },
285
+ {
286
+ "epoch": 0.7339449541284404,
287
+ "grad_norm": 0.023867376148700714,
288
+ "learning_rate": 0.0001841999188095224,
289
+ "loss": 4.1809,
290
+ "step": 400
291
+ },
292
+ {
293
+ "epoch": 0.7522935779816514,
294
+ "grad_norm": 0.027158165350556374,
295
+ "learning_rate": 0.00016124158186777676,
296
+ "loss": 4.1797,
297
+ "step": 410
298
+ },
299
+ {
300
+ "epoch": 0.7706422018348624,
301
+ "grad_norm": 0.029092445969581604,
302
+ "learning_rate": 0.0001395337202205211,
303
+ "loss": 4.1686,
304
+ "step": 420
305
+ },
306
+ {
307
+ "epoch": 0.7889908256880734,
308
+ "grad_norm": 0.0232634749263525,
309
+ "learning_rate": 0.00011915646514489292,
310
+ "loss": 4.1756,
311
+ "step": 430
312
+ },
313
+ {
314
+ "epoch": 0.8073394495412844,
315
+ "grad_norm": 0.02425857074558735,
316
+ "learning_rate": 0.00010018503618578818,
317
+ "loss": 4.1831,
318
+ "step": 440
319
+ },
320
+ {
321
+ "epoch": 0.8256880733944955,
322
+ "grad_norm": 0.02306993491947651,
323
+ "learning_rate": 8.268946349432583e-05,
324
+ "loss": 4.1745,
325
+ "step": 450
326
+ },
327
+ {
328
+ "epoch": 0.8440366972477065,
329
+ "grad_norm": 0.02201121672987938,
330
+ "learning_rate": 6.673432932217117e-05,
331
+ "loss": 4.1639,
332
+ "step": 460
333
+ },
334
+ {
335
+ "epoch": 0.8623853211009175,
336
+ "grad_norm": 0.02361287921667099,
337
+ "learning_rate": 5.2378529625954684e-05,
338
+ "loss": 4.1549,
339
+ "step": 470
340
+ },
341
+ {
342
+ "epoch": 0.8807339449541285,
343
+ "grad_norm": 0.020879125222563744,
344
+ "learning_rate": 3.9675056661785556e-05,
345
+ "loss": 4.1644,
346
+ "step": 480
347
+ },
348
+ {
349
+ "epoch": 0.8990825688073395,
350
+ "grad_norm": 0.020139949396252632,
351
+ "learning_rate": 2.8670803372379538e-05,
352
+ "loss": 4.1632,
353
+ "step": 490
354
+ },
355
+ {
356
+ "epoch": 0.9174311926605505,
357
+ "grad_norm": 0.020516378805041313,
358
+ "learning_rate": 1.9406390288876586e-05,
359
+ "loss": 4.1797,
360
+ "step": 500
361
+ },
362
+ {
363
+ "epoch": 0.9357798165137615,
364
+ "grad_norm": 0.019294647499918938,
365
+ "learning_rate": 1.191601558631461e-05,
366
+ "loss": 4.1658,
367
+ "step": 510
368
+ },
369
+ {
370
+ "epoch": 0.9541284403669725,
371
+ "grad_norm": 0.018471471965312958,
372
+ "learning_rate": 6.2273288462559305e-06,
373
+ "loss": 4.1564,
374
+ "step": 520
375
+ },
376
+ {
377
+ "epoch": 0.9724770642201835,
378
+ "grad_norm": 0.020287733525037766,
379
+ "learning_rate": 2.361328992552314e-06,
380
+ "loss": 4.1599,
381
+ "step": 530
382
+ },
383
+ {
384
+ "epoch": 0.9908256880733946,
385
+ "grad_norm": 0.022673217579722404,
386
+ "learning_rate": 3.322867770044202e-07,
387
+ "loss": 4.1669,
388
+ "step": 540
389
+ },
390
+ {
391
+ "epoch": 1.0,
392
+ "step": 545,
393
+ "total_flos": 3.035728330935501e+17,
394
+ "train_loss": 4.405280604931193,
395
+ "train_runtime": 8084.6432,
396
+ "train_samples_per_second": 69.036,
397
+ "train_steps_per_second": 0.067
398
+ }
399
+ ],
400
+ "logging_steps": 10,
401
+ "max_steps": 545,
402
+ "num_input_tokens_seen": 0,
403
+ "num_train_epochs": 1,
404
+ "save_steps": 500,
405
+ "stateful_callbacks": {
406
+ "TrainerControl": {
407
+ "args": {
408
+ "should_epoch_stop": false,
409
+ "should_evaluate": false,
410
+ "should_log": false,
411
+ "should_save": false,
412
+ "should_training_stop": false
413
+ },
414
+ "attributes": {}
415
+ }
416
+ },
417
+ "total_flos": 3.035728330935501e+17,
418
+ "train_batch_size": 8,
419
+ "trial_name": null,
420
+ "trial_params": null
421
+ }
continuous/qwen3/siglip2_sm14_224/finetune/added_tokens.json ADDED
@@ -0,0 +1,28 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "</think>": 151668,
3
+ "</tool_call>": 151658,
4
+ "</tool_response>": 151666,
5
+ "<think>": 151667,
6
+ "<tool_call>": 151657,
7
+ "<tool_response>": 151665,
8
+ "<|box_end|>": 151649,
9
+ "<|box_start|>": 151648,
10
+ "<|endoftext|>": 151643,
11
+ "<|file_sep|>": 151664,
12
+ "<|fim_middle|>": 151660,
13
+ "<|fim_pad|>": 151662,
14
+ "<|fim_prefix|>": 151659,
15
+ "<|fim_suffix|>": 151661,
16
+ "<|im_end|>": 151645,
17
+ "<|im_start|>": 151644,
18
+ "<|image_pad|>": 151655,
19
+ "<|object_ref_end|>": 151647,
20
+ "<|object_ref_start|>": 151646,
21
+ "<|quad_end|>": 151651,
22
+ "<|quad_start|>": 151650,
23
+ "<|repo_name|>": 151663,
24
+ "<|video_pad|>": 151656,
25
+ "<|vision_end|>": 151653,
26
+ "<|vision_pad|>": 151654,
27
+ "<|vision_start|>": 151652
28
+ }
continuous/qwen3/siglip2_sm14_224/finetune/config.json ADDED
@@ -0,0 +1,59 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "add_faster_video": false,
3
+ "add_time_instruction": false,
4
+ "architectures": [
5
+ "LlavaQwen3ForCausalLM"
6
+ ],
7
+ "attention_bias": false,
8
+ "attention_dropout": 0.0,
9
+ "bos_token_id": 151643,
10
+ "eos_token_id": 151645,
11
+ "faster_token_stride": 10,
12
+ "force_sample": false,
13
+ "head_dim": 128,
14
+ "hidden_act": "silu",
15
+ "hidden_size": 2048,
16
+ "image_aspect_ratio": "square",
17
+ "image_crop_resolution": null,
18
+ "image_grid_pinpoints": null,
19
+ "image_split_resolution": null,
20
+ "initializer_range": 0.02,
21
+ "intermediate_size": 6144,
22
+ "max_position_embeddings": 40960,
23
+ "max_window_layers": 28,
24
+ "mm_hidden_size": 1152,
25
+ "mm_newline_position": "grid",
26
+ "mm_patch_merge_type": "flat",
27
+ "mm_projector_lr": 0.0001,
28
+ "mm_projector_type": "mlp2x_gelu",
29
+ "mm_resampler_type": null,
30
+ "mm_spatial_pool_mode": "bilinear",
31
+ "mm_spatial_pool_stride": null,
32
+ "mm_tunable_parts": "mm_mlp_adapter,mm_language_model",
33
+ "mm_use_im_patch_token": false,
34
+ "mm_use_im_start_end": false,
35
+ "mm_vision_select_feature": "patch",
36
+ "mm_vision_select_layer": -2,
37
+ "mm_vision_tower": "open_clip_hub:ViT-SO400M-14-SigLIP2",
38
+ "mm_vision_tower_lr": null,
39
+ "model_type": "qwen3",
40
+ "num_attention_heads": 16,
41
+ "num_hidden_layers": 28,
42
+ "num_key_value_heads": 8,
43
+ "pos_skipping_range": 4096,
44
+ "rms_norm_eps": 1e-06,
45
+ "rope_scaling": null,
46
+ "rope_theta": 1000000,
47
+ "sliding_window": null,
48
+ "tie_word_embeddings": true,
49
+ "tokenizer_model_max_length": 8192,
50
+ "tokenizer_padding_side": "right",
51
+ "torch_dtype": "bfloat16",
52
+ "transformers_version": "4.51.3",
53
+ "use_cache": true,
54
+ "use_mm_proj": true,
55
+ "use_pos_skipping": false,
56
+ "use_sliding_window": false,
57
+ "vision_tower_pretrained": "/cache/ckpt/download/tokenizer/continuous/siglip2/siglip2_sm14_224.npz",
58
+ "vocab_size": 151936
59
+ }
continuous/qwen3/siglip2_sm14_224/finetune/generation_config.json ADDED
@@ -0,0 +1,13 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "bos_token_id": 151643,
3
+ "do_sample": true,
4
+ "eos_token_id": [
5
+ 151645,
6
+ 151643
7
+ ],
8
+ "pad_token_id": 151643,
9
+ "temperature": 0.6,
10
+ "top_k": 20,
11
+ "top_p": 0.95,
12
+ "transformers_version": "4.51.3"
13
+ }
continuous/qwen3/siglip2_sm14_224/finetune/merges.txt ADDED
The diff for this file is too large to render. See raw diff
 
continuous/qwen3/siglip2_sm14_224/finetune/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:884ecd82a6569ab2f29d9500e84bbdb863bcd0a0588302db1e1cc98d8a0dc14a
3
+ size 4945154704
continuous/qwen3/siglip2_sm14_224/finetune/special_tokens_map.json ADDED
@@ -0,0 +1,31 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "additional_special_tokens": [
3
+ "<|im_start|>",
4
+ "<|im_end|>",
5
+ "<|object_ref_start|>",
6
+ "<|object_ref_end|>",
7
+ "<|box_start|>",
8
+ "<|box_end|>",
9
+ "<|quad_start|>",
10
+ "<|quad_end|>",
11
+ "<|vision_start|>",
12
+ "<|vision_end|>",
13
+ "<|vision_pad|>",
14
+ "<|image_pad|>",
15
+ "<|video_pad|>"
16
+ ],
17
+ "eos_token": {
18
+ "content": "<|im_end|>",
19
+ "lstrip": false,
20
+ "normalized": false,
21
+ "rstrip": false,
22
+ "single_word": false
23
+ },
24
+ "pad_token": {
25
+ "content": "<|endoftext|>",
26
+ "lstrip": false,
27
+ "normalized": false,
28
+ "rstrip": false,
29
+ "single_word": false
30
+ }
31
+ }
continuous/qwen3/siglip2_sm14_224/finetune/tokenizer.json ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:aeb13307a71acd8fe81861d94ad54ab689df773318809eed3cbe794b4492dae4
3
+ size 11422654
continuous/qwen3/siglip2_sm14_224/finetune/tokenizer_config.json ADDED
@@ -0,0 +1,241 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "add_bos_token": false,
3
+ "add_prefix_space": false,
4
+ "added_tokens_decoder": {
5
+ "151643": {
6
+ "content": "<|endoftext|>",
7
+ "lstrip": false,
8
+ "normalized": false,
9
+ "rstrip": false,
10
+ "single_word": false,
11
+ "special": true
12
+ },
13
+ "151644": {
14
+ "content": "<|im_start|>",
15
+ "lstrip": false,
16
+ "normalized": false,
17
+ "rstrip": false,
18
+ "single_word": false,
19
+ "special": true
20
+ },
21
+ "151645": {
22
+ "content": "<|im_end|>",
23
+ "lstrip": false,
24
+ "normalized": false,
25
+ "rstrip": false,
26
+ "single_word": false,
27
+ "special": true
28
+ },
29
+ "151646": {
30
+ "content": "<|object_ref_start|>",
31
+ "lstrip": false,
32
+ "normalized": false,
33
+ "rstrip": false,
34
+ "single_word": false,
35
+ "special": true
36
+ },
37
+ "151647": {
38
+ "content": "<|object_ref_end|>",
39
+ "lstrip": false,
40
+ "normalized": false,
41
+ "rstrip": false,
42
+ "single_word": false,
43
+ "special": true
44
+ },
45
+ "151648": {
46
+ "content": "<|box_start|>",
47
+ "lstrip": false,
48
+ "normalized": false,
49
+ "rstrip": false,
50
+ "single_word": false,
51
+ "special": true
52
+ },
53
+ "151649": {
54
+ "content": "<|box_end|>",
55
+ "lstrip": false,
56
+ "normalized": false,
57
+ "rstrip": false,
58
+ "single_word": false,
59
+ "special": true
60
+ },
61
+ "151650": {
62
+ "content": "<|quad_start|>",
63
+ "lstrip": false,
64
+ "normalized": false,
65
+ "rstrip": false,
66
+ "single_word": false,
67
+ "special": true
68
+ },
69
+ "151651": {
70
+ "content": "<|quad_end|>",
71
+ "lstrip": false,
72
+ "normalized": false,
73
+ "rstrip": false,
74
+ "single_word": false,
75
+ "special": true
76
+ },
77
+ "151652": {
78
+ "content": "<|vision_start|>",
79
+ "lstrip": false,
80
+ "normalized": false,
81
+ "rstrip": false,
82
+ "single_word": false,
83
+ "special": true
84
+ },
85
+ "151653": {
86
+ "content": "<|vision_end|>",
87
+ "lstrip": false,
88
+ "normalized": false,
89
+ "rstrip": false,
90
+ "single_word": false,
91
+ "special": true
92
+ },
93
+ "151654": {
94
+ "content": "<|vision_pad|>",
95
+ "lstrip": false,
96
+ "normalized": false,
97
+ "rstrip": false,
98
+ "single_word": false,
99
+ "special": true
100
+ },
101
+ "151655": {
102
+ "content": "<|image_pad|>",
103
+ "lstrip": false,
104
+ "normalized": false,
105
+ "rstrip": false,
106
+ "single_word": false,
107
+ "special": true
108
+ },
109
+ "151656": {
110
+ "content": "<|video_pad|>",
111
+ "lstrip": false,
112
+ "normalized": false,
113
+ "rstrip": false,
114
+ "single_word": false,
115
+ "special": true
116
+ },
117
+ "151657": {
118
+ "content": "<tool_call>",
119
+ "lstrip": false,
120
+ "normalized": false,
121
+ "rstrip": false,
122
+ "single_word": false,
123
+ "special": false
124
+ },
125
+ "151658": {
126
+ "content": "</tool_call>",
127
+ "lstrip": false,
128
+ "normalized": false,
129
+ "rstrip": false,
130
+ "single_word": false,
131
+ "special": false
132
+ },
133
+ "151659": {
134
+ "content": "<|fim_prefix|>",
135
+ "lstrip": false,
136
+ "normalized": false,
137
+ "rstrip": false,
138
+ "single_word": false,
139
+ "special": false
140
+ },
141
+ "151660": {
142
+ "content": "<|fim_middle|>",
143
+ "lstrip": false,
144
+ "normalized": false,
145
+ "rstrip": false,
146
+ "single_word": false,
147
+ "special": false
148
+ },
149
+ "151661": {
150
+ "content": "<|fim_suffix|>",
151
+ "lstrip": false,
152
+ "normalized": false,
153
+ "rstrip": false,
154
+ "single_word": false,
155
+ "special": false
156
+ },
157
+ "151662": {
158
+ "content": "<|fim_pad|>",
159
+ "lstrip": false,
160
+ "normalized": false,
161
+ "rstrip": false,
162
+ "single_word": false,
163
+ "special": false
164
+ },
165
+ "151663": {
166
+ "content": "<|repo_name|>",
167
+ "lstrip": false,
168
+ "normalized": false,
169
+ "rstrip": false,
170
+ "single_word": false,
171
+ "special": false
172
+ },
173
+ "151664": {
174
+ "content": "<|file_sep|>",
175
+ "lstrip": false,
176
+ "normalized": false,
177
+ "rstrip": false,
178
+ "single_word": false,
179
+ "special": false
180
+ },
181
+ "151665": {
182
+ "content": "<tool_response>",
183
+ "lstrip": false,
184
+ "normalized": false,
185
+ "rstrip": false,
186
+ "single_word": false,
187
+ "special": false
188
+ },
189
+ "151666": {
190
+ "content": "</tool_response>",
191
+ "lstrip": false,
192
+ "normalized": false,
193
+ "rstrip": false,
194
+ "single_word": false,
195
+ "special": false
196
+ },
197
+ "151667": {
198
+ "content": "<think>",
199
+ "lstrip": false,
200
+ "normalized": false,
201
+ "rstrip": false,
202
+ "single_word": false,
203
+ "special": false
204
+ },
205
+ "151668": {
206
+ "content": "</think>",
207
+ "lstrip": false,
208
+ "normalized": false,
209
+ "rstrip": false,
210
+ "single_word": false,
211
+ "special": false
212
+ }
213
+ },
214
+ "additional_special_tokens": [
215
+ "<|im_start|>",
216
+ "<|im_end|>",
217
+ "<|object_ref_start|>",
218
+ "<|object_ref_end|>",
219
+ "<|box_start|>",
220
+ "<|box_end|>",
221
+ "<|quad_start|>",
222
+ "<|quad_end|>",
223
+ "<|vision_start|>",
224
+ "<|vision_end|>",
225
+ "<|vision_pad|>",
226
+ "<|image_pad|>",
227
+ "<|video_pad|>"
228
+ ],
229
+ "bos_token": null,
230
+ "chat_template": "{%- if tools %}\n {{- '<|im_start|>system\\n' }}\n {%- if messages[0].role == 'system' %}\n {{- messages[0].content + '\\n\\n' }}\n {%- endif %}\n {{- \"# Tools\\n\\nYou may call one or more functions to assist with the user query.\\n\\nYou are provided with function signatures within <tools></tools> XML tags:\\n<tools>\" }}\n {%- for tool in tools %}\n {{- \"\\n\" }}\n {{- tool | tojson }}\n {%- endfor %}\n {{- \"\\n</tools>\\n\\nFor each function call, return a json object with function name and arguments within <tool_call></tool_call> XML tags:\\n<tool_call>\\n{\\\"name\\\": <function-name>, \\\"arguments\\\": <args-json-object>}\\n</tool_call><|im_end|>\\n\" }}\n{%- else %}\n {%- if messages[0].role == 'system' %}\n {{- '<|im_start|>system\\n' + messages[0].content + '<|im_end|>\\n' }}\n {%- endif %}\n{%- endif %}\n{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}\n{%- for message in messages[::-1] %}\n {%- set index = (messages|length - 1) - loop.index0 %}\n {%- if ns.multi_step_tool and message.role == \"user\" and message.content is string and not(message.content.startswith('<tool_response>') and message.content.endswith('</tool_response>')) %}\n {%- set ns.multi_step_tool = false %}\n {%- set ns.last_query_index = index %}\n {%- endif %}\n{%- endfor %}\n{%- for message in messages %}\n {%- if message.content is string %}\n {%- set content = message.content %}\n {%- else %}\n {%- set content = '' %}\n {%- endif %}\n {%- if (message.role == \"user\") or (message.role == \"system\" and not loop.first) %}\n {{- '<|im_start|>' + message.role + '\\n' + content + '<|im_end|>' + '\\n' }}\n {%- elif message.role == \"assistant\" %}\n {%- set reasoning_content = '' %}\n {%- if message.reasoning_content is string %}\n {%- set reasoning_content = message.reasoning_content %}\n {%- else %}\n {%- if '</think>' in content %}\n {%- set reasoning_content = content.split('</think>')[0].rstrip('\\n').split('<think>')[-1].lstrip('\\n') %}\n {%- set content = content.split('</think>')[-1].lstrip('\\n') %}\n {%- endif %}\n {%- endif %}\n {%- if loop.index0 > ns.last_query_index %}\n {%- if loop.last or (not loop.last and reasoning_content) %}\n {{- '<|im_start|>' + message.role + '\\n<think>\\n' + reasoning_content.strip('\\n') + '\\n</think>\\n\\n' + content.lstrip('\\n') }}\n {%- else %}\n {{- '<|im_start|>' + message.role + '\\n' + content }}\n {%- endif %}\n {%- else %}\n {{- '<|im_start|>' + message.role + '\\n' + content }}\n {%- endif %}\n {%- if message.tool_calls %}\n {%- for tool_call in message.tool_calls %}\n {%- if (loop.first and content) or (not loop.first) %}\n {{- '\\n' }}\n {%- endif %}\n {%- if tool_call.function %}\n {%- set tool_call = tool_call.function %}\n {%- endif %}\n {{- '<tool_call>\\n{\"name\": \"' }}\n {{- tool_call.name }}\n {{- '\", \"arguments\": ' }}\n {%- if tool_call.arguments is string %}\n {{- tool_call.arguments }}\n {%- else %}\n {{- tool_call.arguments | tojson }}\n {%- endif %}\n {{- '}\\n</tool_call>' }}\n {%- endfor %}\n {%- endif %}\n {{- '<|im_end|>\\n' }}\n {%- elif message.role == \"tool\" %}\n {%- if loop.first or (messages[loop.index0 - 1].role != \"tool\") %}\n {{- '<|im_start|>user' }}\n {%- endif %}\n {{- '\\n<tool_response>\\n' }}\n {{- content }}\n {{- '\\n</tool_response>' }}\n {%- if loop.last or (messages[loop.index0 + 1].role != \"tool\") %}\n {{- '<|im_end|>\\n' }}\n {%- endif %}\n {%- endif %}\n{%- endfor %}\n{%- if add_generation_prompt %}\n {{- '<|im_start|>assistant\\n' }}\n {%- if enable_thinking is defined and enable_thinking is false %}\n {{- '<think>\\n\\n</think>\\n\\n' }}\n {%- endif %}\n{%- endif %}",
231
+ "clean_up_tokenization_spaces": false,
232
+ "eos_token": "<|im_end|>",
233
+ "errors": "replace",
234
+ "extra_special_tokens": {},
235
+ "model_max_length": 8192,
236
+ "pad_token": "<|endoftext|>",
237
+ "padding_side": "right",
238
+ "split_special_tokens": false,
239
+ "tokenizer_class": "Qwen2Tokenizer",
240
+ "unk_token": null
241
+ }
continuous/qwen3/siglip2_sm14_224/finetune/trainer_state.json ADDED
@@ -0,0 +1,1856 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_global_step": null,
3
+ "best_metric": null,
4
+ "best_model_checkpoint": null,
5
+ "epoch": 0.9999037443449803,
6
+ "eval_steps": 500,
7
+ "global_step": 2597,
8
+ "is_hyper_param_search": false,
9
+ "is_local_process_zero": true,
10
+ "is_world_process_zero": true,
11
+ "log_history": [
12
+ {
13
+ "epoch": 0.003850226200789296,
14
+ "grad_norm": 20.822731018066406,
15
+ "learning_rate": 1.153846153846154e-06,
16
+ "loss": 2.5623,
17
+ "step": 10
18
+ },
19
+ {
20
+ "epoch": 0.007700452401578592,
21
+ "grad_norm": 8.860735893249512,
22
+ "learning_rate": 2.435897435897436e-06,
23
+ "loss": 1.9749,
24
+ "step": 20
25
+ },
26
+ {
27
+ "epoch": 0.01155067860236789,
28
+ "grad_norm": 4.52546501159668,
29
+ "learning_rate": 3.7179487179487184e-06,
30
+ "loss": 1.4603,
31
+ "step": 30
32
+ },
33
+ {
34
+ "epoch": 0.015400904803157185,
35
+ "grad_norm": 1.5605366230010986,
36
+ "learning_rate": 5e-06,
37
+ "loss": 1.122,
38
+ "step": 40
39
+ },
40
+ {
41
+ "epoch": 0.019251131003946483,
42
+ "grad_norm": 0.8273071646690369,
43
+ "learning_rate": 6.282051282051282e-06,
44
+ "loss": 1.0098,
45
+ "step": 50
46
+ },
47
+ {
48
+ "epoch": 0.02310135720473578,
49
+ "grad_norm": 0.4240211248397827,
50
+ "learning_rate": 7.564102564102564e-06,
51
+ "loss": 0.9227,
52
+ "step": 60
53
+ },
54
+ {
55
+ "epoch": 0.026951583405525073,
56
+ "grad_norm": 0.31568580865859985,
57
+ "learning_rate": 8.846153846153847e-06,
58
+ "loss": 0.8772,
59
+ "step": 70
60
+ },
61
+ {
62
+ "epoch": 0.03080180960631437,
63
+ "grad_norm": 0.30856403708457947,
64
+ "learning_rate": 9.999996111488723e-06,
65
+ "loss": 0.8745,
66
+ "step": 80
67
+ },
68
+ {
69
+ "epoch": 0.034652035807103666,
70
+ "grad_norm": 0.2912237346172333,
71
+ "learning_rate": 9.999529497453782e-06,
72
+ "loss": 0.8452,
73
+ "step": 90
74
+ },
75
+ {
76
+ "epoch": 0.038502262007892966,
77
+ "grad_norm": 0.26034387946128845,
78
+ "learning_rate": 9.998285264324193e-06,
79
+ "loss": 0.818,
80
+ "step": 100
81
+ },
82
+ {
83
+ "epoch": 0.04235248820868226,
84
+ "grad_norm": 0.26798075437545776,
85
+ "learning_rate": 9.996263605626052e-06,
86
+ "loss": 0.806,
87
+ "step": 110
88
+ },
89
+ {
90
+ "epoch": 0.04620271440947156,
91
+ "grad_norm": 0.2721961736679077,
92
+ "learning_rate": 9.99346483580503e-06,
93
+ "loss": 0.815,
94
+ "step": 120
95
+ },
96
+ {
97
+ "epoch": 0.05005294061026085,
98
+ "grad_norm": 0.26553991436958313,
99
+ "learning_rate": 9.989889390177467e-06,
100
+ "loss": 0.8083,
101
+ "step": 130
102
+ },
103
+ {
104
+ "epoch": 0.053903166811050146,
105
+ "grad_norm": 0.30027005076408386,
106
+ "learning_rate": 9.985537824862647e-06,
107
+ "loss": 0.8067,
108
+ "step": 140
109
+ },
110
+ {
111
+ "epoch": 0.057753393011839446,
112
+ "grad_norm": 0.2986396849155426,
113
+ "learning_rate": 9.980410816696317e-06,
114
+ "loss": 0.7881,
115
+ "step": 150
116
+ },
117
+ {
118
+ "epoch": 0.06160361921262874,
119
+ "grad_norm": 0.27153223752975464,
120
+ "learning_rate": 9.97450916312541e-06,
121
+ "loss": 0.7979,
122
+ "step": 160
123
+ },
124
+ {
125
+ "epoch": 0.06545384541341803,
126
+ "grad_norm": 0.24366115033626556,
127
+ "learning_rate": 9.967833782084004e-06,
128
+ "loss": 0.7993,
129
+ "step": 170
130
+ },
131
+ {
132
+ "epoch": 0.06930407161420733,
133
+ "grad_norm": 0.28412872552871704,
134
+ "learning_rate": 9.960385711850554e-06,
135
+ "loss": 0.8227,
136
+ "step": 180
137
+ },
138
+ {
139
+ "epoch": 0.07315429781499663,
140
+ "grad_norm": 0.2919147312641144,
141
+ "learning_rate": 9.952166110886394e-06,
142
+ "loss": 0.8141,
143
+ "step": 190
144
+ },
145
+ {
146
+ "epoch": 0.07700452401578593,
147
+ "grad_norm": 0.2793526351451874,
148
+ "learning_rate": 9.943176257655567e-06,
149
+ "loss": 0.8006,
150
+ "step": 200
151
+ },
152
+ {
153
+ "epoch": 0.08085475021657522,
154
+ "grad_norm": 0.2634764611721039,
155
+ "learning_rate": 9.933417550425952e-06,
156
+ "loss": 0.8028,
157
+ "step": 210
158
+ },
159
+ {
160
+ "epoch": 0.08470497641736452,
161
+ "grad_norm": 0.28198230266571045,
162
+ "learning_rate": 9.9228915070518e-06,
163
+ "loss": 0.7819,
164
+ "step": 220
165
+ },
166
+ {
167
+ "epoch": 0.08855520261815382,
168
+ "grad_norm": 0.25833502411842346,
169
+ "learning_rate": 9.911599764737632e-06,
170
+ "loss": 0.7829,
171
+ "step": 230
172
+ },
173
+ {
174
+ "epoch": 0.09240542881894312,
175
+ "grad_norm": 0.2708854377269745,
176
+ "learning_rate": 9.899544079783606e-06,
177
+ "loss": 0.7802,
178
+ "step": 240
179
+ },
180
+ {
181
+ "epoch": 0.0962556550197324,
182
+ "grad_norm": 0.30411165952682495,
183
+ "learning_rate": 9.886726327312338e-06,
184
+ "loss": 0.7987,
185
+ "step": 250
186
+ },
187
+ {
188
+ "epoch": 0.1001058812205217,
189
+ "grad_norm": 0.2554266154766083,
190
+ "learning_rate": 9.87314850097724e-06,
191
+ "loss": 0.7906,
192
+ "step": 260
193
+ },
194
+ {
195
+ "epoch": 0.103956107421311,
196
+ "grad_norm": 0.24927367269992828,
197
+ "learning_rate": 9.858812712652452e-06,
198
+ "loss": 0.7799,
199
+ "step": 270
200
+ },
201
+ {
202
+ "epoch": 0.10780633362210029,
203
+ "grad_norm": 0.24568291008472443,
204
+ "learning_rate": 9.843721192104341e-06,
205
+ "loss": 0.7781,
206
+ "step": 280
207
+ },
208
+ {
209
+ "epoch": 0.11165655982288959,
210
+ "grad_norm": 0.2502061128616333,
211
+ "learning_rate": 9.827876286644698e-06,
212
+ "loss": 0.7978,
213
+ "step": 290
214
+ },
215
+ {
216
+ "epoch": 0.11550678602367889,
217
+ "grad_norm": 0.23090818524360657,
218
+ "learning_rate": 9.811280460765643e-06,
219
+ "loss": 0.7596,
220
+ "step": 300
221
+ },
222
+ {
223
+ "epoch": 0.11935701222446819,
224
+ "grad_norm": 0.23904722929000854,
225
+ "learning_rate": 9.793936295756292e-06,
226
+ "loss": 0.7689,
227
+ "step": 310
228
+ },
229
+ {
230
+ "epoch": 0.12320723842525748,
231
+ "grad_norm": 0.2455524206161499,
232
+ "learning_rate": 9.775846489301279e-06,
233
+ "loss": 0.7754,
234
+ "step": 320
235
+ },
236
+ {
237
+ "epoch": 0.12705746462604678,
238
+ "grad_norm": 0.26052621006965637,
239
+ "learning_rate": 9.757013855061157e-06,
240
+ "loss": 0.7722,
241
+ "step": 330
242
+ },
243
+ {
244
+ "epoch": 0.13090769082683607,
245
+ "grad_norm": 0.2839418947696686,
246
+ "learning_rate": 9.737441322234758e-06,
247
+ "loss": 0.7827,
248
+ "step": 340
249
+ },
250
+ {
251
+ "epoch": 0.13475791702762538,
252
+ "grad_norm": 0.2858501374721527,
253
+ "learning_rate": 9.717131935103605e-06,
254
+ "loss": 0.7846,
255
+ "step": 350
256
+ },
257
+ {
258
+ "epoch": 0.13860814322841467,
259
+ "grad_norm": 0.25717398524284363,
260
+ "learning_rate": 9.696088852558398e-06,
261
+ "loss": 0.7713,
262
+ "step": 360
263
+ },
264
+ {
265
+ "epoch": 0.14245836942920398,
266
+ "grad_norm": 0.2643663287162781,
267
+ "learning_rate": 9.674315347607686e-06,
268
+ "loss": 0.7995,
269
+ "step": 370
270
+ },
271
+ {
272
+ "epoch": 0.14630859562999327,
273
+ "grad_norm": 0.2540455758571625,
274
+ "learning_rate": 9.651814806868796e-06,
275
+ "loss": 0.7593,
276
+ "step": 380
277
+ },
278
+ {
279
+ "epoch": 0.15015882183078255,
280
+ "grad_norm": 0.2415592223405838,
281
+ "learning_rate": 9.62859073004107e-06,
282
+ "loss": 0.7615,
283
+ "step": 390
284
+ },
285
+ {
286
+ "epoch": 0.15400904803157187,
287
+ "grad_norm": 0.24444164335727692,
288
+ "learning_rate": 9.604646729361547e-06,
289
+ "loss": 0.7611,
290
+ "step": 400
291
+ },
292
+ {
293
+ "epoch": 0.15785927423236115,
294
+ "grad_norm": 0.2445029467344284,
295
+ "learning_rate": 9.579986529043099e-06,
296
+ "loss": 0.7504,
297
+ "step": 410
298
+ },
299
+ {
300
+ "epoch": 0.16170950043315044,
301
+ "grad_norm": 0.24462901055812836,
302
+ "learning_rate": 9.554613964695189e-06,
303
+ "loss": 0.7743,
304
+ "step": 420
305
+ },
306
+ {
307
+ "epoch": 0.16555972663393975,
308
+ "grad_norm": 0.2832423746585846,
309
+ "learning_rate": 9.528532982727281e-06,
310
+ "loss": 0.7659,
311
+ "step": 430
312
+ },
313
+ {
314
+ "epoch": 0.16940995283472904,
315
+ "grad_norm": 0.25782066583633423,
316
+ "learning_rate": 9.501747639735018e-06,
317
+ "loss": 0.7577,
318
+ "step": 440
319
+ },
320
+ {
321
+ "epoch": 0.17326017903551832,
322
+ "grad_norm": 0.2520759403705597,
323
+ "learning_rate": 9.474262101869272e-06,
324
+ "loss": 0.749,
325
+ "step": 450
326
+ },
327
+ {
328
+ "epoch": 0.17711040523630764,
329
+ "grad_norm": 0.2584396302700043,
330
+ "learning_rate": 9.446080644188141e-06,
331
+ "loss": 0.7567,
332
+ "step": 460
333
+ },
334
+ {
335
+ "epoch": 0.18096063143709692,
336
+ "grad_norm": 0.2432314157485962,
337
+ "learning_rate": 9.417207649992017e-06,
338
+ "loss": 0.7596,
339
+ "step": 470
340
+ },
341
+ {
342
+ "epoch": 0.18481085763788624,
343
+ "grad_norm": 0.2489367127418518,
344
+ "learning_rate": 9.387647610141816e-06,
345
+ "loss": 0.757,
346
+ "step": 480
347
+ },
348
+ {
349
+ "epoch": 0.18866108383867553,
350
+ "grad_norm": 0.23927755653858185,
351
+ "learning_rate": 9.35740512236047e-06,
352
+ "loss": 0.7472,
353
+ "step": 490
354
+ },
355
+ {
356
+ "epoch": 0.1925113100394648,
357
+ "grad_norm": 0.24385707080364227,
358
+ "learning_rate": 9.326484890517809e-06,
359
+ "loss": 0.7569,
360
+ "step": 500
361
+ },
362
+ {
363
+ "epoch": 0.19636153624025413,
364
+ "grad_norm": 0.24530303478240967,
365
+ "learning_rate": 9.294891723898926e-06,
366
+ "loss": 0.7466,
367
+ "step": 510
368
+ },
369
+ {
370
+ "epoch": 0.2002117624410434,
371
+ "grad_norm": 0.2543795108795166,
372
+ "learning_rate": 9.262630536456153e-06,
373
+ "loss": 0.7773,
374
+ "step": 520
375
+ },
376
+ {
377
+ "epoch": 0.2040619886418327,
378
+ "grad_norm": 0.26317375898361206,
379
+ "learning_rate": 9.229706346044749e-06,
380
+ "loss": 0.7506,
381
+ "step": 530
382
+ },
383
+ {
384
+ "epoch": 0.207912214842622,
385
+ "grad_norm": 0.25899821519851685,
386
+ "learning_rate": 9.196124273642422e-06,
387
+ "loss": 0.7616,
388
+ "step": 540
389
+ },
390
+ {
391
+ "epoch": 0.2117624410434113,
392
+ "grad_norm": 0.23583026230335236,
393
+ "learning_rate": 9.161889542552842e-06,
394
+ "loss": 0.7704,
395
+ "step": 550
396
+ },
397
+ {
398
+ "epoch": 0.21561266724420058,
399
+ "grad_norm": 0.22977587580680847,
400
+ "learning_rate": 9.127007477593193e-06,
401
+ "loss": 0.7687,
402
+ "step": 560
403
+ },
404
+ {
405
+ "epoch": 0.2194628934449899,
406
+ "grad_norm": 0.28237593173980713,
407
+ "learning_rate": 9.091483504265974e-06,
408
+ "loss": 0.7711,
409
+ "step": 570
410
+ },
411
+ {
412
+ "epoch": 0.22331311964577918,
413
+ "grad_norm": 0.25540873408317566,
414
+ "learning_rate": 9.055323147915116e-06,
415
+ "loss": 0.7572,
416
+ "step": 580
417
+ },
418
+ {
419
+ "epoch": 0.2271633458465685,
420
+ "grad_norm": 0.2665945291519165,
421
+ "learning_rate": 9.018532032866585e-06,
422
+ "loss": 0.763,
423
+ "step": 590
424
+ },
425
+ {
426
+ "epoch": 0.23101357204735778,
427
+ "grad_norm": 0.23430664837360382,
428
+ "learning_rate": 8.981115881553579e-06,
429
+ "loss": 0.7561,
430
+ "step": 600
431
+ },
432
+ {
433
+ "epoch": 0.23486379824814707,
434
+ "grad_norm": 0.2692226469516754,
435
+ "learning_rate": 8.943080513626476e-06,
436
+ "loss": 0.7755,
437
+ "step": 610
438
+ },
439
+ {
440
+ "epoch": 0.23871402444893638,
441
+ "grad_norm": 0.24797184765338898,
442
+ "learning_rate": 8.904431845047641e-06,
443
+ "loss": 0.7626,
444
+ "step": 620
445
+ },
446
+ {
447
+ "epoch": 0.24256425064972567,
448
+ "grad_norm": 0.25342825055122375,
449
+ "learning_rate": 8.86517588717128e-06,
450
+ "loss": 0.7775,
451
+ "step": 630
452
+ },
453
+ {
454
+ "epoch": 0.24641447685051496,
455
+ "grad_norm": 0.23582082986831665,
456
+ "learning_rate": 8.82531874580844e-06,
457
+ "loss": 0.7737,
458
+ "step": 640
459
+ },
460
+ {
461
+ "epoch": 0.25026470305130427,
462
+ "grad_norm": 0.23786808550357819,
463
+ "learning_rate": 8.784866620277314e-06,
464
+ "loss": 0.7274,
465
+ "step": 650
466
+ },
467
+ {
468
+ "epoch": 0.25411492925209356,
469
+ "grad_norm": 0.2628067433834076,
470
+ "learning_rate": 8.743825802439018e-06,
471
+ "loss": 0.7564,
472
+ "step": 660
473
+ },
474
+ {
475
+ "epoch": 0.25796515545288284,
476
+ "grad_norm": 0.2781405448913574,
477
+ "learning_rate": 8.702202675718956e-06,
478
+ "loss": 0.7692,
479
+ "step": 670
480
+ },
481
+ {
482
+ "epoch": 0.26181538165367213,
483
+ "grad_norm": 0.226299449801445,
484
+ "learning_rate": 8.660003714113959e-06,
485
+ "loss": 0.734,
486
+ "step": 680
487
+ },
488
+ {
489
+ "epoch": 0.26566560785446147,
490
+ "grad_norm": 0.23770830035209656,
491
+ "learning_rate": 8.617235481185323e-06,
492
+ "loss": 0.7574,
493
+ "step": 690
494
+ },
495
+ {
496
+ "epoch": 0.26951583405525076,
497
+ "grad_norm": 0.2601287364959717,
498
+ "learning_rate": 8.573904629037931e-06,
499
+ "loss": 0.7659,
500
+ "step": 700
501
+ },
502
+ {
503
+ "epoch": 0.27336606025604004,
504
+ "grad_norm": 0.2455611526966095,
505
+ "learning_rate": 8.530017897285585e-06,
506
+ "loss": 0.7493,
507
+ "step": 710
508
+ },
509
+ {
510
+ "epoch": 0.27721628645682933,
511
+ "grad_norm": 0.22639057040214539,
512
+ "learning_rate": 8.485582112002746e-06,
513
+ "loss": 0.7554,
514
+ "step": 720
515
+ },
516
+ {
517
+ "epoch": 0.2810665126576186,
518
+ "grad_norm": 0.24754133820533752,
519
+ "learning_rate": 8.440604184662811e-06,
520
+ "loss": 0.7544,
521
+ "step": 730
522
+ },
523
+ {
524
+ "epoch": 0.28491673885840796,
525
+ "grad_norm": 0.2363588511943817,
526
+ "learning_rate": 8.395091111063118e-06,
527
+ "loss": 0.7425,
528
+ "step": 740
529
+ },
530
+ {
531
+ "epoch": 0.28876696505919724,
532
+ "grad_norm": 0.22622808814048767,
533
+ "learning_rate": 8.349049970236822e-06,
534
+ "loss": 0.7455,
535
+ "step": 750
536
+ },
537
+ {
538
+ "epoch": 0.29261719125998653,
539
+ "grad_norm": 0.2666724622249603,
540
+ "learning_rate": 8.302487923351842e-06,
541
+ "loss": 0.7374,
542
+ "step": 760
543
+ },
544
+ {
545
+ "epoch": 0.2964674174607758,
546
+ "grad_norm": 0.24705751240253448,
547
+ "learning_rate": 8.255412212597021e-06,
548
+ "loss": 0.7566,
549
+ "step": 770
550
+ },
551
+ {
552
+ "epoch": 0.3003176436615651,
553
+ "grad_norm": 0.24594572186470032,
554
+ "learning_rate": 8.207830160055688e-06,
555
+ "loss": 0.7525,
556
+ "step": 780
557
+ },
558
+ {
559
+ "epoch": 0.3041678698623544,
560
+ "grad_norm": 0.24235892295837402,
561
+ "learning_rate": 8.159749166566786e-06,
562
+ "loss": 0.7354,
563
+ "step": 790
564
+ },
565
+ {
566
+ "epoch": 0.30801809606314373,
567
+ "grad_norm": 0.23181109130382538,
568
+ "learning_rate": 8.11117671057377e-06,
569
+ "loss": 0.76,
570
+ "step": 800
571
+ },
572
+ {
573
+ "epoch": 0.311868322263933,
574
+ "grad_norm": 0.2528509199619293,
575
+ "learning_rate": 8.062120346961404e-06,
576
+ "loss": 0.7512,
577
+ "step": 810
578
+ },
579
+ {
580
+ "epoch": 0.3157185484647223,
581
+ "grad_norm": 0.2378137856721878,
582
+ "learning_rate": 8.012587705880705e-06,
583
+ "loss": 0.7454,
584
+ "step": 820
585
+ },
586
+ {
587
+ "epoch": 0.3195687746655116,
588
+ "grad_norm": 0.24435830116271973,
589
+ "learning_rate": 7.962586491562146e-06,
590
+ "loss": 0.7328,
591
+ "step": 830
592
+ },
593
+ {
594
+ "epoch": 0.3234190008663009,
595
+ "grad_norm": 0.24372529983520508,
596
+ "learning_rate": 7.912124481117357e-06,
597
+ "loss": 0.7377,
598
+ "step": 840
599
+ },
600
+ {
601
+ "epoch": 0.3272692270670902,
602
+ "grad_norm": 0.23847636580467224,
603
+ "learning_rate": 7.861209523329494e-06,
604
+ "loss": 0.7544,
605
+ "step": 850
606
+ },
607
+ {
608
+ "epoch": 0.3311194532678795,
609
+ "grad_norm": 0.23524430394172668,
610
+ "learning_rate": 7.809849537432432e-06,
611
+ "loss": 0.7323,
612
+ "step": 860
613
+ },
614
+ {
615
+ "epoch": 0.3349696794686688,
616
+ "grad_norm": 0.2379012405872345,
617
+ "learning_rate": 7.758052511879035e-06,
618
+ "loss": 0.744,
619
+ "step": 870
620
+ },
621
+ {
622
+ "epoch": 0.3388199056694581,
623
+ "grad_norm": 0.23893901705741882,
624
+ "learning_rate": 7.705826503098642e-06,
625
+ "loss": 0.7376,
626
+ "step": 880
627
+ },
628
+ {
629
+ "epoch": 0.34267013187024736,
630
+ "grad_norm": 0.23630915582180023,
631
+ "learning_rate": 7.653179634243979e-06,
632
+ "loss": 0.7423,
633
+ "step": 890
634
+ },
635
+ {
636
+ "epoch": 0.34652035807103665,
637
+ "grad_norm": 0.22671861946582794,
638
+ "learning_rate": 7.600120093927696e-06,
639
+ "loss": 0.7404,
640
+ "step": 900
641
+ },
642
+ {
643
+ "epoch": 0.350370584271826,
644
+ "grad_norm": 0.2483118772506714,
645
+ "learning_rate": 7.546656134948729e-06,
646
+ "loss": 0.7304,
647
+ "step": 910
648
+ },
649
+ {
650
+ "epoch": 0.3542208104726153,
651
+ "grad_norm": 0.2507372796535492,
652
+ "learning_rate": 7.492796073008661e-06,
653
+ "loss": 0.7332,
654
+ "step": 920
655
+ },
656
+ {
657
+ "epoch": 0.35807103667340456,
658
+ "grad_norm": 0.2523139417171478,
659
+ "learning_rate": 7.438548285418326e-06,
660
+ "loss": 0.7182,
661
+ "step": 930
662
+ },
663
+ {
664
+ "epoch": 0.36192126287419385,
665
+ "grad_norm": 0.2668459415435791,
666
+ "learning_rate": 7.3839212097948e-06,
667
+ "loss": 0.7535,
668
+ "step": 940
669
+ },
670
+ {
671
+ "epoch": 0.36577148907498314,
672
+ "grad_norm": 0.24314750730991364,
673
+ "learning_rate": 7.328923342749039e-06,
674
+ "loss": 0.7714,
675
+ "step": 950
676
+ },
677
+ {
678
+ "epoch": 0.3696217152757725,
679
+ "grad_norm": 0.25559529662132263,
680
+ "learning_rate": 7.273563238564319e-06,
681
+ "loss": 0.7531,
682
+ "step": 960
683
+ },
684
+ {
685
+ "epoch": 0.37347194147656176,
686
+ "grad_norm": 0.2461807280778885,
687
+ "learning_rate": 7.217849507865724e-06,
688
+ "loss": 0.7284,
689
+ "step": 970
690
+ },
691
+ {
692
+ "epoch": 0.37732216767735105,
693
+ "grad_norm": 0.24866263568401337,
694
+ "learning_rate": 7.161790816280858e-06,
695
+ "loss": 0.7418,
696
+ "step": 980
697
+ },
698
+ {
699
+ "epoch": 0.38117239387814034,
700
+ "grad_norm": 0.24930378794670105,
701
+ "learning_rate": 7.1053958830920125e-06,
702
+ "loss": 0.7171,
703
+ "step": 990
704
+ },
705
+ {
706
+ "epoch": 0.3850226200789296,
707
+ "grad_norm": 0.2504958510398865,
708
+ "learning_rate": 7.0486734798799715e-06,
709
+ "loss": 0.7414,
710
+ "step": 1000
711
+ },
712
+ {
713
+ "epoch": 0.3888728462797189,
714
+ "grad_norm": 0.23553283512592316,
715
+ "learning_rate": 6.991632429159708e-06,
716
+ "loss": 0.7654,
717
+ "step": 1010
718
+ },
719
+ {
720
+ "epoch": 0.39272307248050825,
721
+ "grad_norm": 0.26473990082740784,
722
+ "learning_rate": 6.934281603008132e-06,
723
+ "loss": 0.7507,
724
+ "step": 1020
725
+ },
726
+ {
727
+ "epoch": 0.39657329868129754,
728
+ "grad_norm": 0.24581706523895264,
729
+ "learning_rate": 6.876629921684143e-06,
730
+ "loss": 0.7509,
731
+ "step": 1030
732
+ },
733
+ {
734
+ "epoch": 0.4004235248820868,
735
+ "grad_norm": 0.2467559278011322,
736
+ "learning_rate": 6.818686352241199e-06,
737
+ "loss": 0.7386,
738
+ "step": 1040
739
+ },
740
+ {
741
+ "epoch": 0.4042737510828761,
742
+ "grad_norm": 0.2161765843629837,
743
+ "learning_rate": 6.760459907132573e-06,
744
+ "loss": 0.7364,
745
+ "step": 1050
746
+ },
747
+ {
748
+ "epoch": 0.4081239772836654,
749
+ "grad_norm": 0.23059023916721344,
750
+ "learning_rate": 6.701959642809589e-06,
751
+ "loss": 0.7405,
752
+ "step": 1060
753
+ },
754
+ {
755
+ "epoch": 0.41197420348445474,
756
+ "grad_norm": 0.24997533857822418,
757
+ "learning_rate": 6.643194658312987e-06,
758
+ "loss": 0.732,
759
+ "step": 1070
760
+ },
761
+ {
762
+ "epoch": 0.415824429685244,
763
+ "grad_norm": 0.23343941569328308,
764
+ "learning_rate": 6.584174093857676e-06,
765
+ "loss": 0.7589,
766
+ "step": 1080
767
+ },
768
+ {
769
+ "epoch": 0.4196746558860333,
770
+ "grad_norm": 0.27916407585144043,
771
+ "learning_rate": 6.5249071294110735e-06,
772
+ "loss": 0.7398,
773
+ "step": 1090
774
+ },
775
+ {
776
+ "epoch": 0.4235248820868226,
777
+ "grad_norm": 0.2595101296901703,
778
+ "learning_rate": 6.465402983265274e-06,
779
+ "loss": 0.7535,
780
+ "step": 1100
781
+ },
782
+ {
783
+ "epoch": 0.4273751082876119,
784
+ "grad_norm": 0.24531684815883636,
785
+ "learning_rate": 6.405670910603243e-06,
786
+ "loss": 0.7489,
787
+ "step": 1110
788
+ },
789
+ {
790
+ "epoch": 0.43122533448840117,
791
+ "grad_norm": 0.2551664710044861,
792
+ "learning_rate": 6.345720202059291e-06,
793
+ "loss": 0.745,
794
+ "step": 1120
795
+ },
796
+ {
797
+ "epoch": 0.4350755606891905,
798
+ "grad_norm": 0.24974071979522705,
799
+ "learning_rate": 6.285560182274008e-06,
800
+ "loss": 0.7226,
801
+ "step": 1130
802
+ },
803
+ {
804
+ "epoch": 0.4389257868899798,
805
+ "grad_norm": 0.25124430656433105,
806
+ "learning_rate": 6.225200208443936e-06,
807
+ "loss": 0.7483,
808
+ "step": 1140
809
+ },
810
+ {
811
+ "epoch": 0.4427760130907691,
812
+ "grad_norm": 0.23269502818584442,
813
+ "learning_rate": 6.16464966886616e-06,
814
+ "loss": 0.7327,
815
+ "step": 1150
816
+ },
817
+ {
818
+ "epoch": 0.44662623929155837,
819
+ "grad_norm": 0.2840500771999359,
820
+ "learning_rate": 6.103917981478065e-06,
821
+ "loss": 0.7593,
822
+ "step": 1160
823
+ },
824
+ {
825
+ "epoch": 0.45047646549234766,
826
+ "grad_norm": 0.23583635687828064,
827
+ "learning_rate": 6.043014592392489e-06,
828
+ "loss": 0.7692,
829
+ "step": 1170
830
+ },
831
+ {
832
+ "epoch": 0.454326691693137,
833
+ "grad_norm": 0.23781178891658783,
834
+ "learning_rate": 5.981948974428485e-06,
835
+ "loss": 0.7456,
836
+ "step": 1180
837
+ },
838
+ {
839
+ "epoch": 0.4581769178939263,
840
+ "grad_norm": 0.26153862476348877,
841
+ "learning_rate": 5.920730625637934e-06,
842
+ "loss": 0.721,
843
+ "step": 1190
844
+ },
845
+ {
846
+ "epoch": 0.46202714409471557,
847
+ "grad_norm": 0.25586453080177307,
848
+ "learning_rate": 5.8593690678282355e-06,
849
+ "loss": 0.7362,
850
+ "step": 1200
851
+ },
852
+ {
853
+ "epoch": 0.46587737029550486,
854
+ "grad_norm": 0.24769704043865204,
855
+ "learning_rate": 5.797873845081301e-06,
856
+ "loss": 0.7395,
857
+ "step": 1210
858
+ },
859
+ {
860
+ "epoch": 0.46972759649629414,
861
+ "grad_norm": 0.24773752689361572,
862
+ "learning_rate": 5.736254522269075e-06,
863
+ "loss": 0.7552,
864
+ "step": 1220
865
+ },
866
+ {
867
+ "epoch": 0.47357782269708343,
868
+ "grad_norm": 0.243848517537117,
869
+ "learning_rate": 5.6745206835658386e-06,
870
+ "loss": 0.7581,
871
+ "step": 1230
872
+ },
873
+ {
874
+ "epoch": 0.47742804889787277,
875
+ "grad_norm": 0.2369810938835144,
876
+ "learning_rate": 5.612681930957495e-06,
877
+ "loss": 0.7289,
878
+ "step": 1240
879
+ },
880
+ {
881
+ "epoch": 0.48127827509866206,
882
+ "grad_norm": 0.23909157514572144,
883
+ "learning_rate": 5.550747882748095e-06,
884
+ "loss": 0.7243,
885
+ "step": 1250
886
+ },
887
+ {
888
+ "epoch": 0.48512850129945134,
889
+ "grad_norm": 0.2631821036338806,
890
+ "learning_rate": 5.4887281720638195e-06,
891
+ "loss": 0.7353,
892
+ "step": 1260
893
+ },
894
+ {
895
+ "epoch": 0.48897872750024063,
896
+ "grad_norm": 0.26492393016815186,
897
+ "learning_rate": 5.42663244535466e-06,
898
+ "loss": 0.717,
899
+ "step": 1270
900
+ },
901
+ {
902
+ "epoch": 0.4928289537010299,
903
+ "grad_norm": 0.264731228351593,
904
+ "learning_rate": 5.3644703608940205e-06,
905
+ "loss": 0.7279,
906
+ "step": 1280
907
+ },
908
+ {
909
+ "epoch": 0.49667917990181926,
910
+ "grad_norm": 0.2398008555173874,
911
+ "learning_rate": 5.302251587276491e-06,
912
+ "loss": 0.7369,
913
+ "step": 1290
914
+ },
915
+ {
916
+ "epoch": 0.5005294061026085,
917
+ "grad_norm": 0.2304096817970276,
918
+ "learning_rate": 5.2399858019140005e-06,
919
+ "loss": 0.7373,
920
+ "step": 1300
921
+ },
922
+ {
923
+ "epoch": 0.5043796323033978,
924
+ "grad_norm": 0.23684367537498474,
925
+ "learning_rate": 5.177682689530615e-06,
926
+ "loss": 0.7566,
927
+ "step": 1310
928
+ },
929
+ {
930
+ "epoch": 0.5082298585041871,
931
+ "grad_norm": 0.25382277369499207,
932
+ "learning_rate": 5.11535194065619e-06,
933
+ "loss": 0.7182,
934
+ "step": 1320
935
+ },
936
+ {
937
+ "epoch": 0.5120800847049765,
938
+ "grad_norm": 0.26454922556877136,
939
+ "learning_rate": 5.053003250119113e-06,
940
+ "loss": 0.7279,
941
+ "step": 1330
942
+ },
943
+ {
944
+ "epoch": 0.5159303109057657,
945
+ "grad_norm": 0.25216519832611084,
946
+ "learning_rate": 4.9906463155383925e-06,
947
+ "loss": 0.7383,
948
+ "step": 1340
949
+ },
950
+ {
951
+ "epoch": 0.519780537106555,
952
+ "grad_norm": 0.24496489763259888,
953
+ "learning_rate": 4.9282908358153046e-06,
954
+ "loss": 0.7339,
955
+ "step": 1350
956
+ },
957
+ {
958
+ "epoch": 0.5236307633073443,
959
+ "grad_norm": 0.26152753829956055,
960
+ "learning_rate": 4.8659465096248345e-06,
961
+ "loss": 0.7454,
962
+ "step": 1360
963
+ },
964
+ {
965
+ "epoch": 0.5274809895081336,
966
+ "grad_norm": 0.25739532709121704,
967
+ "learning_rate": 4.803623033907169e-06,
968
+ "loss": 0.7419,
969
+ "step": 1370
970
+ },
971
+ {
972
+ "epoch": 0.5313312157089229,
973
+ "grad_norm": 0.23496253788471222,
974
+ "learning_rate": 4.741330102359441e-06,
975
+ "loss": 0.7522,
976
+ "step": 1380
977
+ },
978
+ {
979
+ "epoch": 0.5351814419097122,
980
+ "grad_norm": 0.25455138087272644,
981
+ "learning_rate": 4.679077403927996e-06,
982
+ "loss": 0.7305,
983
+ "step": 1390
984
+ },
985
+ {
986
+ "epoch": 0.5390316681105015,
987
+ "grad_norm": 0.2368665337562561,
988
+ "learning_rate": 4.616874621301373e-06,
989
+ "loss": 0.732,
990
+ "step": 1400
991
+ },
992
+ {
993
+ "epoch": 0.5428818943112907,
994
+ "grad_norm": 0.2543937563896179,
995
+ "learning_rate": 4.554731429404293e-06,
996
+ "loss": 0.7471,
997
+ "step": 1410
998
+ },
999
+ {
1000
+ "epoch": 0.5467321205120801,
1001
+ "grad_norm": 0.2678263187408447,
1002
+ "learning_rate": 4.492657493892822e-06,
1003
+ "loss": 0.7271,
1004
+ "step": 1420
1005
+ },
1006
+ {
1007
+ "epoch": 0.5505823467128694,
1008
+ "grad_norm": 0.26733070611953735,
1009
+ "learning_rate": 4.430662469650999e-06,
1010
+ "loss": 0.7438,
1011
+ "step": 1430
1012
+ },
1013
+ {
1014
+ "epoch": 0.5544325729136587,
1015
+ "grad_norm": 0.27192339301109314,
1016
+ "learning_rate": 4.368755999289118e-06,
1017
+ "loss": 0.7513,
1018
+ "step": 1440
1019
+ },
1020
+ {
1021
+ "epoch": 0.558282799114448,
1022
+ "grad_norm": 0.263380229473114,
1023
+ "learning_rate": 4.306947711643949e-06,
1024
+ "loss": 0.7324,
1025
+ "step": 1450
1026
+ },
1027
+ {
1028
+ "epoch": 0.5621330253152372,
1029
+ "grad_norm": 0.2527351379394531,
1030
+ "learning_rate": 4.245247220281068e-06,
1031
+ "loss": 0.7359,
1032
+ "step": 1460
1033
+ },
1034
+ {
1035
+ "epoch": 0.5659832515160266,
1036
+ "grad_norm": 0.2406824380159378,
1037
+ "learning_rate": 4.183664121999587e-06,
1038
+ "loss": 0.7303,
1039
+ "step": 1470
1040
+ },
1041
+ {
1042
+ "epoch": 0.5698334777168159,
1043
+ "grad_norm": 0.24566097557544708,
1044
+ "learning_rate": 4.122207995339474e-06,
1045
+ "loss": 0.731,
1046
+ "step": 1480
1047
+ },
1048
+ {
1049
+ "epoch": 0.5736837039176051,
1050
+ "grad_norm": 0.2358175665140152,
1051
+ "learning_rate": 4.060888399091734e-06,
1052
+ "loss": 0.7335,
1053
+ "step": 1490
1054
+ },
1055
+ {
1056
+ "epoch": 0.5775339301183945,
1057
+ "grad_norm": 0.24320851266384125,
1058
+ "learning_rate": 3.999714870811639e-06,
1059
+ "loss": 0.7464,
1060
+ "step": 1500
1061
+ },
1062
+ {
1063
+ "epoch": 0.5813841563191837,
1064
+ "grad_norm": 0.24450035393238068,
1065
+ "learning_rate": 3.938696925335275e-06,
1066
+ "loss": 0.7384,
1067
+ "step": 1510
1068
+ },
1069
+ {
1070
+ "epoch": 0.5852343825199731,
1071
+ "grad_norm": 0.2543749511241913,
1072
+ "learning_rate": 3.87784405329962e-06,
1073
+ "loss": 0.7448,
1074
+ "step": 1520
1075
+ },
1076
+ {
1077
+ "epoch": 0.5890846087207623,
1078
+ "grad_norm": 0.24193929135799408,
1079
+ "learning_rate": 3.817165719666382e-06,
1080
+ "loss": 0.745,
1081
+ "step": 1530
1082
+ },
1083
+ {
1084
+ "epoch": 0.5929348349215516,
1085
+ "grad_norm": 0.2424173206090927,
1086
+ "learning_rate": 3.756671362249841e-06,
1087
+ "loss": 0.7539,
1088
+ "step": 1540
1089
+ },
1090
+ {
1091
+ "epoch": 0.596785061122341,
1092
+ "grad_norm": 0.2362832874059677,
1093
+ "learning_rate": 3.6963703902488907e-06,
1094
+ "loss": 0.7371,
1095
+ "step": 1550
1096
+ },
1097
+ {
1098
+ "epoch": 0.6006352873231302,
1099
+ "grad_norm": 0.23754023015499115,
1100
+ "learning_rate": 3.6362721827835655e-06,
1101
+ "loss": 0.7488,
1102
+ "step": 1560
1103
+ },
1104
+ {
1105
+ "epoch": 0.6044855135239195,
1106
+ "grad_norm": 0.2548673152923584,
1107
+ "learning_rate": 3.576386087436212e-06,
1108
+ "loss": 0.7488,
1109
+ "step": 1570
1110
+ },
1111
+ {
1112
+ "epoch": 0.6083357397247088,
1113
+ "grad_norm": 0.2547430694103241,
1114
+ "learning_rate": 3.5167214187975913e-06,
1115
+ "loss": 0.7382,
1116
+ "step": 1580
1117
+ },
1118
+ {
1119
+ "epoch": 0.6121859659254981,
1120
+ "grad_norm": 0.22359825670719147,
1121
+ "learning_rate": 3.457287457018086e-06,
1122
+ "loss": 0.7409,
1123
+ "step": 1590
1124
+ },
1125
+ {
1126
+ "epoch": 0.6160361921262875,
1127
+ "grad_norm": 0.23000743985176086,
1128
+ "learning_rate": 3.3980934463643003e-06,
1129
+ "loss": 0.7315,
1130
+ "step": 1600
1131
+ },
1132
+ {
1133
+ "epoch": 0.6198864183270767,
1134
+ "grad_norm": 0.2249031662940979,
1135
+ "learning_rate": 3.3391485937812053e-06,
1136
+ "loss": 0.7119,
1137
+ "step": 1610
1138
+ },
1139
+ {
1140
+ "epoch": 0.623736644527866,
1141
+ "grad_norm": 0.2596883773803711,
1142
+ "learning_rate": 3.280462067460115e-06,
1143
+ "loss": 0.7114,
1144
+ "step": 1620
1145
+ },
1146
+ {
1147
+ "epoch": 0.6275868707286553,
1148
+ "grad_norm": 0.23822768032550812,
1149
+ "learning_rate": 3.222042995412669e-06,
1150
+ "loss": 0.7419,
1151
+ "step": 1630
1152
+ },
1153
+ {
1154
+ "epoch": 0.6314370969294446,
1155
+ "grad_norm": 0.22608095407485962,
1156
+ "learning_rate": 3.163900464051089e-06,
1157
+ "loss": 0.7295,
1158
+ "step": 1640
1159
+ },
1160
+ {
1161
+ "epoch": 0.635287323130234,
1162
+ "grad_norm": 0.25906655192375183,
1163
+ "learning_rate": 3.1060435167748826e-06,
1164
+ "loss": 0.742,
1165
+ "step": 1650
1166
+ },
1167
+ {
1168
+ "epoch": 0.6391375493310232,
1169
+ "grad_norm": 0.23130592703819275,
1170
+ "learning_rate": 3.048481152564253e-06,
1171
+ "loss": 0.7229,
1172
+ "step": 1660
1173
+ },
1174
+ {
1175
+ "epoch": 0.6429877755318125,
1176
+ "grad_norm": 0.2520008981227875,
1177
+ "learning_rate": 2.9912223245804017e-06,
1178
+ "loss": 0.7481,
1179
+ "step": 1670
1180
+ },
1181
+ {
1182
+ "epoch": 0.6468380017326018,
1183
+ "grad_norm": 0.25257977843284607,
1184
+ "learning_rate": 2.934275938772981e-06,
1185
+ "loss": 0.7372,
1186
+ "step": 1680
1187
+ },
1188
+ {
1189
+ "epoch": 0.6506882279333911,
1190
+ "grad_norm": 0.2665594518184662,
1191
+ "learning_rate": 2.8776508524948676e-06,
1192
+ "loss": 0.7289,
1193
+ "step": 1690
1194
+ },
1195
+ {
1196
+ "epoch": 0.6545384541341804,
1197
+ "grad_norm": 0.22724343836307526,
1198
+ "learning_rate": 2.821355873124503e-06,
1199
+ "loss": 0.7255,
1200
+ "step": 1700
1201
+ },
1202
+ {
1203
+ "epoch": 0.6583886803349697,
1204
+ "grad_norm": 0.2574631869792938,
1205
+ "learning_rate": 2.7653997566960144e-06,
1206
+ "loss": 0.7439,
1207
+ "step": 1710
1208
+ },
1209
+ {
1210
+ "epoch": 0.662238906535759,
1211
+ "grad_norm": 0.25911903381347656,
1212
+ "learning_rate": 2.709791206537309e-06,
1213
+ "loss": 0.7331,
1214
+ "step": 1720
1215
+ },
1216
+ {
1217
+ "epoch": 0.6660891327365482,
1218
+ "grad_norm": 0.24707308411598206,
1219
+ "learning_rate": 2.6545388719163747e-06,
1220
+ "loss": 0.7321,
1221
+ "step": 1730
1222
+ },
1223
+ {
1224
+ "epoch": 0.6699393589373376,
1225
+ "grad_norm": 0.24382266402244568,
1226
+ "learning_rate": 2.599651346695979e-06,
1227
+ "loss": 0.734,
1228
+ "step": 1740
1229
+ },
1230
+ {
1231
+ "epoch": 0.6737895851381268,
1232
+ "grad_norm": 0.24484798312187195,
1233
+ "learning_rate": 2.545137167997007e-06,
1234
+ "loss": 0.7335,
1235
+ "step": 1750
1236
+ },
1237
+ {
1238
+ "epoch": 0.6776398113389162,
1239
+ "grad_norm": 0.24545250833034515,
1240
+ "learning_rate": 2.4910048148705972e-06,
1241
+ "loss": 0.7166,
1242
+ "step": 1760
1243
+ },
1244
+ {
1245
+ "epoch": 0.6814900375397055,
1246
+ "grad_norm": 0.257656991481781,
1247
+ "learning_rate": 2.4372627069793327e-06,
1248
+ "loss": 0.764,
1249
+ "step": 1770
1250
+ },
1251
+ {
1252
+ "epoch": 0.6853402637404947,
1253
+ "grad_norm": 0.2497720569372177,
1254
+ "learning_rate": 2.3839192032876486e-06,
1255
+ "loss": 0.7489,
1256
+ "step": 1780
1257
+ },
1258
+ {
1259
+ "epoch": 0.6891904899412841,
1260
+ "grad_norm": 0.22988678514957428,
1261
+ "learning_rate": 2.3309826007617047e-06,
1262
+ "loss": 0.7602,
1263
+ "step": 1790
1264
+ },
1265
+ {
1266
+ "epoch": 0.6930407161420733,
1267
+ "grad_norm": 0.24952177703380585,
1268
+ "learning_rate": 2.278461133078881e-06,
1269
+ "loss": 0.7237,
1270
+ "step": 1800
1271
+ },
1272
+ {
1273
+ "epoch": 0.6968909423428626,
1274
+ "grad_norm": 0.2502732574939728,
1275
+ "learning_rate": 2.2263629693471258e-06,
1276
+ "loss": 0.7133,
1277
+ "step": 1810
1278
+ },
1279
+ {
1280
+ "epoch": 0.700741168543652,
1281
+ "grad_norm": 0.23524925112724304,
1282
+ "learning_rate": 2.1746962128343414e-06,
1283
+ "loss": 0.7376,
1284
+ "step": 1820
1285
+ },
1286
+ {
1287
+ "epoch": 0.7045913947444412,
1288
+ "grad_norm": 0.238727867603302,
1289
+ "learning_rate": 2.1234688997080183e-06,
1290
+ "loss": 0.7286,
1291
+ "step": 1830
1292
+ },
1293
+ {
1294
+ "epoch": 0.7084416209452306,
1295
+ "grad_norm": 0.2582331895828247,
1296
+ "learning_rate": 2.0726889977853053e-06,
1297
+ "loss": 0.7326,
1298
+ "step": 1840
1299
+ },
1300
+ {
1301
+ "epoch": 0.7122918471460198,
1302
+ "grad_norm": 0.24908247590065002,
1303
+ "learning_rate": 2.022364405293703e-06,
1304
+ "loss": 0.7476,
1305
+ "step": 1850
1306
+ },
1307
+ {
1308
+ "epoch": 0.7161420733468091,
1309
+ "grad_norm": 0.23934200406074524,
1310
+ "learning_rate": 1.9725029496425814e-06,
1311
+ "loss": 0.7383,
1312
+ "step": 1860
1313
+ },
1314
+ {
1315
+ "epoch": 0.7199922995475985,
1316
+ "grad_norm": 0.24828778207302094,
1317
+ "learning_rate": 1.923112386205727e-06,
1318
+ "loss": 0.7356,
1319
+ "step": 1870
1320
+ },
1321
+ {
1322
+ "epoch": 0.7238425257483877,
1323
+ "grad_norm": 0.24286973476409912,
1324
+ "learning_rate": 1.874200397115079e-06,
1325
+ "loss": 0.7431,
1326
+ "step": 1880
1327
+ },
1328
+ {
1329
+ "epoch": 0.727692751949177,
1330
+ "grad_norm": 0.24799053370952606,
1331
+ "learning_rate": 1.825774590065857e-06,
1332
+ "loss": 0.7266,
1333
+ "step": 1890
1334
+ },
1335
+ {
1336
+ "epoch": 0.7315429781499663,
1337
+ "grad_norm": 0.24640041589736938,
1338
+ "learning_rate": 1.7778424971332813e-06,
1339
+ "loss": 0.7332,
1340
+ "step": 1900
1341
+ },
1342
+ {
1343
+ "epoch": 0.7353932043507556,
1344
+ "grad_norm": 0.23751963675022125,
1345
+ "learning_rate": 1.7304115736010407e-06,
1346
+ "loss": 0.7309,
1347
+ "step": 1910
1348
+ },
1349
+ {
1350
+ "epoch": 0.739243430551545,
1351
+ "grad_norm": 0.24287831783294678,
1352
+ "learning_rate": 1.6834891968017108e-06,
1353
+ "loss": 0.7256,
1354
+ "step": 1920
1355
+ },
1356
+ {
1357
+ "epoch": 0.7430936567523342,
1358
+ "grad_norm": 0.25131744146347046,
1359
+ "learning_rate": 1.6370826649692884e-06,
1360
+ "loss": 0.739,
1361
+ "step": 1930
1362
+ },
1363
+ {
1364
+ "epoch": 0.7469438829531235,
1365
+ "grad_norm": 0.24572177231311798,
1366
+ "learning_rate": 1.5911991961040463e-06,
1367
+ "loss": 0.7522,
1368
+ "step": 1940
1369
+ },
1370
+ {
1371
+ "epoch": 0.7507941091539128,
1372
+ "grad_norm": 0.24601101875305176,
1373
+ "learning_rate": 1.545845926849851e-06,
1374
+ "loss": 0.7404,
1375
+ "step": 1950
1376
+ },
1377
+ {
1378
+ "epoch": 0.7546443353547021,
1379
+ "grad_norm": 0.26765137910842896,
1380
+ "learning_rate": 1.5010299113841397e-06,
1381
+ "loss": 0.7558,
1382
+ "step": 1960
1383
+ },
1384
+ {
1385
+ "epoch": 0.7584945615554913,
1386
+ "grad_norm": 0.2414122372865677,
1387
+ "learning_rate": 1.4567581203207238e-06,
1388
+ "loss": 0.7456,
1389
+ "step": 1970
1390
+ },
1391
+ {
1392
+ "epoch": 0.7623447877562807,
1393
+ "grad_norm": 0.2603057026863098,
1394
+ "learning_rate": 1.413037439625593e-06,
1395
+ "loss": 0.7416,
1396
+ "step": 1980
1397
+ },
1398
+ {
1399
+ "epoch": 0.76619501395707,
1400
+ "grad_norm": 0.2654842436313629,
1401
+ "learning_rate": 1.3698746695458847e-06,
1402
+ "loss": 0.7389,
1403
+ "step": 1990
1404
+ },
1405
+ {
1406
+ "epoch": 0.7700452401578592,
1407
+ "grad_norm": 0.24445505440235138,
1408
+ "learning_rate": 1.3272765235521745e-06,
1409
+ "loss": 0.73,
1410
+ "step": 2000
1411
+ },
1412
+ {
1413
+ "epoch": 0.7738954663586486,
1414
+ "grad_norm": 0.23773328959941864,
1415
+ "learning_rate": 1.2852496272942877e-06,
1416
+ "loss": 0.7399,
1417
+ "step": 2010
1418
+ },
1419
+ {
1420
+ "epoch": 0.7777456925594378,
1421
+ "grad_norm": 0.2343347817659378,
1422
+ "learning_rate": 1.2438005175707396e-06,
1423
+ "loss": 0.7229,
1424
+ "step": 2020
1425
+ },
1426
+ {
1427
+ "epoch": 0.7815959187602272,
1428
+ "grad_norm": 0.2397432178258896,
1429
+ "learning_rate": 1.2029356413120297e-06,
1430
+ "loss": 0.7084,
1431
+ "step": 2030
1432
+ },
1433
+ {
1434
+ "epoch": 0.7854461449610165,
1435
+ "grad_norm": 0.24255838990211487,
1436
+ "learning_rate": 1.1626613545778786e-06,
1437
+ "loss": 0.7259,
1438
+ "step": 2040
1439
+ },
1440
+ {
1441
+ "epoch": 0.7892963711618057,
1442
+ "grad_norm": 0.2379896193742752,
1443
+ "learning_rate": 1.1229839215686294e-06,
1444
+ "loss": 0.7316,
1445
+ "step": 2050
1446
+ },
1447
+ {
1448
+ "epoch": 0.7931465973625951,
1449
+ "grad_norm": 0.22665193676948547,
1450
+ "learning_rate": 1.0839095136509188e-06,
1451
+ "loss": 0.7136,
1452
+ "step": 2060
1453
+ },
1454
+ {
1455
+ "epoch": 0.7969968235633843,
1456
+ "grad_norm": 0.2591674029827118,
1457
+ "learning_rate": 1.045444208397791e-06,
1458
+ "loss": 0.7349,
1459
+ "step": 2070
1460
+ },
1461
+ {
1462
+ "epoch": 0.8008470497641736,
1463
+ "grad_norm": 0.24025781452655792,
1464
+ "learning_rate": 1.007593988643402e-06,
1465
+ "loss": 0.7175,
1466
+ "step": 2080
1467
+ },
1468
+ {
1469
+ "epoch": 0.804697275964963,
1470
+ "grad_norm": 0.23568974435329437,
1471
+ "learning_rate": 9.703647415524613e-07,
1472
+ "loss": 0.7409,
1473
+ "step": 2090
1474
+ },
1475
+ {
1476
+ "epoch": 0.8085475021657522,
1477
+ "grad_norm": 0.23237770795822144,
1478
+ "learning_rate": 9.337622577045485e-07,
1479
+ "loss": 0.7276,
1480
+ "step": 2100
1481
+ },
1482
+ {
1483
+ "epoch": 0.8123977283665416,
1484
+ "grad_norm": 0.24259373545646667,
1485
+ "learning_rate": 8.977922301934588e-07,
1486
+ "loss": 0.7416,
1487
+ "step": 2110
1488
+ },
1489
+ {
1490
+ "epoch": 0.8162479545673308,
1491
+ "grad_norm": 0.24287734925746918,
1492
+ "learning_rate": 8.624602537417009e-07,
1493
+ "loss": 0.7444,
1494
+ "step": 2120
1495
+ },
1496
+ {
1497
+ "epoch": 0.8200981807681201,
1498
+ "grad_norm": 0.2603750228881836,
1499
+ "learning_rate": 8.277718238303123e-07,
1500
+ "loss": 0.7294,
1501
+ "step": 2130
1502
+ },
1503
+ {
1504
+ "epoch": 0.8239484069689095,
1505
+ "grad_norm": 0.23923136293888092,
1506
+ "learning_rate": 7.937323358440935e-07,
1507
+ "loss": 0.7321,
1508
+ "step": 2140
1509
+ },
1510
+ {
1511
+ "epoch": 0.8277986331696987,
1512
+ "grad_norm": 0.2624675929546356,
1513
+ "learning_rate": 7.603470842324223e-07,
1514
+ "loss": 0.7242,
1515
+ "step": 2150
1516
+ },
1517
+ {
1518
+ "epoch": 0.831648859370488,
1519
+ "grad_norm": 0.22938279807567596,
1520
+ "learning_rate": 7.276212616857575e-07,
1521
+ "loss": 0.741,
1522
+ "step": 2160
1523
+ },
1524
+ {
1525
+ "epoch": 0.8354990855712773,
1526
+ "grad_norm": 0.2362944483757019,
1527
+ "learning_rate": 6.955599583279837e-07,
1528
+ "loss": 0.7392,
1529
+ "step": 2170
1530
+ },
1531
+ {
1532
+ "epoch": 0.8393493117720666,
1533
+ "grad_norm": 0.24612662196159363,
1534
+ "learning_rate": 6.641681609246981e-07,
1535
+ "loss": 0.7344,
1536
+ "step": 2180
1537
+ },
1538
+ {
1539
+ "epoch": 0.8431995379728559,
1540
+ "grad_norm": 0.22963689267635345,
1541
+ "learning_rate": 6.334507521075711e-07,
1542
+ "loss": 0.7362,
1543
+ "step": 2190
1544
+ },
1545
+ {
1546
+ "epoch": 0.8470497641736452,
1547
+ "grad_norm": 0.2511405944824219,
1548
+ "learning_rate": 6.034125096149179e-07,
1549
+ "loss": 0.7347,
1550
+ "step": 2200
1551
+ },
1552
+ {
1553
+ "epoch": 0.8508999903744345,
1554
+ "grad_norm": 0.2456102818250656,
1555
+ "learning_rate": 5.740581055485722e-07,
1556
+ "loss": 0.7058,
1557
+ "step": 2210
1558
+ },
1559
+ {
1560
+ "epoch": 0.8547502165752238,
1561
+ "grad_norm": 0.23535743355751038,
1562
+ "learning_rate": 5.453921056471961e-07,
1563
+ "loss": 0.7391,
1564
+ "step": 2220
1565
+ },
1566
+ {
1567
+ "epoch": 0.8586004427760131,
1568
+ "grad_norm": 0.24837130308151245,
1569
+ "learning_rate": 5.174189685761288e-07,
1570
+ "loss": 0.7334,
1571
+ "step": 2230
1572
+ },
1573
+ {
1574
+ "epoch": 0.8624506689768023,
1575
+ "grad_norm": 0.23152261972427368,
1576
+ "learning_rate": 4.901430452338979e-07,
1577
+ "loss": 0.7474,
1578
+ "step": 2240
1579
+ },
1580
+ {
1581
+ "epoch": 0.8663008951775917,
1582
+ "grad_norm": 0.2381637841463089,
1583
+ "learning_rate": 4.6356857807548286e-07,
1584
+ "loss": 0.7267,
1585
+ "step": 2250
1586
+ },
1587
+ {
1588
+ "epoch": 0.870151121378381,
1589
+ "grad_norm": 0.2432614117860794,
1590
+ "learning_rate": 4.376997004524541e-07,
1591
+ "loss": 0.7161,
1592
+ "step": 2260
1593
+ },
1594
+ {
1595
+ "epoch": 0.8740013475791703,
1596
+ "grad_norm": 0.2498764544725418,
1597
+ "learning_rate": 4.12540435970073e-07,
1598
+ "loss": 0.724,
1599
+ "step": 2270
1600
+ },
1601
+ {
1602
+ "epoch": 0.8778515737799596,
1603
+ "grad_norm": 0.23934809863567352,
1604
+ "learning_rate": 3.880946978614703e-07,
1605
+ "loss": 0.7426,
1606
+ "step": 2280
1607
+ },
1608
+ {
1609
+ "epoch": 0.8817017999807488,
1610
+ "grad_norm": 0.23980608582496643,
1611
+ "learning_rate": 3.643662883789878e-07,
1612
+ "loss": 0.7381,
1613
+ "step": 2290
1614
+ },
1615
+ {
1616
+ "epoch": 0.8855520261815382,
1617
+ "grad_norm": 0.22927387058734894,
1618
+ "learning_rate": 3.41358898202781e-07,
1619
+ "loss": 0.7182,
1620
+ "step": 2300
1621
+ },
1622
+ {
1623
+ "epoch": 0.8894022523823275,
1624
+ "grad_norm": 0.22735606133937836,
1625
+ "learning_rate": 3.190761058667741e-07,
1626
+ "loss": 0.7283,
1627
+ "step": 2310
1628
+ },
1629
+ {
1630
+ "epoch": 0.8932524785831167,
1631
+ "grad_norm": 0.24090701341629028,
1632
+ "learning_rate": 2.9752137720206463e-07,
1633
+ "loss": 0.7341,
1634
+ "step": 2320
1635
+ },
1636
+ {
1637
+ "epoch": 0.8971027047839061,
1638
+ "grad_norm": 0.24678541719913483,
1639
+ "learning_rate": 2.7669806479785067e-07,
1640
+ "loss": 0.7405,
1641
+ "step": 2330
1642
+ },
1643
+ {
1644
+ "epoch": 0.9009529309846953,
1645
+ "grad_norm": 0.2500489354133606,
1646
+ "learning_rate": 2.566094074799763e-07,
1647
+ "loss": 0.7327,
1648
+ "step": 2340
1649
+ },
1650
+ {
1651
+ "epoch": 0.9048031571854847,
1652
+ "grad_norm": 0.2355518788099289,
1653
+ "learning_rate": 2.3725852980716557e-07,
1654
+ "loss": 0.7183,
1655
+ "step": 2350
1656
+ },
1657
+ {
1658
+ "epoch": 0.908653383386274,
1659
+ "grad_norm": 0.23016060888767242,
1660
+ "learning_rate": 2.1864844158503983e-07,
1661
+ "loss": 0.7282,
1662
+ "step": 2360
1663
+ },
1664
+ {
1665
+ "epoch": 0.9125036095870632,
1666
+ "grad_norm": 0.2680000066757202,
1667
+ "learning_rate": 2.007820373979702e-07,
1668
+ "loss": 0.72,
1669
+ "step": 2370
1670
+ },
1671
+ {
1672
+ "epoch": 0.9163538357878526,
1673
+ "grad_norm": 0.24530255794525146,
1674
+ "learning_rate": 1.8366209615886532e-07,
1675
+ "loss": 0.7308,
1676
+ "step": 2380
1677
+ },
1678
+ {
1679
+ "epoch": 0.9202040619886418,
1680
+ "grad_norm": 0.2361113280057907,
1681
+ "learning_rate": 1.6729128067693613e-07,
1682
+ "loss": 0.7454,
1683
+ "step": 2390
1684
+ },
1685
+ {
1686
+ "epoch": 0.9240542881894311,
1687
+ "grad_norm": 0.24870836734771729,
1688
+ "learning_rate": 1.5167213724353426e-07,
1689
+ "loss": 0.7445,
1690
+ "step": 2400
1691
+ },
1692
+ {
1693
+ "epoch": 0.9279045143902205,
1694
+ "grad_norm": 0.23738086223602295,
1695
+ "learning_rate": 1.3680709523610058e-07,
1696
+ "loss": 0.7255,
1697
+ "step": 2410
1698
+ },
1699
+ {
1700
+ "epoch": 0.9317547405910097,
1701
+ "grad_norm": 0.23620250821113586,
1702
+ "learning_rate": 1.2269846674030418e-07,
1703
+ "loss": 0.7456,
1704
+ "step": 2420
1705
+ },
1706
+ {
1707
+ "epoch": 0.935604966791799,
1708
+ "grad_norm": 0.24729779362678528,
1709
+ "learning_rate": 1.0934844619042595e-07,
1710
+ "loss": 0.7466,
1711
+ "step": 2430
1712
+ },
1713
+ {
1714
+ "epoch": 0.9394551929925883,
1715
+ "grad_norm": 0.24688898026943207,
1716
+ "learning_rate": 9.675911002803961e-08,
1717
+ "loss": 0.7301,
1718
+ "step": 2440
1719
+ },
1720
+ {
1721
+ "epoch": 0.9433054191933776,
1722
+ "grad_norm": 0.250814288854599,
1723
+ "learning_rate": 8.493241637904381e-08,
1724
+ "loss": 0.7273,
1725
+ "step": 2450
1726
+ },
1727
+ {
1728
+ "epoch": 0.9471556453941669,
1729
+ "grad_norm": 0.24807293713092804,
1730
+ "learning_rate": 7.387020474909689e-08,
1731
+ "loss": 0.7507,
1732
+ "step": 2460
1733
+ },
1734
+ {
1735
+ "epoch": 0.9510058715949562,
1736
+ "grad_norm": 0.26027828454971313,
1737
+ "learning_rate": 6.357419573750856e-08,
1738
+ "loss": 0.7444,
1739
+ "step": 2470
1740
+ },
1741
+ {
1742
+ "epoch": 0.9548560977957455,
1743
+ "grad_norm": 0.2493964582681656,
1744
+ "learning_rate": 5.4045990769613365e-08,
1745
+ "loss": 0.7271,
1746
+ "step": 2480
1747
+ },
1748
+ {
1749
+ "epoch": 0.9587063239965348,
1750
+ "grad_norm": 0.2913523316383362,
1751
+ "learning_rate": 4.5287071847694434e-08,
1752
+ "loss": 0.7306,
1753
+ "step": 2490
1754
+ },
1755
+ {
1756
+ "epoch": 0.9625565501973241,
1757
+ "grad_norm": 0.22955387830734253,
1758
+ "learning_rate": 3.7298801320467817e-08,
1759
+ "loss": 0.7279,
1760
+ "step": 2500
1761
+ },
1762
+ {
1763
+ "epoch": 0.9664067763981133,
1764
+ "grad_norm": 0.24414069950580597,
1765
+ "learning_rate": 3.0082421671192576e-08,
1766
+ "loss": 0.7214,
1767
+ "step": 2510
1768
+ },
1769
+ {
1770
+ "epoch": 0.9702570025989027,
1771
+ "grad_norm": 0.2446526736021042,
1772
+ "learning_rate": 2.3639055324410355e-08,
1773
+ "loss": 0.746,
1774
+ "step": 2520
1775
+ },
1776
+ {
1777
+ "epoch": 0.974107228799692,
1778
+ "grad_norm": 0.23625919222831726,
1779
+ "learning_rate": 1.7969704471369477e-08,
1780
+ "loss": 0.7252,
1781
+ "step": 2530
1782
+ },
1783
+ {
1784
+ "epoch": 0.9779574550004813,
1785
+ "grad_norm": 0.23089419305324554,
1786
+ "learning_rate": 1.3075250914142413e-08,
1787
+ "loss": 0.7322,
1788
+ "step": 2540
1789
+ },
1790
+ {
1791
+ "epoch": 0.9818076812012706,
1792
+ "grad_norm": 0.23270326852798462,
1793
+ "learning_rate": 8.956455928476049e-09,
1794
+ "loss": 0.7374,
1795
+ "step": 2550
1796
+ },
1797
+ {
1798
+ "epoch": 0.9856579074020598,
1799
+ "grad_norm": 0.24237456917762756,
1800
+ "learning_rate": 5.6139601453786275e-09,
1801
+ "loss": 0.7287,
1802
+ "step": 2560
1803
+ },
1804
+ {
1805
+ "epoch": 0.9895081336028492,
1806
+ "grad_norm": 0.23908966779708862,
1807
+ "learning_rate": 3.0482834514805602e-09,
1808
+ "loss": 0.7188,
1809
+ "step": 2570
1810
+ },
1811
+ {
1812
+ "epoch": 0.9933583598036385,
1813
+ "grad_norm": 0.22261138260364532,
1814
+ "learning_rate": 1.2598249081713366e-09,
1815
+ "loss": 0.7105,
1816
+ "step": 2580
1817
+ },
1818
+ {
1819
+ "epoch": 0.9972085860044277,
1820
+ "grad_norm": 0.24466122686862946,
1821
+ "learning_rate": 2.488626895302826e-10,
1822
+ "loss": 0.7467,
1823
+ "step": 2590
1824
+ },
1825
+ {
1826
+ "epoch": 0.9999037443449803,
1827
+ "step": 2597,
1828
+ "total_flos": 4.633332116728316e+18,
1829
+ "train_loss": 0.765134797446949,
1830
+ "train_runtime": 10841.995,
1831
+ "train_samples_per_second": 61.327,
1832
+ "train_steps_per_second": 0.24
1833
+ }
1834
+ ],
1835
+ "logging_steps": 10,
1836
+ "max_steps": 2597,
1837
+ "num_input_tokens_seen": 0,
1838
+ "num_train_epochs": 1,
1839
+ "save_steps": 500,
1840
+ "stateful_callbacks": {
1841
+ "TrainerControl": {
1842
+ "args": {
1843
+ "should_epoch_stop": false,
1844
+ "should_evaluate": false,
1845
+ "should_log": false,
1846
+ "should_save": true,
1847
+ "should_training_stop": true
1848
+ },
1849
+ "attributes": {}
1850
+ }
1851
+ },
1852
+ "total_flos": 4.633332116728316e+18,
1853
+ "train_batch_size": 8,
1854
+ "trial_name": null,
1855
+ "trial_params": null
1856
+ }
continuous/qwen3/siglip2_sm14_224/finetune/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:1b2c1bd90ea17fc0d80b130e7b9e04279a7b63961532b3ad8b12eb1346677258
3
+ size 6136
continuous/qwen3/siglip2_sm14_224/finetune/vocab.json ADDED
The diff for this file is too large to render. See raw diff
 
continuous/qwen3/siglip2_sm14_224/pretrain/config.json ADDED
@@ -0,0 +1,60 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_attn_implementation_autoset": true,
3
+ "add_faster_video": false,
4
+ "add_time_instruction": false,
5
+ "architectures": [
6
+ "Qwen3ForCausalLM"
7
+ ],
8
+ "attention_bias": false,
9
+ "attention_dropout": 0.0,
10
+ "bos_token_id": 151643,
11
+ "eos_token_id": 151645,
12
+ "faster_token_stride": 10,
13
+ "force_sample": false,
14
+ "head_dim": 128,
15
+ "hidden_act": "silu",
16
+ "hidden_size": 2048,
17
+ "image_aspect_ratio": "square",
18
+ "image_crop_resolution": null,
19
+ "image_grid_pinpoints": null,
20
+ "image_split_resolution": null,
21
+ "initializer_range": 0.02,
22
+ "intermediate_size": 6144,
23
+ "max_position_embeddings": 40960,
24
+ "max_window_layers": 28,
25
+ "mm_hidden_size": 1152,
26
+ "mm_newline_position": "grid",
27
+ "mm_patch_merge_type": "flat",
28
+ "mm_projector_lr": null,
29
+ "mm_projector_type": "mlp2x_gelu",
30
+ "mm_resampler_type": null,
31
+ "mm_spatial_pool_mode": "bilinear",
32
+ "mm_spatial_pool_stride": null,
33
+ "mm_tunable_parts": "mm_mlp_adapter",
34
+ "mm_use_im_patch_token": false,
35
+ "mm_use_im_start_end": false,
36
+ "mm_vision_select_feature": "patch",
37
+ "mm_vision_select_layer": -2,
38
+ "mm_vision_tower": "open_clip_hub:ViT-SO400M-14-SigLIP2",
39
+ "mm_vision_tower_lr": null,
40
+ "model_type": "qwen3",
41
+ "num_attention_heads": 16,
42
+ "num_hidden_layers": 28,
43
+ "num_key_value_heads": 8,
44
+ "pos_skipping_range": 4096,
45
+ "rms_norm_eps": 1e-06,
46
+ "rope_scaling": null,
47
+ "rope_theta": 1000000,
48
+ "sliding_window": null,
49
+ "tie_word_embeddings": true,
50
+ "tokenizer_model_max_length": 8192,
51
+ "tokenizer_padding_side": "right",
52
+ "torch_dtype": "bfloat16",
53
+ "transformers_version": "4.51.3",
54
+ "use_cache": true,
55
+ "use_mm_proj": true,
56
+ "use_pos_skipping": false,
57
+ "use_sliding_window": false,
58
+ "vision_tower_pretrained": "/cache/ckpt/download/tokenizer/continuous/siglip2/siglip2_sm14_224.npz",
59
+ "vocab_size": 151936
60
+ }
continuous/qwen3/siglip2_sm14_224/pretrain/mm_projector.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:446595bc0c9ef2d914b421a849825f4d4a5c256d259044cdec147dd29a2c6017
3
+ size 26232888
continuous/qwen3/siglip2_sm14_224/pretrain/trainer_state.json ADDED
@@ -0,0 +1,421 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_global_step": null,
3
+ "best_metric": null,
4
+ "best_model_checkpoint": null,
5
+ "epoch": 1.0,
6
+ "eval_steps": 500,
7
+ "global_step": 545,
8
+ "is_hyper_param_search": false,
9
+ "is_local_process_zero": true,
10
+ "is_world_process_zero": true,
11
+ "log_history": [
12
+ {
13
+ "epoch": 0.01834862385321101,
14
+ "grad_norm": 7.482438087463379,
15
+ "learning_rate": 0.00032142857142857147,
16
+ "loss": 8.8604,
17
+ "step": 10
18
+ },
19
+ {
20
+ "epoch": 0.03669724770642202,
21
+ "grad_norm": 1.2543365955352783,
22
+ "learning_rate": 0.0006785714285714287,
23
+ "loss": 6.0576,
24
+ "step": 20
25
+ },
26
+ {
27
+ "epoch": 0.05504587155963303,
28
+ "grad_norm": 0.21129053831100464,
29
+ "learning_rate": 0.0009999907688176173,
30
+ "loss": 4.9013,
31
+ "step": 30
32
+ },
33
+ {
34
+ "epoch": 0.07339449541284404,
35
+ "grad_norm": 0.18647241592407227,
36
+ "learning_rate": 0.0009988834393115766,
37
+ "loss": 4.7082,
38
+ "step": 40
39
+ },
40
+ {
41
+ "epoch": 0.09174311926605505,
42
+ "grad_norm": 0.1863221824169159,
43
+ "learning_rate": 0.0009959345572940772,
44
+ "loss": 4.6388,
45
+ "step": 50
46
+ },
47
+ {
48
+ "epoch": 0.11009174311926606,
49
+ "grad_norm": 0.4343675374984741,
50
+ "learning_rate": 0.0009911550081155983,
51
+ "loss": 4.6068,
52
+ "step": 60
53
+ },
54
+ {
55
+ "epoch": 0.12844036697247707,
56
+ "grad_norm": 0.1967705339193344,
57
+ "learning_rate": 0.0009845624347566433,
58
+ "loss": 4.553,
59
+ "step": 70
60
+ },
61
+ {
62
+ "epoch": 0.14678899082568808,
63
+ "grad_norm": 0.20485320687294006,
64
+ "learning_rate": 0.0009761811727013547,
65
+ "loss": 4.5092,
66
+ "step": 80
67
+ },
68
+ {
69
+ "epoch": 0.1651376146788991,
70
+ "grad_norm": 0.13904982805252075,
71
+ "learning_rate": 0.0009660421601067666,
72
+ "loss": 4.4561,
73
+ "step": 90
74
+ },
75
+ {
76
+ "epoch": 0.1834862385321101,
77
+ "grad_norm": 0.1629440188407898,
78
+ "learning_rate": 0.0009541828235992926,
79
+ "loss": 4.4296,
80
+ "step": 100
81
+ },
82
+ {
83
+ "epoch": 0.2018348623853211,
84
+ "grad_norm": 0.22176125645637512,
85
+ "learning_rate": 0.0009406469401200151,
86
+ "loss": 4.3909,
87
+ "step": 110
88
+ },
89
+ {
90
+ "epoch": 0.22018348623853212,
91
+ "grad_norm": 0.26596155762672424,
92
+ "learning_rate": 0.0009254844753287492,
93
+ "loss": 4.3754,
94
+ "step": 120
95
+ },
96
+ {
97
+ "epoch": 0.23853211009174313,
98
+ "grad_norm": 0.23705574870109558,
99
+ "learning_rate": 0.0009087513991633924,
100
+ "loss": 4.334,
101
+ "step": 130
102
+ },
103
+ {
104
+ "epoch": 0.25688073394495414,
105
+ "grad_norm": 0.19574818015098572,
106
+ "learning_rate": 0.0008905094792353917,
107
+ "loss": 4.3044,
108
+ "step": 140
109
+ },
110
+ {
111
+ "epoch": 0.27522935779816515,
112
+ "grad_norm": 0.13468025624752045,
113
+ "learning_rate": 0.0008708260528239789,
114
+ "loss": 4.2872,
115
+ "step": 150
116
+ },
117
+ {
118
+ "epoch": 0.29357798165137616,
119
+ "grad_norm": 0.1859128177165985,
120
+ "learning_rate": 0.0008497737783108237,
121
+ "loss": 4.252,
122
+ "step": 160
123
+ },
124
+ {
125
+ "epoch": 0.3119266055045872,
126
+ "grad_norm": 0.18957847356796265,
127
+ "learning_rate": 0.0008274303669726426,
128
+ "loss": 4.226,
129
+ "step": 170
130
+ },
131
+ {
132
+ "epoch": 0.3302752293577982,
133
+ "grad_norm": 0.14642402529716492,
134
+ "learning_rate": 0.0008038782961218137,
135
+ "loss": 4.2038,
136
+ "step": 180
137
+ },
138
+ {
139
+ "epoch": 0.3486238532110092,
140
+ "grad_norm": 0.14492210745811462,
141
+ "learning_rate": 0.000779204504653894,
142
+ "loss": 4.1812,
143
+ "step": 190
144
+ },
145
+ {
146
+ "epoch": 0.3669724770642202,
147
+ "grad_norm": 0.16114751994609833,
148
+ "learning_rate": 0.0007535000721258791,
149
+ "loss": 4.1464,
150
+ "step": 200
151
+ },
152
+ {
153
+ "epoch": 0.3853211009174312,
154
+ "grad_norm": 0.1807613968849182,
155
+ "learning_rate": 0.00072685988254984,
156
+ "loss": 4.129,
157
+ "step": 210
158
+ },
159
+ {
160
+ "epoch": 0.4036697247706422,
161
+ "grad_norm": 0.18491551280021667,
162
+ "learning_rate": 0.0006993822741429906,
163
+ "loss": 4.134,
164
+ "step": 220
165
+ },
166
+ {
167
+ "epoch": 0.42201834862385323,
168
+ "grad_norm": 0.1542729139328003,
169
+ "learning_rate": 0.0006711686763270818,
170
+ "loss": 4.1132,
171
+ "step": 230
172
+ },
173
+ {
174
+ "epoch": 0.44036697247706424,
175
+ "grad_norm": 0.1544124037027359,
176
+ "learning_rate": 0.0006423232353170831,
177
+ "loss": 4.0868,
178
+ "step": 240
179
+ },
180
+ {
181
+ "epoch": 0.45871559633027525,
182
+ "grad_norm": 0.16166554391384125,
183
+ "learning_rate": 0.0006129524296812335,
184
+ "loss": 4.0501,
185
+ "step": 250
186
+ },
187
+ {
188
+ "epoch": 0.47706422018348627,
189
+ "grad_norm": 0.23510634899139404,
190
+ "learning_rate": 0.0005831646772915651,
191
+ "loss": 4.037,
192
+ "step": 260
193
+ },
194
+ {
195
+ "epoch": 0.4954128440366973,
196
+ "grad_norm": 0.17621584236621857,
197
+ "learning_rate": 0.0005530699351157809,
198
+ "loss": 3.9915,
199
+ "step": 270
200
+ },
201
+ {
202
+ "epoch": 0.5137614678899083,
203
+ "grad_norm": 0.1644352525472641,
204
+ "learning_rate": 0.0005227792933277943,
205
+ "loss": 3.9607,
206
+ "step": 280
207
+ },
208
+ {
209
+ "epoch": 0.5321100917431193,
210
+ "grad_norm": 0.1596604287624359,
211
+ "learning_rate": 0.0004924045652352092,
212
+ "loss": 3.9322,
213
+ "step": 290
214
+ },
215
+ {
216
+ "epoch": 0.5504587155963303,
217
+ "grad_norm": 0.13874907791614532,
218
+ "learning_rate": 0.0004620578745374633,
219
+ "loss": 3.9168,
220
+ "step": 300
221
+ },
222
+ {
223
+ "epoch": 0.5688073394495413,
224
+ "grad_norm": 0.15355391800403595,
225
+ "learning_rate": 0.0004318512414382059,
226
+ "loss": 3.9182,
227
+ "step": 310
228
+ },
229
+ {
230
+ "epoch": 0.5871559633027523,
231
+ "grad_norm": 0.15171395242214203,
232
+ "learning_rate": 0.0004018961691397148,
233
+ "loss": 3.9037,
234
+ "step": 320
235
+ },
236
+ {
237
+ "epoch": 0.6055045871559633,
238
+ "grad_norm": 0.15213322639465332,
239
+ "learning_rate": 0.0003723032322457458,
240
+ "loss": 3.8802,
241
+ "step": 330
242
+ },
243
+ {
244
+ "epoch": 0.6238532110091743,
245
+ "grad_norm": 0.14570190012454987,
246
+ "learning_rate": 0.0003431816685921625,
247
+ "loss": 3.8744,
248
+ "step": 340
249
+ },
250
+ {
251
+ "epoch": 0.6422018348623854,
252
+ "grad_norm": 0.14543983340263367,
253
+ "learning_rate": 0.0003146389760120416,
254
+ "loss": 3.8685,
255
+ "step": 350
256
+ },
257
+ {
258
+ "epoch": 0.6605504587155964,
259
+ "grad_norm": 0.15356746315956116,
260
+ "learning_rate": 0.00028678051552373485,
261
+ "loss": 3.8739,
262
+ "step": 360
263
+ },
264
+ {
265
+ "epoch": 0.6788990825688074,
266
+ "grad_norm": 0.13074801862239838,
267
+ "learning_rate": 0.0002597091224066581,
268
+ "loss": 3.8584,
269
+ "step": 370
270
+ },
271
+ {
272
+ "epoch": 0.6972477064220184,
273
+ "grad_norm": 0.1379108875989914,
274
+ "learning_rate": 0.0002335247266004629,
275
+ "loss": 3.8671,
276
+ "step": 380
277
+ },
278
+ {
279
+ "epoch": 0.7155963302752294,
280
+ "grad_norm": 0.13431651890277863,
281
+ "learning_rate": 0.000208323983828828,
282
+ "loss": 3.8513,
283
+ "step": 390
284
+ },
285
+ {
286
+ "epoch": 0.7339449541284404,
287
+ "grad_norm": 0.1298692226409912,
288
+ "learning_rate": 0.0001841999188095224,
289
+ "loss": 3.8468,
290
+ "step": 400
291
+ },
292
+ {
293
+ "epoch": 0.7522935779816514,
294
+ "grad_norm": 0.123204305768013,
295
+ "learning_rate": 0.00016124158186777676,
296
+ "loss": 3.8516,
297
+ "step": 410
298
+ },
299
+ {
300
+ "epoch": 0.7706422018348624,
301
+ "grad_norm": 0.1227988451719284,
302
+ "learning_rate": 0.0001395337202205211,
303
+ "loss": 3.8421,
304
+ "step": 420
305
+ },
306
+ {
307
+ "epoch": 0.7889908256880734,
308
+ "grad_norm": 0.12320256233215332,
309
+ "learning_rate": 0.00011915646514489292,
310
+ "loss": 3.8437,
311
+ "step": 430
312
+ },
313
+ {
314
+ "epoch": 0.8073394495412844,
315
+ "grad_norm": 0.14015628397464752,
316
+ "learning_rate": 0.00010018503618578818,
317
+ "loss": 3.8479,
318
+ "step": 440
319
+ },
320
+ {
321
+ "epoch": 0.8256880733944955,
322
+ "grad_norm": 0.12411060184240341,
323
+ "learning_rate": 8.268946349432583e-05,
324
+ "loss": 3.8372,
325
+ "step": 450
326
+ },
327
+ {
328
+ "epoch": 0.8440366972477065,
329
+ "grad_norm": 0.1256406158208847,
330
+ "learning_rate": 6.673432932217117e-05,
331
+ "loss": 3.8335,
332
+ "step": 460
333
+ },
334
+ {
335
+ "epoch": 0.8623853211009175,
336
+ "grad_norm": 0.12184158712625504,
337
+ "learning_rate": 5.2378529625954684e-05,
338
+ "loss": 3.8197,
339
+ "step": 470
340
+ },
341
+ {
342
+ "epoch": 0.8807339449541285,
343
+ "grad_norm": 0.11440487205982208,
344
+ "learning_rate": 3.9675056661785556e-05,
345
+ "loss": 3.8294,
346
+ "step": 480
347
+ },
348
+ {
349
+ "epoch": 0.8990825688073395,
350
+ "grad_norm": 0.12027451395988464,
351
+ "learning_rate": 2.8670803372379538e-05,
352
+ "loss": 3.8275,
353
+ "step": 490
354
+ },
355
+ {
356
+ "epoch": 0.9174311926605505,
357
+ "grad_norm": 0.12016978859901428,
358
+ "learning_rate": 1.9406390288876586e-05,
359
+ "loss": 3.8396,
360
+ "step": 500
361
+ },
362
+ {
363
+ "epoch": 0.9357798165137615,
364
+ "grad_norm": 0.11837776750326157,
365
+ "learning_rate": 1.191601558631461e-05,
366
+ "loss": 3.8321,
367
+ "step": 510
368
+ },
369
+ {
370
+ "epoch": 0.9541284403669725,
371
+ "grad_norm": 0.11801979690790176,
372
+ "learning_rate": 6.2273288462559305e-06,
373
+ "loss": 3.8249,
374
+ "step": 520
375
+ },
376
+ {
377
+ "epoch": 0.9724770642201835,
378
+ "grad_norm": 0.11723949760198593,
379
+ "learning_rate": 2.361328992552314e-06,
380
+ "loss": 3.8284,
381
+ "step": 530
382
+ },
383
+ {
384
+ "epoch": 0.9908256880733946,
385
+ "grad_norm": 0.11505445837974548,
386
+ "learning_rate": 3.322867770044202e-07,
387
+ "loss": 3.827,
388
+ "step": 540
389
+ },
390
+ {
391
+ "epoch": 1.0,
392
+ "step": 545,
393
+ "total_flos": 3.6662326867892634e+17,
394
+ "train_loss": 4.2043135091799115,
395
+ "train_runtime": 5594.5203,
396
+ "train_samples_per_second": 99.763,
397
+ "train_steps_per_second": 0.097
398
+ }
399
+ ],
400
+ "logging_steps": 10,
401
+ "max_steps": 545,
402
+ "num_input_tokens_seen": 0,
403
+ "num_train_epochs": 1,
404
+ "save_steps": 500,
405
+ "stateful_callbacks": {
406
+ "TrainerControl": {
407
+ "args": {
408
+ "should_epoch_stop": false,
409
+ "should_evaluate": false,
410
+ "should_log": false,
411
+ "should_save": false,
412
+ "should_training_stop": false
413
+ },
414
+ "attributes": {}
415
+ }
416
+ },
417
+ "total_flos": 3.6662326867892634e+17,
418
+ "train_batch_size": 64,
419
+ "trial_name": null,
420
+ "trial_params": null
421
+ }
continuous/qwen3/siglip2_sm14_384/finetune/added_tokens.json ADDED
@@ -0,0 +1,28 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "</think>": 151668,
3
+ "</tool_call>": 151658,
4
+ "</tool_response>": 151666,
5
+ "<think>": 151667,
6
+ "<tool_call>": 151657,
7
+ "<tool_response>": 151665,
8
+ "<|box_end|>": 151649,
9
+ "<|box_start|>": 151648,
10
+ "<|endoftext|>": 151643,
11
+ "<|file_sep|>": 151664,
12
+ "<|fim_middle|>": 151660,
13
+ "<|fim_pad|>": 151662,
14
+ "<|fim_prefix|>": 151659,
15
+ "<|fim_suffix|>": 151661,
16
+ "<|im_end|>": 151645,
17
+ "<|im_start|>": 151644,
18
+ "<|image_pad|>": 151655,
19
+ "<|object_ref_end|>": 151647,
20
+ "<|object_ref_start|>": 151646,
21
+ "<|quad_end|>": 151651,
22
+ "<|quad_start|>": 151650,
23
+ "<|repo_name|>": 151663,
24
+ "<|video_pad|>": 151656,
25
+ "<|vision_end|>": 151653,
26
+ "<|vision_pad|>": 151654,
27
+ "<|vision_start|>": 151652
28
+ }
continuous/qwen3/siglip2_sm14_384/finetune/config.json ADDED
@@ -0,0 +1,59 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "add_faster_video": false,
3
+ "add_time_instruction": false,
4
+ "architectures": [
5
+ "LlavaQwen3ForCausalLM"
6
+ ],
7
+ "attention_bias": false,
8
+ "attention_dropout": 0.0,
9
+ "bos_token_id": 151643,
10
+ "eos_token_id": 151645,
11
+ "faster_token_stride": 10,
12
+ "force_sample": false,
13
+ "head_dim": 128,
14
+ "hidden_act": "silu",
15
+ "hidden_size": 2048,
16
+ "image_aspect_ratio": "square",
17
+ "image_crop_resolution": null,
18
+ "image_grid_pinpoints": null,
19
+ "image_split_resolution": null,
20
+ "initializer_range": 0.02,
21
+ "intermediate_size": 6144,
22
+ "max_position_embeddings": 40960,
23
+ "max_window_layers": 28,
24
+ "mm_hidden_size": 1152,
25
+ "mm_newline_position": "grid",
26
+ "mm_patch_merge_type": "flat",
27
+ "mm_projector_lr": 0.0001,
28
+ "mm_projector_type": "mlp2x_gelu",
29
+ "mm_resampler_type": null,
30
+ "mm_spatial_pool_mode": "bilinear",
31
+ "mm_spatial_pool_stride": null,
32
+ "mm_tunable_parts": "mm_mlp_adapter,mm_language_model",
33
+ "mm_use_im_patch_token": false,
34
+ "mm_use_im_start_end": false,
35
+ "mm_vision_select_feature": "patch",
36
+ "mm_vision_select_layer": -2,
37
+ "mm_vision_tower": "open_clip_hub:ViT-SO400M-14-SigLIP2-378",
38
+ "mm_vision_tower_lr": null,
39
+ "model_type": "qwen3",
40
+ "num_attention_heads": 16,
41
+ "num_hidden_layers": 28,
42
+ "num_key_value_heads": 8,
43
+ "pos_skipping_range": 4096,
44
+ "rms_norm_eps": 1e-06,
45
+ "rope_scaling": null,
46
+ "rope_theta": 1000000,
47
+ "sliding_window": null,
48
+ "tie_word_embeddings": true,
49
+ "tokenizer_model_max_length": 8192,
50
+ "tokenizer_padding_side": "right",
51
+ "torch_dtype": "bfloat16",
52
+ "transformers_version": "4.51.3",
53
+ "use_cache": true,
54
+ "use_mm_proj": true,
55
+ "use_pos_skipping": false,
56
+ "use_sliding_window": false,
57
+ "vision_tower_pretrained": "/cache/ckpt/download/tokenizer/continuous/siglip2/siglip2_sm14_384.npz",
58
+ "vocab_size": 151936
59
+ }
continuous/qwen3/siglip2_sm14_384/finetune/generation_config.json ADDED
@@ -0,0 +1,13 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "bos_token_id": 151643,
3
+ "do_sample": true,
4
+ "eos_token_id": [
5
+ 151645,
6
+ 151643
7
+ ],
8
+ "pad_token_id": 151643,
9
+ "temperature": 0.6,
10
+ "top_k": 20,
11
+ "top_p": 0.95,
12
+ "transformers_version": "4.51.3"
13
+ }
continuous/qwen3/siglip2_sm14_384/finetune/merges.txt ADDED
The diff for this file is too large to render. See raw diff
 
continuous/qwen3/siglip2_sm14_384/finetune/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4626453da87ba96756191aa52526a5c99fa0cbf46fdce407a97fb7f04b50a92b
3
+ size 4946244496
continuous/qwen3/siglip2_sm14_384/finetune/special_tokens_map.json ADDED
@@ -0,0 +1,31 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "additional_special_tokens": [
3
+ "<|im_start|>",
4
+ "<|im_end|>",
5
+ "<|object_ref_start|>",
6
+ "<|object_ref_end|>",
7
+ "<|box_start|>",
8
+ "<|box_end|>",
9
+ "<|quad_start|>",
10
+ "<|quad_end|>",
11
+ "<|vision_start|>",
12
+ "<|vision_end|>",
13
+ "<|vision_pad|>",
14
+ "<|image_pad|>",
15
+ "<|video_pad|>"
16
+ ],
17
+ "eos_token": {
18
+ "content": "<|im_end|>",
19
+ "lstrip": false,
20
+ "normalized": false,
21
+ "rstrip": false,
22
+ "single_word": false
23
+ },
24
+ "pad_token": {
25
+ "content": "<|endoftext|>",
26
+ "lstrip": false,
27
+ "normalized": false,
28
+ "rstrip": false,
29
+ "single_word": false
30
+ }
31
+ }
continuous/qwen3/siglip2_sm14_384/finetune/tokenizer.json ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:aeb13307a71acd8fe81861d94ad54ab689df773318809eed3cbe794b4492dae4
3
+ size 11422654
continuous/qwen3/siglip2_sm14_384/finetune/tokenizer_config.json ADDED
@@ -0,0 +1,241 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "add_bos_token": false,
3
+ "add_prefix_space": false,
4
+ "added_tokens_decoder": {
5
+ "151643": {
6
+ "content": "<|endoftext|>",
7
+ "lstrip": false,
8
+ "normalized": false,
9
+ "rstrip": false,
10
+ "single_word": false,
11
+ "special": true
12
+ },
13
+ "151644": {
14
+ "content": "<|im_start|>",
15
+ "lstrip": false,
16
+ "normalized": false,
17
+ "rstrip": false,
18
+ "single_word": false,
19
+ "special": true
20
+ },
21
+ "151645": {
22
+ "content": "<|im_end|>",
23
+ "lstrip": false,
24
+ "normalized": false,
25
+ "rstrip": false,
26
+ "single_word": false,
27
+ "special": true
28
+ },
29
+ "151646": {
30
+ "content": "<|object_ref_start|>",
31
+ "lstrip": false,
32
+ "normalized": false,
33
+ "rstrip": false,
34
+ "single_word": false,
35
+ "special": true
36
+ },
37
+ "151647": {
38
+ "content": "<|object_ref_end|>",
39
+ "lstrip": false,
40
+ "normalized": false,
41
+ "rstrip": false,
42
+ "single_word": false,
43
+ "special": true
44
+ },
45
+ "151648": {
46
+ "content": "<|box_start|>",
47
+ "lstrip": false,
48
+ "normalized": false,
49
+ "rstrip": false,
50
+ "single_word": false,
51
+ "special": true
52
+ },
53
+ "151649": {
54
+ "content": "<|box_end|>",
55
+ "lstrip": false,
56
+ "normalized": false,
57
+ "rstrip": false,
58
+ "single_word": false,
59
+ "special": true
60
+ },
61
+ "151650": {
62
+ "content": "<|quad_start|>",
63
+ "lstrip": false,
64
+ "normalized": false,
65
+ "rstrip": false,
66
+ "single_word": false,
67
+ "special": true
68
+ },
69
+ "151651": {
70
+ "content": "<|quad_end|>",
71
+ "lstrip": false,
72
+ "normalized": false,
73
+ "rstrip": false,
74
+ "single_word": false,
75
+ "special": true
76
+ },
77
+ "151652": {
78
+ "content": "<|vision_start|>",
79
+ "lstrip": false,
80
+ "normalized": false,
81
+ "rstrip": false,
82
+ "single_word": false,
83
+ "special": true
84
+ },
85
+ "151653": {
86
+ "content": "<|vision_end|>",
87
+ "lstrip": false,
88
+ "normalized": false,
89
+ "rstrip": false,
90
+ "single_word": false,
91
+ "special": true
92
+ },
93
+ "151654": {
94
+ "content": "<|vision_pad|>",
95
+ "lstrip": false,
96
+ "normalized": false,
97
+ "rstrip": false,
98
+ "single_word": false,
99
+ "special": true
100
+ },
101
+ "151655": {
102
+ "content": "<|image_pad|>",
103
+ "lstrip": false,
104
+ "normalized": false,
105
+ "rstrip": false,
106
+ "single_word": false,
107
+ "special": true
108
+ },
109
+ "151656": {
110
+ "content": "<|video_pad|>",
111
+ "lstrip": false,
112
+ "normalized": false,
113
+ "rstrip": false,
114
+ "single_word": false,
115
+ "special": true
116
+ },
117
+ "151657": {
118
+ "content": "<tool_call>",
119
+ "lstrip": false,
120
+ "normalized": false,
121
+ "rstrip": false,
122
+ "single_word": false,
123
+ "special": false
124
+ },
125
+ "151658": {
126
+ "content": "</tool_call>",
127
+ "lstrip": false,
128
+ "normalized": false,
129
+ "rstrip": false,
130
+ "single_word": false,
131
+ "special": false
132
+ },
133
+ "151659": {
134
+ "content": "<|fim_prefix|>",
135
+ "lstrip": false,
136
+ "normalized": false,
137
+ "rstrip": false,
138
+ "single_word": false,
139
+ "special": false
140
+ },
141
+ "151660": {
142
+ "content": "<|fim_middle|>",
143
+ "lstrip": false,
144
+ "normalized": false,
145
+ "rstrip": false,
146
+ "single_word": false,
147
+ "special": false
148
+ },
149
+ "151661": {
150
+ "content": "<|fim_suffix|>",
151
+ "lstrip": false,
152
+ "normalized": false,
153
+ "rstrip": false,
154
+ "single_word": false,
155
+ "special": false
156
+ },
157
+ "151662": {
158
+ "content": "<|fim_pad|>",
159
+ "lstrip": false,
160
+ "normalized": false,
161
+ "rstrip": false,
162
+ "single_word": false,
163
+ "special": false
164
+ },
165
+ "151663": {
166
+ "content": "<|repo_name|>",
167
+ "lstrip": false,
168
+ "normalized": false,
169
+ "rstrip": false,
170
+ "single_word": false,
171
+ "special": false
172
+ },
173
+ "151664": {
174
+ "content": "<|file_sep|>",
175
+ "lstrip": false,
176
+ "normalized": false,
177
+ "rstrip": false,
178
+ "single_word": false,
179
+ "special": false
180
+ },
181
+ "151665": {
182
+ "content": "<tool_response>",
183
+ "lstrip": false,
184
+ "normalized": false,
185
+ "rstrip": false,
186
+ "single_word": false,
187
+ "special": false
188
+ },
189
+ "151666": {
190
+ "content": "</tool_response>",
191
+ "lstrip": false,
192
+ "normalized": false,
193
+ "rstrip": false,
194
+ "single_word": false,
195
+ "special": false
196
+ },
197
+ "151667": {
198
+ "content": "<think>",
199
+ "lstrip": false,
200
+ "normalized": false,
201
+ "rstrip": false,
202
+ "single_word": false,
203
+ "special": false
204
+ },
205
+ "151668": {
206
+ "content": "</think>",
207
+ "lstrip": false,
208
+ "normalized": false,
209
+ "rstrip": false,
210
+ "single_word": false,
211
+ "special": false
212
+ }
213
+ },
214
+ "additional_special_tokens": [
215
+ "<|im_start|>",
216
+ "<|im_end|>",
217
+ "<|object_ref_start|>",
218
+ "<|object_ref_end|>",
219
+ "<|box_start|>",
220
+ "<|box_end|>",
221
+ "<|quad_start|>",
222
+ "<|quad_end|>",
223
+ "<|vision_start|>",
224
+ "<|vision_end|>",
225
+ "<|vision_pad|>",
226
+ "<|image_pad|>",
227
+ "<|video_pad|>"
228
+ ],
229
+ "bos_token": null,
230
+ "chat_template": "{%- if tools %}\n {{- '<|im_start|>system\\n' }}\n {%- if messages[0].role == 'system' %}\n {{- messages[0].content + '\\n\\n' }}\n {%- endif %}\n {{- \"# Tools\\n\\nYou may call one or more functions to assist with the user query.\\n\\nYou are provided with function signatures within <tools></tools> XML tags:\\n<tools>\" }}\n {%- for tool in tools %}\n {{- \"\\n\" }}\n {{- tool | tojson }}\n {%- endfor %}\n {{- \"\\n</tools>\\n\\nFor each function call, return a json object with function name and arguments within <tool_call></tool_call> XML tags:\\n<tool_call>\\n{\\\"name\\\": <function-name>, \\\"arguments\\\": <args-json-object>}\\n</tool_call><|im_end|>\\n\" }}\n{%- else %}\n {%- if messages[0].role == 'system' %}\n {{- '<|im_start|>system\\n' + messages[0].content + '<|im_end|>\\n' }}\n {%- endif %}\n{%- endif %}\n{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}\n{%- for message in messages[::-1] %}\n {%- set index = (messages|length - 1) - loop.index0 %}\n {%- if ns.multi_step_tool and message.role == \"user\" and message.content is string and not(message.content.startswith('<tool_response>') and message.content.endswith('</tool_response>')) %}\n {%- set ns.multi_step_tool = false %}\n {%- set ns.last_query_index = index %}\n {%- endif %}\n{%- endfor %}\n{%- for message in messages %}\n {%- if message.content is string %}\n {%- set content = message.content %}\n {%- else %}\n {%- set content = '' %}\n {%- endif %}\n {%- if (message.role == \"user\") or (message.role == \"system\" and not loop.first) %}\n {{- '<|im_start|>' + message.role + '\\n' + content + '<|im_end|>' + '\\n' }}\n {%- elif message.role == \"assistant\" %}\n {%- set reasoning_content = '' %}\n {%- if message.reasoning_content is string %}\n {%- set reasoning_content = message.reasoning_content %}\n {%- else %}\n {%- if '</think>' in content %}\n {%- set reasoning_content = content.split('</think>')[0].rstrip('\\n').split('<think>')[-1].lstrip('\\n') %}\n {%- set content = content.split('</think>')[-1].lstrip('\\n') %}\n {%- endif %}\n {%- endif %}\n {%- if loop.index0 > ns.last_query_index %}\n {%- if loop.last or (not loop.last and reasoning_content) %}\n {{- '<|im_start|>' + message.role + '\\n<think>\\n' + reasoning_content.strip('\\n') + '\\n</think>\\n\\n' + content.lstrip('\\n') }}\n {%- else %}\n {{- '<|im_start|>' + message.role + '\\n' + content }}\n {%- endif %}\n {%- else %}\n {{- '<|im_start|>' + message.role + '\\n' + content }}\n {%- endif %}\n {%- if message.tool_calls %}\n {%- for tool_call in message.tool_calls %}\n {%- if (loop.first and content) or (not loop.first) %}\n {{- '\\n' }}\n {%- endif %}\n {%- if tool_call.function %}\n {%- set tool_call = tool_call.function %}\n {%- endif %}\n {{- '<tool_call>\\n{\"name\": \"' }}\n {{- tool_call.name }}\n {{- '\", \"arguments\": ' }}\n {%- if tool_call.arguments is string %}\n {{- tool_call.arguments }}\n {%- else %}\n {{- tool_call.arguments | tojson }}\n {%- endif %}\n {{- '}\\n</tool_call>' }}\n {%- endfor %}\n {%- endif %}\n {{- '<|im_end|>\\n' }}\n {%- elif message.role == \"tool\" %}\n {%- if loop.first or (messages[loop.index0 - 1].role != \"tool\") %}\n {{- '<|im_start|>user' }}\n {%- endif %}\n {{- '\\n<tool_response>\\n' }}\n {{- content }}\n {{- '\\n</tool_response>' }}\n {%- if loop.last or (messages[loop.index0 + 1].role != \"tool\") %}\n {{- '<|im_end|>\\n' }}\n {%- endif %}\n {%- endif %}\n{%- endfor %}\n{%- if add_generation_prompt %}\n {{- '<|im_start|>assistant\\n' }}\n {%- if enable_thinking is defined and enable_thinking is false %}\n {{- '<think>\\n\\n</think>\\n\\n' }}\n {%- endif %}\n{%- endif %}",
231
+ "clean_up_tokenization_spaces": false,
232
+ "eos_token": "<|im_end|>",
233
+ "errors": "replace",
234
+ "extra_special_tokens": {},
235
+ "model_max_length": 8192,
236
+ "pad_token": "<|endoftext|>",
237
+ "padding_side": "right",
238
+ "split_special_tokens": false,
239
+ "tokenizer_class": "Qwen2Tokenizer",
240
+ "unk_token": null
241
+ }
continuous/qwen3/siglip2_sm14_384/finetune/trainer_state.json ADDED
@@ -0,0 +1,1856 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_global_step": null,
3
+ "best_metric": null,
4
+ "best_model_checkpoint": null,
5
+ "epoch": 0.9999037443449803,
6
+ "eval_steps": 500,
7
+ "global_step": 2597,
8
+ "is_hyper_param_search": false,
9
+ "is_local_process_zero": true,
10
+ "is_world_process_zero": true,
11
+ "log_history": [
12
+ {
13
+ "epoch": 0.003850226200789296,
14
+ "grad_norm": 8.169692039489746,
15
+ "learning_rate": 1.153846153846154e-06,
16
+ "loss": 2.1466,
17
+ "step": 10
18
+ },
19
+ {
20
+ "epoch": 0.007700452401578592,
21
+ "grad_norm": 3.3677866458892822,
22
+ "learning_rate": 2.435897435897436e-06,
23
+ "loss": 1.7477,
24
+ "step": 20
25
+ },
26
+ {
27
+ "epoch": 0.01155067860236789,
28
+ "grad_norm": 2.1976206302642822,
29
+ "learning_rate": 3.7179487179487184e-06,
30
+ "loss": 1.3827,
31
+ "step": 30
32
+ },
33
+ {
34
+ "epoch": 0.015400904803157185,
35
+ "grad_norm": 0.43533825874328613,
36
+ "learning_rate": 5e-06,
37
+ "loss": 1.0465,
38
+ "step": 40
39
+ },
40
+ {
41
+ "epoch": 0.019251131003946483,
42
+ "grad_norm": 0.4479895532131195,
43
+ "learning_rate": 6.282051282051282e-06,
44
+ "loss": 0.97,
45
+ "step": 50
46
+ },
47
+ {
48
+ "epoch": 0.02310135720473578,
49
+ "grad_norm": 0.2332359403371811,
50
+ "learning_rate": 7.564102564102564e-06,
51
+ "loss": 0.8981,
52
+ "step": 60
53
+ },
54
+ {
55
+ "epoch": 0.026951583405525073,
56
+ "grad_norm": 0.21232149004936218,
57
+ "learning_rate": 8.846153846153847e-06,
58
+ "loss": 0.8607,
59
+ "step": 70
60
+ },
61
+ {
62
+ "epoch": 0.03080180960631437,
63
+ "grad_norm": 0.175906240940094,
64
+ "learning_rate": 9.999996111488723e-06,
65
+ "loss": 0.8634,
66
+ "step": 80
67
+ },
68
+ {
69
+ "epoch": 0.034652035807103666,
70
+ "grad_norm": 0.17599861323833466,
71
+ "learning_rate": 9.999529497453782e-06,
72
+ "loss": 0.8414,
73
+ "step": 90
74
+ },
75
+ {
76
+ "epoch": 0.038502262007892966,
77
+ "grad_norm": 0.15849661827087402,
78
+ "learning_rate": 9.998285264324193e-06,
79
+ "loss": 0.8066,
80
+ "step": 100
81
+ },
82
+ {
83
+ "epoch": 0.04235248820868226,
84
+ "grad_norm": 0.16509424149990082,
85
+ "learning_rate": 9.996263605626052e-06,
86
+ "loss": 0.7913,
87
+ "step": 110
88
+ },
89
+ {
90
+ "epoch": 0.04620271440947156,
91
+ "grad_norm": 0.14889495074748993,
92
+ "learning_rate": 9.99346483580503e-06,
93
+ "loss": 0.8007,
94
+ "step": 120
95
+ },
96
+ {
97
+ "epoch": 0.05005294061026085,
98
+ "grad_norm": 0.1441938877105713,
99
+ "learning_rate": 9.989889390177467e-06,
100
+ "loss": 0.7896,
101
+ "step": 130
102
+ },
103
+ {
104
+ "epoch": 0.053903166811050146,
105
+ "grad_norm": 0.15888895094394684,
106
+ "learning_rate": 9.985537824862647e-06,
107
+ "loss": 0.7919,
108
+ "step": 140
109
+ },
110
+ {
111
+ "epoch": 0.057753393011839446,
112
+ "grad_norm": 0.17070317268371582,
113
+ "learning_rate": 9.980410816696317e-06,
114
+ "loss": 0.7811,
115
+ "step": 150
116
+ },
117
+ {
118
+ "epoch": 0.06160361921262874,
119
+ "grad_norm": 0.14946454763412476,
120
+ "learning_rate": 9.97450916312541e-06,
121
+ "loss": 0.7793,
122
+ "step": 160
123
+ },
124
+ {
125
+ "epoch": 0.06545384541341803,
126
+ "grad_norm": 0.1282702535390854,
127
+ "learning_rate": 9.967833782084004e-06,
128
+ "loss": 0.78,
129
+ "step": 170
130
+ },
131
+ {
132
+ "epoch": 0.06930407161420733,
133
+ "grad_norm": 0.14094731211662292,
134
+ "learning_rate": 9.960385711850554e-06,
135
+ "loss": 0.7985,
136
+ "step": 180
137
+ },
138
+ {
139
+ "epoch": 0.07315429781499663,
140
+ "grad_norm": 0.1605401188135147,
141
+ "learning_rate": 9.952166110886394e-06,
142
+ "loss": 0.8014,
143
+ "step": 190
144
+ },
145
+ {
146
+ "epoch": 0.07700452401578593,
147
+ "grad_norm": 0.14923018217086792,
148
+ "learning_rate": 9.943176257655567e-06,
149
+ "loss": 0.7773,
150
+ "step": 200
151
+ },
152
+ {
153
+ "epoch": 0.08085475021657522,
154
+ "grad_norm": 0.1422853320837021,
155
+ "learning_rate": 9.933417550425952e-06,
156
+ "loss": 0.7904,
157
+ "step": 210
158
+ },
159
+ {
160
+ "epoch": 0.08470497641736452,
161
+ "grad_norm": 0.1399507075548172,
162
+ "learning_rate": 9.9228915070518e-06,
163
+ "loss": 0.7618,
164
+ "step": 220
165
+ },
166
+ {
167
+ "epoch": 0.08855520261815382,
168
+ "grad_norm": 0.15950343012809753,
169
+ "learning_rate": 9.911599764737632e-06,
170
+ "loss": 0.7675,
171
+ "step": 230
172
+ },
173
+ {
174
+ "epoch": 0.09240542881894312,
175
+ "grad_norm": 0.16157372295856476,
176
+ "learning_rate": 9.899544079783606e-06,
177
+ "loss": 0.7537,
178
+ "step": 240
179
+ },
180
+ {
181
+ "epoch": 0.0962556550197324,
182
+ "grad_norm": 0.142499640583992,
183
+ "learning_rate": 9.886726327312338e-06,
184
+ "loss": 0.7758,
185
+ "step": 250
186
+ },
187
+ {
188
+ "epoch": 0.1001058812205217,
189
+ "grad_norm": 0.1363229602575302,
190
+ "learning_rate": 9.87314850097724e-06,
191
+ "loss": 0.7755,
192
+ "step": 260
193
+ },
194
+ {
195
+ "epoch": 0.103956107421311,
196
+ "grad_norm": 0.14040303230285645,
197
+ "learning_rate": 9.858812712652452e-06,
198
+ "loss": 0.7603,
199
+ "step": 270
200
+ },
201
+ {
202
+ "epoch": 0.10780633362210029,
203
+ "grad_norm": 0.14070791006088257,
204
+ "learning_rate": 9.843721192104341e-06,
205
+ "loss": 0.761,
206
+ "step": 280
207
+ },
208
+ {
209
+ "epoch": 0.11165655982288959,
210
+ "grad_norm": 0.1377442181110382,
211
+ "learning_rate": 9.827876286644698e-06,
212
+ "loss": 0.7776,
213
+ "step": 290
214
+ },
215
+ {
216
+ "epoch": 0.11550678602367889,
217
+ "grad_norm": 0.13174867630004883,
218
+ "learning_rate": 9.811280460765643e-06,
219
+ "loss": 0.7399,
220
+ "step": 300
221
+ },
222
+ {
223
+ "epoch": 0.11935701222446819,
224
+ "grad_norm": 0.1321926712989807,
225
+ "learning_rate": 9.793936295756292e-06,
226
+ "loss": 0.7529,
227
+ "step": 310
228
+ },
229
+ {
230
+ "epoch": 0.12320723842525748,
231
+ "grad_norm": 0.13814984261989594,
232
+ "learning_rate": 9.775846489301279e-06,
233
+ "loss": 0.7559,
234
+ "step": 320
235
+ },
236
+ {
237
+ "epoch": 0.12705746462604678,
238
+ "grad_norm": 0.14317971467971802,
239
+ "learning_rate": 9.757013855061157e-06,
240
+ "loss": 0.7561,
241
+ "step": 330
242
+ },
243
+ {
244
+ "epoch": 0.13090769082683607,
245
+ "grad_norm": 0.1345769762992859,
246
+ "learning_rate": 9.737441322234758e-06,
247
+ "loss": 0.7623,
248
+ "step": 340
249
+ },
250
+ {
251
+ "epoch": 0.13475791702762538,
252
+ "grad_norm": 0.15007136762142181,
253
+ "learning_rate": 9.717131935103605e-06,
254
+ "loss": 0.7697,
255
+ "step": 350
256
+ },
257
+ {
258
+ "epoch": 0.13860814322841467,
259
+ "grad_norm": 0.13195142149925232,
260
+ "learning_rate": 9.696088852558398e-06,
261
+ "loss": 0.7541,
262
+ "step": 360
263
+ },
264
+ {
265
+ "epoch": 0.14245836942920398,
266
+ "grad_norm": 0.13849051296710968,
267
+ "learning_rate": 9.674315347607686e-06,
268
+ "loss": 0.7801,
269
+ "step": 370
270
+ },
271
+ {
272
+ "epoch": 0.14630859562999327,
273
+ "grad_norm": 0.13180379569530487,
274
+ "learning_rate": 9.651814806868796e-06,
275
+ "loss": 0.7512,
276
+ "step": 380
277
+ },
278
+ {
279
+ "epoch": 0.15015882183078255,
280
+ "grad_norm": 0.14003239572048187,
281
+ "learning_rate": 9.62859073004107e-06,
282
+ "loss": 0.7488,
283
+ "step": 390
284
+ },
285
+ {
286
+ "epoch": 0.15400904803157187,
287
+ "grad_norm": 0.13475312292575836,
288
+ "learning_rate": 9.604646729361547e-06,
289
+ "loss": 0.7495,
290
+ "step": 400
291
+ },
292
+ {
293
+ "epoch": 0.15785927423236115,
294
+ "grad_norm": 0.13502562046051025,
295
+ "learning_rate": 9.579986529043099e-06,
296
+ "loss": 0.7299,
297
+ "step": 410
298
+ },
299
+ {
300
+ "epoch": 0.16170950043315044,
301
+ "grad_norm": 0.13956373929977417,
302
+ "learning_rate": 9.554613964695189e-06,
303
+ "loss": 0.7541,
304
+ "step": 420
305
+ },
306
+ {
307
+ "epoch": 0.16555972663393975,
308
+ "grad_norm": 0.1389012336730957,
309
+ "learning_rate": 9.528532982727281e-06,
310
+ "loss": 0.748,
311
+ "step": 430
312
+ },
313
+ {
314
+ "epoch": 0.16940995283472904,
315
+ "grad_norm": 0.1420457512140274,
316
+ "learning_rate": 9.501747639735018e-06,
317
+ "loss": 0.7402,
318
+ "step": 440
319
+ },
320
+ {
321
+ "epoch": 0.17326017903551832,
322
+ "grad_norm": 0.1352194994688034,
323
+ "learning_rate": 9.474262101869272e-06,
324
+ "loss": 0.7285,
325
+ "step": 450
326
+ },
327
+ {
328
+ "epoch": 0.17711040523630764,
329
+ "grad_norm": 0.13821879029273987,
330
+ "learning_rate": 9.446080644188141e-06,
331
+ "loss": 0.7477,
332
+ "step": 460
333
+ },
334
+ {
335
+ "epoch": 0.18096063143709692,
336
+ "grad_norm": 0.1354229897260666,
337
+ "learning_rate": 9.417207649992017e-06,
338
+ "loss": 0.7466,
339
+ "step": 470
340
+ },
341
+ {
342
+ "epoch": 0.18481085763788624,
343
+ "grad_norm": 0.1266239583492279,
344
+ "learning_rate": 9.387647610141816e-06,
345
+ "loss": 0.7376,
346
+ "step": 480
347
+ },
348
+ {
349
+ "epoch": 0.18866108383867553,
350
+ "grad_norm": 0.12634965777397156,
351
+ "learning_rate": 9.35740512236047e-06,
352
+ "loss": 0.733,
353
+ "step": 490
354
+ },
355
+ {
356
+ "epoch": 0.1925113100394648,
357
+ "grad_norm": 0.1283612698316574,
358
+ "learning_rate": 9.326484890517809e-06,
359
+ "loss": 0.7392,
360
+ "step": 500
361
+ },
362
+ {
363
+ "epoch": 0.19636153624025413,
364
+ "grad_norm": 0.1307617574930191,
365
+ "learning_rate": 9.294891723898926e-06,
366
+ "loss": 0.7216,
367
+ "step": 510
368
+ },
369
+ {
370
+ "epoch": 0.2002117624410434,
371
+ "grad_norm": 0.13645093142986298,
372
+ "learning_rate": 9.262630536456153e-06,
373
+ "loss": 0.7611,
374
+ "step": 520
375
+ },
376
+ {
377
+ "epoch": 0.2040619886418327,
378
+ "grad_norm": 0.14038141071796417,
379
+ "learning_rate": 9.229706346044749e-06,
380
+ "loss": 0.7311,
381
+ "step": 530
382
+ },
383
+ {
384
+ "epoch": 0.207912214842622,
385
+ "grad_norm": 0.13900795578956604,
386
+ "learning_rate": 9.196124273642422e-06,
387
+ "loss": 0.7487,
388
+ "step": 540
389
+ },
390
+ {
391
+ "epoch": 0.2117624410434113,
392
+ "grad_norm": 0.12451288849115372,
393
+ "learning_rate": 9.161889542552842e-06,
394
+ "loss": 0.7597,
395
+ "step": 550
396
+ },
397
+ {
398
+ "epoch": 0.21561266724420058,
399
+ "grad_norm": 0.1350848525762558,
400
+ "learning_rate": 9.127007477593193e-06,
401
+ "loss": 0.7523,
402
+ "step": 560
403
+ },
404
+ {
405
+ "epoch": 0.2194628934449899,
406
+ "grad_norm": 0.14007258415222168,
407
+ "learning_rate": 9.091483504265974e-06,
408
+ "loss": 0.7572,
409
+ "step": 570
410
+ },
411
+ {
412
+ "epoch": 0.22331311964577918,
413
+ "grad_norm": 0.13169001042842865,
414
+ "learning_rate": 9.055323147915116e-06,
415
+ "loss": 0.7371,
416
+ "step": 580
417
+ },
418
+ {
419
+ "epoch": 0.2271633458465685,
420
+ "grad_norm": 0.14079685509204865,
421
+ "learning_rate": 9.018532032866585e-06,
422
+ "loss": 0.7501,
423
+ "step": 590
424
+ },
425
+ {
426
+ "epoch": 0.23101357204735778,
427
+ "grad_norm": 0.13430313766002655,
428
+ "learning_rate": 8.981115881553579e-06,
429
+ "loss": 0.7459,
430
+ "step": 600
431
+ },
432
+ {
433
+ "epoch": 0.23486379824814707,
434
+ "grad_norm": 0.1424923539161682,
435
+ "learning_rate": 8.943080513626476e-06,
436
+ "loss": 0.7602,
437
+ "step": 610
438
+ },
439
+ {
440
+ "epoch": 0.23871402444893638,
441
+ "grad_norm": 0.12634651362895966,
442
+ "learning_rate": 8.904431845047641e-06,
443
+ "loss": 0.7509,
444
+ "step": 620
445
+ },
446
+ {
447
+ "epoch": 0.24256425064972567,
448
+ "grad_norm": 0.1293267458677292,
449
+ "learning_rate": 8.86517588717128e-06,
450
+ "loss": 0.7549,
451
+ "step": 630
452
+ },
453
+ {
454
+ "epoch": 0.24641447685051496,
455
+ "grad_norm": 0.12834975123405457,
456
+ "learning_rate": 8.82531874580844e-06,
457
+ "loss": 0.7581,
458
+ "step": 640
459
+ },
460
+ {
461
+ "epoch": 0.25026470305130427,
462
+ "grad_norm": 0.12090420722961426,
463
+ "learning_rate": 8.784866620277314e-06,
464
+ "loss": 0.7119,
465
+ "step": 650
466
+ },
467
+ {
468
+ "epoch": 0.25411492925209356,
469
+ "grad_norm": 0.13798002898693085,
470
+ "learning_rate": 8.743825802439018e-06,
471
+ "loss": 0.7271,
472
+ "step": 660
473
+ },
474
+ {
475
+ "epoch": 0.25796515545288284,
476
+ "grad_norm": 0.1354415863752365,
477
+ "learning_rate": 8.702202675718956e-06,
478
+ "loss": 0.7526,
479
+ "step": 670
480
+ },
481
+ {
482
+ "epoch": 0.26181538165367213,
483
+ "grad_norm": 0.12562601268291473,
484
+ "learning_rate": 8.660003714113959e-06,
485
+ "loss": 0.7236,
486
+ "step": 680
487
+ },
488
+ {
489
+ "epoch": 0.26566560785446147,
490
+ "grad_norm": 0.13245192170143127,
491
+ "learning_rate": 8.617235481185323e-06,
492
+ "loss": 0.7357,
493
+ "step": 690
494
+ },
495
+ {
496
+ "epoch": 0.26951583405525076,
497
+ "grad_norm": 0.13617967069149017,
498
+ "learning_rate": 8.573904629037931e-06,
499
+ "loss": 0.7448,
500
+ "step": 700
501
+ },
502
+ {
503
+ "epoch": 0.27336606025604004,
504
+ "grad_norm": 0.12710173428058624,
505
+ "learning_rate": 8.530017897285585e-06,
506
+ "loss": 0.7293,
507
+ "step": 710
508
+ },
509
+ {
510
+ "epoch": 0.27721628645682933,
511
+ "grad_norm": 0.12039714306592941,
512
+ "learning_rate": 8.485582112002746e-06,
513
+ "loss": 0.7398,
514
+ "step": 720
515
+ },
516
+ {
517
+ "epoch": 0.2810665126576186,
518
+ "grad_norm": 0.13378754258155823,
519
+ "learning_rate": 8.440604184662811e-06,
520
+ "loss": 0.7405,
521
+ "step": 730
522
+ },
523
+ {
524
+ "epoch": 0.28491673885840796,
525
+ "grad_norm": 0.12661485373973846,
526
+ "learning_rate": 8.395091111063118e-06,
527
+ "loss": 0.7255,
528
+ "step": 740
529
+ },
530
+ {
531
+ "epoch": 0.28876696505919724,
532
+ "grad_norm": 0.12541916966438293,
533
+ "learning_rate": 8.349049970236822e-06,
534
+ "loss": 0.7274,
535
+ "step": 750
536
+ },
537
+ {
538
+ "epoch": 0.29261719125998653,
539
+ "grad_norm": 0.13116496801376343,
540
+ "learning_rate": 8.302487923351842e-06,
541
+ "loss": 0.7172,
542
+ "step": 760
543
+ },
544
+ {
545
+ "epoch": 0.2964674174607758,
546
+ "grad_norm": 0.13135001063346863,
547
+ "learning_rate": 8.255412212597021e-06,
548
+ "loss": 0.7372,
549
+ "step": 770
550
+ },
551
+ {
552
+ "epoch": 0.3003176436615651,
553
+ "grad_norm": 0.13798557221889496,
554
+ "learning_rate": 8.207830160055688e-06,
555
+ "loss": 0.7338,
556
+ "step": 780
557
+ },
558
+ {
559
+ "epoch": 0.3041678698623544,
560
+ "grad_norm": 0.14424997568130493,
561
+ "learning_rate": 8.159749166566786e-06,
562
+ "loss": 0.7171,
563
+ "step": 790
564
+ },
565
+ {
566
+ "epoch": 0.30801809606314373,
567
+ "grad_norm": 0.1423322856426239,
568
+ "learning_rate": 8.11117671057377e-06,
569
+ "loss": 0.741,
570
+ "step": 800
571
+ },
572
+ {
573
+ "epoch": 0.311868322263933,
574
+ "grad_norm": 0.13012933731079102,
575
+ "learning_rate": 8.062120346961404e-06,
576
+ "loss": 0.7268,
577
+ "step": 810
578
+ },
579
+ {
580
+ "epoch": 0.3157185484647223,
581
+ "grad_norm": 0.12097672373056412,
582
+ "learning_rate": 8.012587705880705e-06,
583
+ "loss": 0.7321,
584
+ "step": 820
585
+ },
586
+ {
587
+ "epoch": 0.3195687746655116,
588
+ "grad_norm": 0.13509368896484375,
589
+ "learning_rate": 7.962586491562146e-06,
590
+ "loss": 0.7209,
591
+ "step": 830
592
+ },
593
+ {
594
+ "epoch": 0.3234190008663009,
595
+ "grad_norm": 0.1372847855091095,
596
+ "learning_rate": 7.912124481117357e-06,
597
+ "loss": 0.7233,
598
+ "step": 840
599
+ },
600
+ {
601
+ "epoch": 0.3272692270670902,
602
+ "grad_norm": 0.1235290989279747,
603
+ "learning_rate": 7.861209523329494e-06,
604
+ "loss": 0.7301,
605
+ "step": 850
606
+ },
607
+ {
608
+ "epoch": 0.3311194532678795,
609
+ "grad_norm": 0.12828940153121948,
610
+ "learning_rate": 7.809849537432432e-06,
611
+ "loss": 0.718,
612
+ "step": 860
613
+ },
614
+ {
615
+ "epoch": 0.3349696794686688,
616
+ "grad_norm": 0.11960785835981369,
617
+ "learning_rate": 7.758052511879035e-06,
618
+ "loss": 0.716,
619
+ "step": 870
620
+ },
621
+ {
622
+ "epoch": 0.3388199056694581,
623
+ "grad_norm": 0.11829168349504471,
624
+ "learning_rate": 7.705826503098642e-06,
625
+ "loss": 0.7226,
626
+ "step": 880
627
+ },
628
+ {
629
+ "epoch": 0.34267013187024736,
630
+ "grad_norm": 0.1296880841255188,
631
+ "learning_rate": 7.653179634243979e-06,
632
+ "loss": 0.7189,
633
+ "step": 890
634
+ },
635
+ {
636
+ "epoch": 0.34652035807103665,
637
+ "grad_norm": 0.12574349343776703,
638
+ "learning_rate": 7.600120093927696e-06,
639
+ "loss": 0.7122,
640
+ "step": 900
641
+ },
642
+ {
643
+ "epoch": 0.350370584271826,
644
+ "grad_norm": 0.13373567163944244,
645
+ "learning_rate": 7.546656134948729e-06,
646
+ "loss": 0.7115,
647
+ "step": 910
648
+ },
649
+ {
650
+ "epoch": 0.3542208104726153,
651
+ "grad_norm": 0.12770110368728638,
652
+ "learning_rate": 7.492796073008661e-06,
653
+ "loss": 0.7083,
654
+ "step": 920
655
+ },
656
+ {
657
+ "epoch": 0.35807103667340456,
658
+ "grad_norm": 0.13059262931346893,
659
+ "learning_rate": 7.438548285418326e-06,
660
+ "loss": 0.7001,
661
+ "step": 930
662
+ },
663
+ {
664
+ "epoch": 0.36192126287419385,
665
+ "grad_norm": 0.14298678934574127,
666
+ "learning_rate": 7.3839212097948e-06,
667
+ "loss": 0.739,
668
+ "step": 940
669
+ },
670
+ {
671
+ "epoch": 0.36577148907498314,
672
+ "grad_norm": 0.11955954879522324,
673
+ "learning_rate": 7.328923342749039e-06,
674
+ "loss": 0.7514,
675
+ "step": 950
676
+ },
677
+ {
678
+ "epoch": 0.3696217152757725,
679
+ "grad_norm": 0.135214164853096,
680
+ "learning_rate": 7.273563238564319e-06,
681
+ "loss": 0.7274,
682
+ "step": 960
683
+ },
684
+ {
685
+ "epoch": 0.37347194147656176,
686
+ "grad_norm": 0.1253443956375122,
687
+ "learning_rate": 7.217849507865724e-06,
688
+ "loss": 0.7034,
689
+ "step": 970
690
+ },
691
+ {
692
+ "epoch": 0.37732216767735105,
693
+ "grad_norm": 0.12406131625175476,
694
+ "learning_rate": 7.161790816280858e-06,
695
+ "loss": 0.7215,
696
+ "step": 980
697
+ },
698
+ {
699
+ "epoch": 0.38117239387814034,
700
+ "grad_norm": 0.1286347359418869,
701
+ "learning_rate": 7.1053958830920125e-06,
702
+ "loss": 0.6924,
703
+ "step": 990
704
+ },
705
+ {
706
+ "epoch": 0.3850226200789296,
707
+ "grad_norm": 0.12784385681152344,
708
+ "learning_rate": 7.0486734798799715e-06,
709
+ "loss": 0.7273,
710
+ "step": 1000
711
+ },
712
+ {
713
+ "epoch": 0.3888728462797189,
714
+ "grad_norm": 0.13386671245098114,
715
+ "learning_rate": 6.991632429159708e-06,
716
+ "loss": 0.7493,
717
+ "step": 1010
718
+ },
719
+ {
720
+ "epoch": 0.39272307248050825,
721
+ "grad_norm": 0.14015640318393707,
722
+ "learning_rate": 6.934281603008132e-06,
723
+ "loss": 0.7202,
724
+ "step": 1020
725
+ },
726
+ {
727
+ "epoch": 0.39657329868129754,
728
+ "grad_norm": 0.13139744102954865,
729
+ "learning_rate": 6.876629921684143e-06,
730
+ "loss": 0.7332,
731
+ "step": 1030
732
+ },
733
+ {
734
+ "epoch": 0.4004235248820868,
735
+ "grad_norm": 0.12731000781059265,
736
+ "learning_rate": 6.818686352241199e-06,
737
+ "loss": 0.7283,
738
+ "step": 1040
739
+ },
740
+ {
741
+ "epoch": 0.4042737510828761,
742
+ "grad_norm": 0.11948556452989578,
743
+ "learning_rate": 6.760459907132573e-06,
744
+ "loss": 0.7173,
745
+ "step": 1050
746
+ },
747
+ {
748
+ "epoch": 0.4081239772836654,
749
+ "grad_norm": 0.13406294584274292,
750
+ "learning_rate": 6.701959642809589e-06,
751
+ "loss": 0.7202,
752
+ "step": 1060
753
+ },
754
+ {
755
+ "epoch": 0.41197420348445474,
756
+ "grad_norm": 0.13165104389190674,
757
+ "learning_rate": 6.643194658312987e-06,
758
+ "loss": 0.705,
759
+ "step": 1070
760
+ },
761
+ {
762
+ "epoch": 0.415824429685244,
763
+ "grad_norm": 0.11999239027500153,
764
+ "learning_rate": 6.584174093857676e-06,
765
+ "loss": 0.7417,
766
+ "step": 1080
767
+ },
768
+ {
769
+ "epoch": 0.4196746558860333,
770
+ "grad_norm": 0.1430213749408722,
771
+ "learning_rate": 6.5249071294110735e-06,
772
+ "loss": 0.7132,
773
+ "step": 1090
774
+ },
775
+ {
776
+ "epoch": 0.4235248820868226,
777
+ "grad_norm": 0.14033856987953186,
778
+ "learning_rate": 6.465402983265274e-06,
779
+ "loss": 0.7319,
780
+ "step": 1100
781
+ },
782
+ {
783
+ "epoch": 0.4273751082876119,
784
+ "grad_norm": 0.1317427009344101,
785
+ "learning_rate": 6.405670910603243e-06,
786
+ "loss": 0.7282,
787
+ "step": 1110
788
+ },
789
+ {
790
+ "epoch": 0.43122533448840117,
791
+ "grad_norm": 0.13468050956726074,
792
+ "learning_rate": 6.345720202059291e-06,
793
+ "loss": 0.7199,
794
+ "step": 1120
795
+ },
796
+ {
797
+ "epoch": 0.4350755606891905,
798
+ "grad_norm": 0.12733791768550873,
799
+ "learning_rate": 6.285560182274008e-06,
800
+ "loss": 0.7096,
801
+ "step": 1130
802
+ },
803
+ {
804
+ "epoch": 0.4389257868899798,
805
+ "grad_norm": 0.1251576840877533,
806
+ "learning_rate": 6.225200208443936e-06,
807
+ "loss": 0.7169,
808
+ "step": 1140
809
+ },
810
+ {
811
+ "epoch": 0.4427760130907691,
812
+ "grad_norm": 0.1204310730099678,
813
+ "learning_rate": 6.16464966886616e-06,
814
+ "loss": 0.7093,
815
+ "step": 1150
816
+ },
817
+ {
818
+ "epoch": 0.44662623929155837,
819
+ "grad_norm": 0.15269270539283752,
820
+ "learning_rate": 6.103917981478065e-06,
821
+ "loss": 0.7406,
822
+ "step": 1160
823
+ },
824
+ {
825
+ "epoch": 0.45047646549234766,
826
+ "grad_norm": 0.12224593758583069,
827
+ "learning_rate": 6.043014592392489e-06,
828
+ "loss": 0.7449,
829
+ "step": 1170
830
+ },
831
+ {
832
+ "epoch": 0.454326691693137,
833
+ "grad_norm": 0.1231018453836441,
834
+ "learning_rate": 5.981948974428485e-06,
835
+ "loss": 0.7195,
836
+ "step": 1180
837
+ },
838
+ {
839
+ "epoch": 0.4581769178939263,
840
+ "grad_norm": 0.13324131071567535,
841
+ "learning_rate": 5.920730625637934e-06,
842
+ "loss": 0.6971,
843
+ "step": 1190
844
+ },
845
+ {
846
+ "epoch": 0.46202714409471557,
847
+ "grad_norm": 0.1418658196926117,
848
+ "learning_rate": 5.8593690678282355e-06,
849
+ "loss": 0.7102,
850
+ "step": 1200
851
+ },
852
+ {
853
+ "epoch": 0.46587737029550486,
854
+ "grad_norm": 0.13369235396385193,
855
+ "learning_rate": 5.797873845081301e-06,
856
+ "loss": 0.7129,
857
+ "step": 1210
858
+ },
859
+ {
860
+ "epoch": 0.46972759649629414,
861
+ "grad_norm": 0.12898017466068268,
862
+ "learning_rate": 5.736254522269075e-06,
863
+ "loss": 0.7272,
864
+ "step": 1220
865
+ },
866
+ {
867
+ "epoch": 0.47357782269708343,
868
+ "grad_norm": 0.12865957617759705,
869
+ "learning_rate": 5.6745206835658386e-06,
870
+ "loss": 0.7324,
871
+ "step": 1230
872
+ },
873
+ {
874
+ "epoch": 0.47742804889787277,
875
+ "grad_norm": 0.12598051130771637,
876
+ "learning_rate": 5.612681930957495e-06,
877
+ "loss": 0.7102,
878
+ "step": 1240
879
+ },
880
+ {
881
+ "epoch": 0.48127827509866206,
882
+ "grad_norm": 0.11844313144683838,
883
+ "learning_rate": 5.550747882748095e-06,
884
+ "loss": 0.7025,
885
+ "step": 1250
886
+ },
887
+ {
888
+ "epoch": 0.48512850129945134,
889
+ "grad_norm": 0.13543567061424255,
890
+ "learning_rate": 5.4887281720638195e-06,
891
+ "loss": 0.7092,
892
+ "step": 1260
893
+ },
894
+ {
895
+ "epoch": 0.48897872750024063,
896
+ "grad_norm": 0.13967390358448029,
897
+ "learning_rate": 5.42663244535466e-06,
898
+ "loss": 0.697,
899
+ "step": 1270
900
+ },
901
+ {
902
+ "epoch": 0.4928289537010299,
903
+ "grad_norm": 0.13191789388656616,
904
+ "learning_rate": 5.3644703608940205e-06,
905
+ "loss": 0.7024,
906
+ "step": 1280
907
+ },
908
+ {
909
+ "epoch": 0.49667917990181926,
910
+ "grad_norm": 0.12759451568126678,
911
+ "learning_rate": 5.302251587276491e-06,
912
+ "loss": 0.7245,
913
+ "step": 1290
914
+ },
915
+ {
916
+ "epoch": 0.5005294061026085,
917
+ "grad_norm": 0.1240101158618927,
918
+ "learning_rate": 5.2399858019140005e-06,
919
+ "loss": 0.717,
920
+ "step": 1300
921
+ },
922
+ {
923
+ "epoch": 0.5043796323033978,
924
+ "grad_norm": 0.13553784787654877,
925
+ "learning_rate": 5.177682689530615e-06,
926
+ "loss": 0.7353,
927
+ "step": 1310
928
+ },
929
+ {
930
+ "epoch": 0.5082298585041871,
931
+ "grad_norm": 0.13568830490112305,
932
+ "learning_rate": 5.11535194065619e-06,
933
+ "loss": 0.7035,
934
+ "step": 1320
935
+ },
936
+ {
937
+ "epoch": 0.5120800847049765,
938
+ "grad_norm": 0.14229336380958557,
939
+ "learning_rate": 5.053003250119113e-06,
940
+ "loss": 0.7051,
941
+ "step": 1330
942
+ },
943
+ {
944
+ "epoch": 0.5159303109057657,
945
+ "grad_norm": 0.13243842124938965,
946
+ "learning_rate": 4.9906463155383925e-06,
947
+ "loss": 0.7179,
948
+ "step": 1340
949
+ },
950
+ {
951
+ "epoch": 0.519780537106555,
952
+ "grad_norm": 0.12789613008499146,
953
+ "learning_rate": 4.9282908358153046e-06,
954
+ "loss": 0.7121,
955
+ "step": 1350
956
+ },
957
+ {
958
+ "epoch": 0.5236307633073443,
959
+ "grad_norm": 0.13916464149951935,
960
+ "learning_rate": 4.8659465096248345e-06,
961
+ "loss": 0.7163,
962
+ "step": 1360
963
+ },
964
+ {
965
+ "epoch": 0.5274809895081336,
966
+ "grad_norm": 0.12604093551635742,
967
+ "learning_rate": 4.803623033907169e-06,
968
+ "loss": 0.7206,
969
+ "step": 1370
970
+ },
971
+ {
972
+ "epoch": 0.5313312157089229,
973
+ "grad_norm": 0.1178334429860115,
974
+ "learning_rate": 4.741330102359441e-06,
975
+ "loss": 0.7303,
976
+ "step": 1380
977
+ },
978
+ {
979
+ "epoch": 0.5351814419097122,
980
+ "grad_norm": 0.13467755913734436,
981
+ "learning_rate": 4.679077403927996e-06,
982
+ "loss": 0.703,
983
+ "step": 1390
984
+ },
985
+ {
986
+ "epoch": 0.5390316681105015,
987
+ "grad_norm": 0.1286448985338211,
988
+ "learning_rate": 4.616874621301373e-06,
989
+ "loss": 0.7112,
990
+ "step": 1400
991
+ },
992
+ {
993
+ "epoch": 0.5428818943112907,
994
+ "grad_norm": 0.13142161071300507,
995
+ "learning_rate": 4.554731429404293e-06,
996
+ "loss": 0.7273,
997
+ "step": 1410
998
+ },
999
+ {
1000
+ "epoch": 0.5467321205120801,
1001
+ "grad_norm": 0.1343977153301239,
1002
+ "learning_rate": 4.492657493892822e-06,
1003
+ "loss": 0.7074,
1004
+ "step": 1420
1005
+ },
1006
+ {
1007
+ "epoch": 0.5505823467128694,
1008
+ "grad_norm": 0.13874071836471558,
1009
+ "learning_rate": 4.430662469650999e-06,
1010
+ "loss": 0.7202,
1011
+ "step": 1430
1012
+ },
1013
+ {
1014
+ "epoch": 0.5544325729136587,
1015
+ "grad_norm": 0.14196699857711792,
1016
+ "learning_rate": 4.368755999289118e-06,
1017
+ "loss": 0.7261,
1018
+ "step": 1440
1019
+ },
1020
+ {
1021
+ "epoch": 0.558282799114448,
1022
+ "grad_norm": 0.14229939877986908,
1023
+ "learning_rate": 4.306947711643949e-06,
1024
+ "loss": 0.707,
1025
+ "step": 1450
1026
+ },
1027
+ {
1028
+ "epoch": 0.5621330253152372,
1029
+ "grad_norm": 0.12409143149852753,
1030
+ "learning_rate": 4.245247220281068e-06,
1031
+ "loss": 0.7132,
1032
+ "step": 1460
1033
+ },
1034
+ {
1035
+ "epoch": 0.5659832515160266,
1036
+ "grad_norm": 0.12602756917476654,
1037
+ "learning_rate": 4.183664121999587e-06,
1038
+ "loss": 0.7067,
1039
+ "step": 1470
1040
+ },
1041
+ {
1042
+ "epoch": 0.5698334777168159,
1043
+ "grad_norm": 0.13078296184539795,
1044
+ "learning_rate": 4.122207995339474e-06,
1045
+ "loss": 0.7055,
1046
+ "step": 1480
1047
+ },
1048
+ {
1049
+ "epoch": 0.5736837039176051,
1050
+ "grad_norm": 0.13216908276081085,
1051
+ "learning_rate": 4.060888399091734e-06,
1052
+ "loss": 0.7126,
1053
+ "step": 1490
1054
+ },
1055
+ {
1056
+ "epoch": 0.5775339301183945,
1057
+ "grad_norm": 0.12879057228565216,
1058
+ "learning_rate": 3.999714870811639e-06,
1059
+ "loss": 0.7268,
1060
+ "step": 1500
1061
+ },
1062
+ {
1063
+ "epoch": 0.5813841563191837,
1064
+ "grad_norm": 0.12915633618831635,
1065
+ "learning_rate": 3.938696925335275e-06,
1066
+ "loss": 0.7145,
1067
+ "step": 1510
1068
+ },
1069
+ {
1070
+ "epoch": 0.5852343825199731,
1071
+ "grad_norm": 0.13643813133239746,
1072
+ "learning_rate": 3.87784405329962e-06,
1073
+ "loss": 0.7195,
1074
+ "step": 1520
1075
+ },
1076
+ {
1077
+ "epoch": 0.5890846087207623,
1078
+ "grad_norm": 0.12933865189552307,
1079
+ "learning_rate": 3.817165719666382e-06,
1080
+ "loss": 0.7172,
1081
+ "step": 1530
1082
+ },
1083
+ {
1084
+ "epoch": 0.5929348349215516,
1085
+ "grad_norm": 0.12968064844608307,
1086
+ "learning_rate": 3.756671362249841e-06,
1087
+ "loss": 0.7303,
1088
+ "step": 1540
1089
+ },
1090
+ {
1091
+ "epoch": 0.596785061122341,
1092
+ "grad_norm": 0.1277313381433487,
1093
+ "learning_rate": 3.6963703902488907e-06,
1094
+ "loss": 0.7197,
1095
+ "step": 1550
1096
+ },
1097
+ {
1098
+ "epoch": 0.6006352873231302,
1099
+ "grad_norm": 0.13256987929344177,
1100
+ "learning_rate": 3.6362721827835655e-06,
1101
+ "loss": 0.7267,
1102
+ "step": 1560
1103
+ },
1104
+ {
1105
+ "epoch": 0.6044855135239195,
1106
+ "grad_norm": 0.13368843495845795,
1107
+ "learning_rate": 3.576386087436212e-06,
1108
+ "loss": 0.7187,
1109
+ "step": 1570
1110
+ },
1111
+ {
1112
+ "epoch": 0.6083357397247088,
1113
+ "grad_norm": 0.12877191603183746,
1114
+ "learning_rate": 3.5167214187975913e-06,
1115
+ "loss": 0.7225,
1116
+ "step": 1580
1117
+ },
1118
+ {
1119
+ "epoch": 0.6121859659254981,
1120
+ "grad_norm": 0.11978751420974731,
1121
+ "learning_rate": 3.457287457018086e-06,
1122
+ "loss": 0.7183,
1123
+ "step": 1590
1124
+ },
1125
+ {
1126
+ "epoch": 0.6160361921262875,
1127
+ "grad_norm": 0.12011537700891495,
1128
+ "learning_rate": 3.3980934463643003e-06,
1129
+ "loss": 0.7054,
1130
+ "step": 1600
1131
+ },
1132
+ {
1133
+ "epoch": 0.6198864183270767,
1134
+ "grad_norm": 0.12017708271741867,
1135
+ "learning_rate": 3.3391485937812053e-06,
1136
+ "loss": 0.684,
1137
+ "step": 1610
1138
+ },
1139
+ {
1140
+ "epoch": 0.623736644527866,
1141
+ "grad_norm": 0.1265425980091095,
1142
+ "learning_rate": 3.280462067460115e-06,
1143
+ "loss": 0.6878,
1144
+ "step": 1620
1145
+ },
1146
+ {
1147
+ "epoch": 0.6275868707286553,
1148
+ "grad_norm": 0.1251218467950821,
1149
+ "learning_rate": 3.222042995412669e-06,
1150
+ "loss": 0.726,
1151
+ "step": 1630
1152
+ },
1153
+ {
1154
+ "epoch": 0.6314370969294446,
1155
+ "grad_norm": 0.11994276195764542,
1156
+ "learning_rate": 3.163900464051089e-06,
1157
+ "loss": 0.7057,
1158
+ "step": 1640
1159
+ },
1160
+ {
1161
+ "epoch": 0.635287323130234,
1162
+ "grad_norm": 0.13004422187805176,
1163
+ "learning_rate": 3.1060435167748826e-06,
1164
+ "loss": 0.718,
1165
+ "step": 1650
1166
+ },
1167
+ {
1168
+ "epoch": 0.6391375493310232,
1169
+ "grad_norm": 0.12432219088077545,
1170
+ "learning_rate": 3.048481152564253e-06,
1171
+ "loss": 0.7046,
1172
+ "step": 1660
1173
+ },
1174
+ {
1175
+ "epoch": 0.6429877755318125,
1176
+ "grad_norm": 0.13419856131076813,
1177
+ "learning_rate": 2.9912223245804017e-06,
1178
+ "loss": 0.7298,
1179
+ "step": 1670
1180
+ },
1181
+ {
1182
+ "epoch": 0.6468380017326018,
1183
+ "grad_norm": 0.12393974512815475,
1184
+ "learning_rate": 2.934275938772981e-06,
1185
+ "loss": 0.7075,
1186
+ "step": 1680
1187
+ },
1188
+ {
1189
+ "epoch": 0.6506882279333911,
1190
+ "grad_norm": 0.1330839842557907,
1191
+ "learning_rate": 2.8776508524948676e-06,
1192
+ "loss": 0.6974,
1193
+ "step": 1690
1194
+ },
1195
+ {
1196
+ "epoch": 0.6545384541341804,
1197
+ "grad_norm": 0.11942951381206512,
1198
+ "learning_rate": 2.821355873124503e-06,
1199
+ "loss": 0.6958,
1200
+ "step": 1700
1201
+ },
1202
+ {
1203
+ "epoch": 0.6583886803349697,
1204
+ "grad_norm": 0.12929436564445496,
1205
+ "learning_rate": 2.7653997566960144e-06,
1206
+ "loss": 0.7182,
1207
+ "step": 1710
1208
+ },
1209
+ {
1210
+ "epoch": 0.662238906535759,
1211
+ "grad_norm": 0.12925295531749725,
1212
+ "learning_rate": 2.709791206537309e-06,
1213
+ "loss": 0.7076,
1214
+ "step": 1720
1215
+ },
1216
+ {
1217
+ "epoch": 0.6660891327365482,
1218
+ "grad_norm": 0.1261921525001526,
1219
+ "learning_rate": 2.6545388719163747e-06,
1220
+ "loss": 0.7045,
1221
+ "step": 1730
1222
+ },
1223
+ {
1224
+ "epoch": 0.6699393589373376,
1225
+ "grad_norm": 0.1261424422264099,
1226
+ "learning_rate": 2.599651346695979e-06,
1227
+ "loss": 0.7161,
1228
+ "step": 1740
1229
+ },
1230
+ {
1231
+ "epoch": 0.6737895851381268,
1232
+ "grad_norm": 0.13251735270023346,
1233
+ "learning_rate": 2.545137167997007e-06,
1234
+ "loss": 0.7113,
1235
+ "step": 1750
1236
+ },
1237
+ {
1238
+ "epoch": 0.6776398113389162,
1239
+ "grad_norm": 0.12073896080255508,
1240
+ "learning_rate": 2.4910048148705972e-06,
1241
+ "loss": 0.6941,
1242
+ "step": 1760
1243
+ },
1244
+ {
1245
+ "epoch": 0.6814900375397055,
1246
+ "grad_norm": 0.1349298357963562,
1247
+ "learning_rate": 2.4372627069793327e-06,
1248
+ "loss": 0.7379,
1249
+ "step": 1770
1250
+ },
1251
+ {
1252
+ "epoch": 0.6853402637404947,
1253
+ "grad_norm": 0.1322462558746338,
1254
+ "learning_rate": 2.3839192032876486e-06,
1255
+ "loss": 0.7221,
1256
+ "step": 1780
1257
+ },
1258
+ {
1259
+ "epoch": 0.6891904899412841,
1260
+ "grad_norm": 0.12216746062040329,
1261
+ "learning_rate": 2.3309826007617047e-06,
1262
+ "loss": 0.7334,
1263
+ "step": 1790
1264
+ },
1265
+ {
1266
+ "epoch": 0.6930407161420733,
1267
+ "grad_norm": 0.12539677321910858,
1268
+ "learning_rate": 2.278461133078881e-06,
1269
+ "loss": 0.6975,
1270
+ "step": 1800
1271
+ },
1272
+ {
1273
+ "epoch": 0.6968909423428626,
1274
+ "grad_norm": 0.12262309342622757,
1275
+ "learning_rate": 2.2263629693471258e-06,
1276
+ "loss": 0.6919,
1277
+ "step": 1810
1278
+ },
1279
+ {
1280
+ "epoch": 0.700741168543652,
1281
+ "grad_norm": 0.12308688461780548,
1282
+ "learning_rate": 2.1746962128343414e-06,
1283
+ "loss": 0.7153,
1284
+ "step": 1820
1285
+ },
1286
+ {
1287
+ "epoch": 0.7045913947444412,
1288
+ "grad_norm": 0.12386128306388855,
1289
+ "learning_rate": 2.1234688997080183e-06,
1290
+ "loss": 0.7137,
1291
+ "step": 1830
1292
+ },
1293
+ {
1294
+ "epoch": 0.7084416209452306,
1295
+ "grad_norm": 0.13593868911266327,
1296
+ "learning_rate": 2.0726889977853053e-06,
1297
+ "loss": 0.7099,
1298
+ "step": 1840
1299
+ },
1300
+ {
1301
+ "epoch": 0.7122918471460198,
1302
+ "grad_norm": 0.13037900626659393,
1303
+ "learning_rate": 2.022364405293703e-06,
1304
+ "loss": 0.7237,
1305
+ "step": 1850
1306
+ },
1307
+ {
1308
+ "epoch": 0.7161420733468091,
1309
+ "grad_norm": 0.12843185663223267,
1310
+ "learning_rate": 1.9725029496425814e-06,
1311
+ "loss": 0.7179,
1312
+ "step": 1860
1313
+ },
1314
+ {
1315
+ "epoch": 0.7199922995475985,
1316
+ "grad_norm": 0.13039237260818481,
1317
+ "learning_rate": 1.923112386205727e-06,
1318
+ "loss": 0.7139,
1319
+ "step": 1870
1320
+ },
1321
+ {
1322
+ "epoch": 0.7238425257483877,
1323
+ "grad_norm": 0.13133153319358826,
1324
+ "learning_rate": 1.874200397115079e-06,
1325
+ "loss": 0.7203,
1326
+ "step": 1880
1327
+ },
1328
+ {
1329
+ "epoch": 0.727692751949177,
1330
+ "grad_norm": 0.1334591507911682,
1331
+ "learning_rate": 1.825774590065857e-06,
1332
+ "loss": 0.7057,
1333
+ "step": 1890
1334
+ },
1335
+ {
1336
+ "epoch": 0.7315429781499663,
1337
+ "grad_norm": 0.1435183435678482,
1338
+ "learning_rate": 1.7778424971332813e-06,
1339
+ "loss": 0.7133,
1340
+ "step": 1900
1341
+ },
1342
+ {
1343
+ "epoch": 0.7353932043507556,
1344
+ "grad_norm": 0.12128371000289917,
1345
+ "learning_rate": 1.7304115736010407e-06,
1346
+ "loss": 0.7142,
1347
+ "step": 1910
1348
+ },
1349
+ {
1350
+ "epoch": 0.739243430551545,
1351
+ "grad_norm": 0.12704399228096008,
1352
+ "learning_rate": 1.6834891968017108e-06,
1353
+ "loss": 0.7064,
1354
+ "step": 1920
1355
+ },
1356
+ {
1357
+ "epoch": 0.7430936567523342,
1358
+ "grad_norm": 0.1307741105556488,
1359
+ "learning_rate": 1.6370826649692884e-06,
1360
+ "loss": 0.7163,
1361
+ "step": 1930
1362
+ },
1363
+ {
1364
+ "epoch": 0.7469438829531235,
1365
+ "grad_norm": 0.13048064708709717,
1366
+ "learning_rate": 1.5911991961040463e-06,
1367
+ "loss": 0.731,
1368
+ "step": 1940
1369
+ },
1370
+ {
1371
+ "epoch": 0.7507941091539128,
1372
+ "grad_norm": 0.12764278054237366,
1373
+ "learning_rate": 1.545845926849851e-06,
1374
+ "loss": 0.7176,
1375
+ "step": 1950
1376
+ },
1377
+ {
1378
+ "epoch": 0.7546443353547021,
1379
+ "grad_norm": 0.13470706343650818,
1380
+ "learning_rate": 1.5010299113841397e-06,
1381
+ "loss": 0.7329,
1382
+ "step": 1960
1383
+ },
1384
+ {
1385
+ "epoch": 0.7584945615554913,
1386
+ "grad_norm": 0.1258677840232849,
1387
+ "learning_rate": 1.4567581203207238e-06,
1388
+ "loss": 0.7298,
1389
+ "step": 1970
1390
+ },
1391
+ {
1392
+ "epoch": 0.7623447877562807,
1393
+ "grad_norm": 0.12920954823493958,
1394
+ "learning_rate": 1.413037439625593e-06,
1395
+ "loss": 0.7116,
1396
+ "step": 1980
1397
+ },
1398
+ {
1399
+ "epoch": 0.76619501395707,
1400
+ "grad_norm": 0.1395183503627777,
1401
+ "learning_rate": 1.3698746695458847e-06,
1402
+ "loss": 0.7146,
1403
+ "step": 1990
1404
+ },
1405
+ {
1406
+ "epoch": 0.7700452401578592,
1407
+ "grad_norm": 0.12889355421066284,
1408
+ "learning_rate": 1.3272765235521745e-06,
1409
+ "loss": 0.7156,
1410
+ "step": 2000
1411
+ },
1412
+ {
1413
+ "epoch": 0.7738954663586486,
1414
+ "grad_norm": 0.12393711507320404,
1415
+ "learning_rate": 1.2852496272942877e-06,
1416
+ "loss": 0.715,
1417
+ "step": 2010
1418
+ },
1419
+ {
1420
+ "epoch": 0.7777456925594378,
1421
+ "grad_norm": 0.13096049427986145,
1422
+ "learning_rate": 1.2438005175707396e-06,
1423
+ "loss": 0.699,
1424
+ "step": 2020
1425
+ },
1426
+ {
1427
+ "epoch": 0.7815959187602272,
1428
+ "grad_norm": 0.13400831818580627,
1429
+ "learning_rate": 1.2029356413120297e-06,
1430
+ "loss": 0.6853,
1431
+ "step": 2030
1432
+ },
1433
+ {
1434
+ "epoch": 0.7854461449610165,
1435
+ "grad_norm": 0.13143937289714813,
1436
+ "learning_rate": 1.1626613545778786e-06,
1437
+ "loss": 0.7033,
1438
+ "step": 2040
1439
+ },
1440
+ {
1441
+ "epoch": 0.7892963711618057,
1442
+ "grad_norm": 0.12021194398403168,
1443
+ "learning_rate": 1.1229839215686294e-06,
1444
+ "loss": 0.7121,
1445
+ "step": 2050
1446
+ },
1447
+ {
1448
+ "epoch": 0.7931465973625951,
1449
+ "grad_norm": 0.12233367562294006,
1450
+ "learning_rate": 1.0839095136509188e-06,
1451
+ "loss": 0.6906,
1452
+ "step": 2060
1453
+ },
1454
+ {
1455
+ "epoch": 0.7969968235633843,
1456
+ "grad_norm": 0.13207249343395233,
1457
+ "learning_rate": 1.045444208397791e-06,
1458
+ "loss": 0.7016,
1459
+ "step": 2070
1460
+ },
1461
+ {
1462
+ "epoch": 0.8008470497641736,
1463
+ "grad_norm": 0.13121359050273895,
1464
+ "learning_rate": 1.007593988643402e-06,
1465
+ "loss": 0.6914,
1466
+ "step": 2080
1467
+ },
1468
+ {
1469
+ "epoch": 0.804697275964963,
1470
+ "grad_norm": 0.12072388082742691,
1471
+ "learning_rate": 9.703647415524613e-07,
1472
+ "loss": 0.7147,
1473
+ "step": 2090
1474
+ },
1475
+ {
1476
+ "epoch": 0.8085475021657522,
1477
+ "grad_norm": 0.12888877093791962,
1478
+ "learning_rate": 9.337622577045485e-07,
1479
+ "loss": 0.7081,
1480
+ "step": 2100
1481
+ },
1482
+ {
1483
+ "epoch": 0.8123977283665416,
1484
+ "grad_norm": 0.1268695443868637,
1485
+ "learning_rate": 8.977922301934588e-07,
1486
+ "loss": 0.723,
1487
+ "step": 2110
1488
+ },
1489
+ {
1490
+ "epoch": 0.8162479545673308,
1491
+ "grad_norm": 0.12469799816608429,
1492
+ "learning_rate": 8.624602537417009e-07,
1493
+ "loss": 0.7216,
1494
+ "step": 2120
1495
+ },
1496
+ {
1497
+ "epoch": 0.8200981807681201,
1498
+ "grad_norm": 0.13114531338214874,
1499
+ "learning_rate": 8.277718238303123e-07,
1500
+ "loss": 0.7027,
1501
+ "step": 2130
1502
+ },
1503
+ {
1504
+ "epoch": 0.8239484069689095,
1505
+ "grad_norm": 0.12440069019794464,
1506
+ "learning_rate": 7.937323358440935e-07,
1507
+ "loss": 0.7153,
1508
+ "step": 2140
1509
+ },
1510
+ {
1511
+ "epoch": 0.8277986331696987,
1512
+ "grad_norm": 0.1400829255580902,
1513
+ "learning_rate": 7.603470842324223e-07,
1514
+ "loss": 0.6987,
1515
+ "step": 2150
1516
+ },
1517
+ {
1518
+ "epoch": 0.831648859370488,
1519
+ "grad_norm": 0.12041378766298294,
1520
+ "learning_rate": 7.276212616857575e-07,
1521
+ "loss": 0.7156,
1522
+ "step": 2160
1523
+ },
1524
+ {
1525
+ "epoch": 0.8354990855712773,
1526
+ "grad_norm": 0.1244392842054367,
1527
+ "learning_rate": 6.955599583279837e-07,
1528
+ "loss": 0.7172,
1529
+ "step": 2170
1530
+ },
1531
+ {
1532
+ "epoch": 0.8393493117720666,
1533
+ "grad_norm": 0.12846964597702026,
1534
+ "learning_rate": 6.641681609246981e-07,
1535
+ "loss": 0.7093,
1536
+ "step": 2180
1537
+ },
1538
+ {
1539
+ "epoch": 0.8431995379728559,
1540
+ "grad_norm": 0.12609384953975677,
1541
+ "learning_rate": 6.334507521075711e-07,
1542
+ "loss": 0.7162,
1543
+ "step": 2190
1544
+ },
1545
+ {
1546
+ "epoch": 0.8470497641736452,
1547
+ "grad_norm": 0.13362562656402588,
1548
+ "learning_rate": 6.034125096149179e-07,
1549
+ "loss": 0.7043,
1550
+ "step": 2200
1551
+ },
1552
+ {
1553
+ "epoch": 0.8508999903744345,
1554
+ "grad_norm": 0.14555233716964722,
1555
+ "learning_rate": 5.740581055485722e-07,
1556
+ "loss": 0.692,
1557
+ "step": 2210
1558
+ },
1559
+ {
1560
+ "epoch": 0.8547502165752238,
1561
+ "grad_norm": 0.11968927830457687,
1562
+ "learning_rate": 5.453921056471961e-07,
1563
+ "loss": 0.7134,
1564
+ "step": 2220
1565
+ },
1566
+ {
1567
+ "epoch": 0.8586004427760131,
1568
+ "grad_norm": 0.13524433970451355,
1569
+ "learning_rate": 5.174189685761288e-07,
1570
+ "loss": 0.7057,
1571
+ "step": 2230
1572
+ },
1573
+ {
1574
+ "epoch": 0.8624506689768023,
1575
+ "grad_norm": 0.12289699912071228,
1576
+ "learning_rate": 4.901430452338979e-07,
1577
+ "loss": 0.7261,
1578
+ "step": 2240
1579
+ },
1580
+ {
1581
+ "epoch": 0.8663008951775917,
1582
+ "grad_norm": 0.13232819736003876,
1583
+ "learning_rate": 4.6356857807548286e-07,
1584
+ "loss": 0.7053,
1585
+ "step": 2250
1586
+ },
1587
+ {
1588
+ "epoch": 0.870151121378381,
1589
+ "grad_norm": 0.1268777996301651,
1590
+ "learning_rate": 4.376997004524541e-07,
1591
+ "loss": 0.6935,
1592
+ "step": 2260
1593
+ },
1594
+ {
1595
+ "epoch": 0.8740013475791703,
1596
+ "grad_norm": 0.12253368645906448,
1597
+ "learning_rate": 4.12540435970073e-07,
1598
+ "loss": 0.694,
1599
+ "step": 2270
1600
+ },
1601
+ {
1602
+ "epoch": 0.8778515737799596,
1603
+ "grad_norm": 0.1244310811161995,
1604
+ "learning_rate": 3.880946978614703e-07,
1605
+ "loss": 0.72,
1606
+ "step": 2280
1607
+ },
1608
+ {
1609
+ "epoch": 0.8817017999807488,
1610
+ "grad_norm": 0.12464471906423569,
1611
+ "learning_rate": 3.643662883789878e-07,
1612
+ "loss": 0.7131,
1613
+ "step": 2290
1614
+ },
1615
+ {
1616
+ "epoch": 0.8855520261815382,
1617
+ "grad_norm": 0.12150370329618454,
1618
+ "learning_rate": 3.41358898202781e-07,
1619
+ "loss": 0.6979,
1620
+ "step": 2300
1621
+ },
1622
+ {
1623
+ "epoch": 0.8894022523823275,
1624
+ "grad_norm": 0.12179762870073318,
1625
+ "learning_rate": 3.190761058667741e-07,
1626
+ "loss": 0.7011,
1627
+ "step": 2310
1628
+ },
1629
+ {
1630
+ "epoch": 0.8932524785831167,
1631
+ "grad_norm": 0.12354201078414917,
1632
+ "learning_rate": 2.9752137720206463e-07,
1633
+ "loss": 0.7085,
1634
+ "step": 2320
1635
+ },
1636
+ {
1637
+ "epoch": 0.8971027047839061,
1638
+ "grad_norm": 0.13485844433307648,
1639
+ "learning_rate": 2.7669806479785067e-07,
1640
+ "loss": 0.7204,
1641
+ "step": 2330
1642
+ },
1643
+ {
1644
+ "epoch": 0.9009529309846953,
1645
+ "grad_norm": 0.12673316895961761,
1646
+ "learning_rate": 2.566094074799763e-07,
1647
+ "loss": 0.7101,
1648
+ "step": 2340
1649
+ },
1650
+ {
1651
+ "epoch": 0.9048031571854847,
1652
+ "grad_norm": 0.12431760877370834,
1653
+ "learning_rate": 2.3725852980716557e-07,
1654
+ "loss": 0.698,
1655
+ "step": 2350
1656
+ },
1657
+ {
1658
+ "epoch": 0.908653383386274,
1659
+ "grad_norm": 0.1261068433523178,
1660
+ "learning_rate": 2.1864844158503983e-07,
1661
+ "loss": 0.7098,
1662
+ "step": 2360
1663
+ },
1664
+ {
1665
+ "epoch": 0.9125036095870632,
1666
+ "grad_norm": 0.13956844806671143,
1667
+ "learning_rate": 2.007820373979702e-07,
1668
+ "loss": 0.6979,
1669
+ "step": 2370
1670
+ },
1671
+ {
1672
+ "epoch": 0.9163538357878526,
1673
+ "grad_norm": 0.12262115627527237,
1674
+ "learning_rate": 1.8366209615886532e-07,
1675
+ "loss": 0.7043,
1676
+ "step": 2380
1677
+ },
1678
+ {
1679
+ "epoch": 0.9202040619886418,
1680
+ "grad_norm": 0.1251538097858429,
1681
+ "learning_rate": 1.6729128067693613e-07,
1682
+ "loss": 0.7256,
1683
+ "step": 2390
1684
+ },
1685
+ {
1686
+ "epoch": 0.9240542881894311,
1687
+ "grad_norm": 0.12445379793643951,
1688
+ "learning_rate": 1.5167213724353426e-07,
1689
+ "loss": 0.7141,
1690
+ "step": 2400
1691
+ },
1692
+ {
1693
+ "epoch": 0.9279045143902205,
1694
+ "grad_norm": 0.12638315558433533,
1695
+ "learning_rate": 1.3680709523610058e-07,
1696
+ "loss": 0.7,
1697
+ "step": 2410
1698
+ },
1699
+ {
1700
+ "epoch": 0.9317547405910097,
1701
+ "grad_norm": 0.11964631825685501,
1702
+ "learning_rate": 1.2269846674030418e-07,
1703
+ "loss": 0.7244,
1704
+ "step": 2420
1705
+ },
1706
+ {
1707
+ "epoch": 0.935604966791799,
1708
+ "grad_norm": 0.13032449781894684,
1709
+ "learning_rate": 1.0934844619042595e-07,
1710
+ "loss": 0.7215,
1711
+ "step": 2430
1712
+ },
1713
+ {
1714
+ "epoch": 0.9394551929925883,
1715
+ "grad_norm": 0.1303466260433197,
1716
+ "learning_rate": 9.675911002803961e-08,
1717
+ "loss": 0.7066,
1718
+ "step": 2440
1719
+ },
1720
+ {
1721
+ "epoch": 0.9433054191933776,
1722
+ "grad_norm": 0.12870800495147705,
1723
+ "learning_rate": 8.493241637904381e-08,
1724
+ "loss": 0.7031,
1725
+ "step": 2450
1726
+ },
1727
+ {
1728
+ "epoch": 0.9471556453941669,
1729
+ "grad_norm": 0.14222726225852966,
1730
+ "learning_rate": 7.387020474909689e-08,
1731
+ "loss": 0.7268,
1732
+ "step": 2460
1733
+ },
1734
+ {
1735
+ "epoch": 0.9510058715949562,
1736
+ "grad_norm": 0.12441957741975784,
1737
+ "learning_rate": 6.357419573750856e-08,
1738
+ "loss": 0.7119,
1739
+ "step": 2470
1740
+ },
1741
+ {
1742
+ "epoch": 0.9548560977957455,
1743
+ "grad_norm": 0.12966763973236084,
1744
+ "learning_rate": 5.4045990769613365e-08,
1745
+ "loss": 0.7058,
1746
+ "step": 2480
1747
+ },
1748
+ {
1749
+ "epoch": 0.9587063239965348,
1750
+ "grad_norm": 0.14249569177627563,
1751
+ "learning_rate": 4.5287071847694434e-08,
1752
+ "loss": 0.7041,
1753
+ "step": 2490
1754
+ },
1755
+ {
1756
+ "epoch": 0.9625565501973241,
1757
+ "grad_norm": 0.12608219683170319,
1758
+ "learning_rate": 3.7298801320467817e-08,
1759
+ "loss": 0.7054,
1760
+ "step": 2500
1761
+ },
1762
+ {
1763
+ "epoch": 0.9664067763981133,
1764
+ "grad_norm": 0.12674234807491302,
1765
+ "learning_rate": 3.0082421671192576e-08,
1766
+ "loss": 0.6978,
1767
+ "step": 2510
1768
+ },
1769
+ {
1770
+ "epoch": 0.9702570025989027,
1771
+ "grad_norm": 0.11876851320266724,
1772
+ "learning_rate": 2.3639055324410355e-08,
1773
+ "loss": 0.7178,
1774
+ "step": 2520
1775
+ },
1776
+ {
1777
+ "epoch": 0.974107228799692,
1778
+ "grad_norm": 0.12713654339313507,
1779
+ "learning_rate": 1.7969704471369477e-08,
1780
+ "loss": 0.7012,
1781
+ "step": 2530
1782
+ },
1783
+ {
1784
+ "epoch": 0.9779574550004813,
1785
+ "grad_norm": 0.12086483836174011,
1786
+ "learning_rate": 1.3075250914142413e-08,
1787
+ "loss": 0.7089,
1788
+ "step": 2540
1789
+ },
1790
+ {
1791
+ "epoch": 0.9818076812012706,
1792
+ "grad_norm": 0.12415555119514465,
1793
+ "learning_rate": 8.956455928476049e-09,
1794
+ "loss": 0.7185,
1795
+ "step": 2550
1796
+ },
1797
+ {
1798
+ "epoch": 0.9856579074020598,
1799
+ "grad_norm": 0.11540102958679199,
1800
+ "learning_rate": 5.6139601453786275e-09,
1801
+ "loss": 0.6962,
1802
+ "step": 2560
1803
+ },
1804
+ {
1805
+ "epoch": 0.9895081336028492,
1806
+ "grad_norm": 0.13373278081417084,
1807
+ "learning_rate": 3.0482834514805602e-09,
1808
+ "loss": 0.7014,
1809
+ "step": 2570
1810
+ },
1811
+ {
1812
+ "epoch": 0.9933583598036385,
1813
+ "grad_norm": 0.11378894746303558,
1814
+ "learning_rate": 1.2598249081713366e-09,
1815
+ "loss": 0.6845,
1816
+ "step": 2580
1817
+ },
1818
+ {
1819
+ "epoch": 0.9972085860044277,
1820
+ "grad_norm": 0.13336224853992462,
1821
+ "learning_rate": 2.488626895302826e-10,
1822
+ "loss": 0.7205,
1823
+ "step": 2590
1824
+ },
1825
+ {
1826
+ "epoch": 0.9999037443449803,
1827
+ "step": 2597,
1828
+ "total_flos": 3.675560283088093e+18,
1829
+ "train_loss": 0.7415287249539052,
1830
+ "train_runtime": 14794.8418,
1831
+ "train_samples_per_second": 44.942,
1832
+ "train_steps_per_second": 0.176
1833
+ }
1834
+ ],
1835
+ "logging_steps": 10,
1836
+ "max_steps": 2597,
1837
+ "num_input_tokens_seen": 0,
1838
+ "num_train_epochs": 1,
1839
+ "save_steps": 500,
1840
+ "stateful_callbacks": {
1841
+ "TrainerControl": {
1842
+ "args": {
1843
+ "should_epoch_stop": false,
1844
+ "should_evaluate": false,
1845
+ "should_log": false,
1846
+ "should_save": true,
1847
+ "should_training_stop": true
1848
+ },
1849
+ "attributes": {}
1850
+ }
1851
+ },
1852
+ "total_flos": 3.675560283088093e+18,
1853
+ "train_batch_size": 4,
1854
+ "trial_name": null,
1855
+ "trial_params": null
1856
+ }
continuous/qwen3/siglip2_sm14_384/finetune/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:1bda933e9dc1791bf66d67cf229734ce9d32d85e6f0d484733475da4e17f1341
3
+ size 6136
continuous/qwen3/siglip2_sm14_384/finetune/vocab.json ADDED
The diff for this file is too large to render. See raw diff
 
continuous/qwen3/siglip2_sm14_384/pretrain/config.json ADDED
@@ -0,0 +1,60 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_attn_implementation_autoset": true,
3
+ "add_faster_video": false,
4
+ "add_time_instruction": false,
5
+ "architectures": [
6
+ "Qwen3ForCausalLM"
7
+ ],
8
+ "attention_bias": false,
9
+ "attention_dropout": 0.0,
10
+ "bos_token_id": 151643,
11
+ "eos_token_id": 151645,
12
+ "faster_token_stride": 10,
13
+ "force_sample": false,
14
+ "head_dim": 128,
15
+ "hidden_act": "silu",
16
+ "hidden_size": 2048,
17
+ "image_aspect_ratio": "square",
18
+ "image_crop_resolution": null,
19
+ "image_grid_pinpoints": null,
20
+ "image_split_resolution": null,
21
+ "initializer_range": 0.02,
22
+ "intermediate_size": 6144,
23
+ "max_position_embeddings": 40960,
24
+ "max_window_layers": 28,
25
+ "mm_hidden_size": 1152,
26
+ "mm_newline_position": "grid",
27
+ "mm_patch_merge_type": "flat",
28
+ "mm_projector_lr": null,
29
+ "mm_projector_type": "mlp2x_gelu",
30
+ "mm_resampler_type": null,
31
+ "mm_spatial_pool_mode": "bilinear",
32
+ "mm_spatial_pool_stride": null,
33
+ "mm_tunable_parts": "mm_mlp_adapter",
34
+ "mm_use_im_patch_token": false,
35
+ "mm_use_im_start_end": false,
36
+ "mm_vision_select_feature": "patch",
37
+ "mm_vision_select_layer": -2,
38
+ "mm_vision_tower": "open_clip_hub:ViT-SO400M-14-SigLIP2-378",
39
+ "mm_vision_tower_lr": null,
40
+ "model_type": "qwen3",
41
+ "num_attention_heads": 16,
42
+ "num_hidden_layers": 28,
43
+ "num_key_value_heads": 8,
44
+ "pos_skipping_range": 4096,
45
+ "rms_norm_eps": 1e-06,
46
+ "rope_scaling": null,
47
+ "rope_theta": 1000000,
48
+ "sliding_window": null,
49
+ "tie_word_embeddings": true,
50
+ "tokenizer_model_max_length": 8192,
51
+ "tokenizer_padding_side": "right",
52
+ "torch_dtype": "bfloat16",
53
+ "transformers_version": "4.51.3",
54
+ "use_cache": true,
55
+ "use_mm_proj": true,
56
+ "use_pos_skipping": false,
57
+ "use_sliding_window": false,
58
+ "vision_tower_pretrained": "/cache/ckpt/download/tokenizer/continuous/siglip2/siglip2_sm14_384.npz",
59
+ "vocab_size": 151936
60
+ }
continuous/qwen3/siglip2_sm14_384/pretrain/mm_projector.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3b3f8a5223e4efc767521bacf530184ac4f8184e40892bed61852395ab016031
3
+ size 26232888
continuous/qwen3/siglip2_sm14_384/pretrain/trainer_state.json ADDED
@@ -0,0 +1,421 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_global_step": null,
3
+ "best_metric": null,
4
+ "best_model_checkpoint": null,
5
+ "epoch": 1.0,
6
+ "eval_steps": 500,
7
+ "global_step": 545,
8
+ "is_hyper_param_search": false,
9
+ "is_local_process_zero": true,
10
+ "is_world_process_zero": true,
11
+ "log_history": [
12
+ {
13
+ "epoch": 0.01834862385321101,
14
+ "grad_norm": 1.8941501379013062,
15
+ "learning_rate": 0.00032142857142857147,
16
+ "loss": 8.846,
17
+ "step": 10
18
+ },
19
+ {
20
+ "epoch": 0.03669724770642202,
21
+ "grad_norm": 0.5077146291732788,
22
+ "learning_rate": 0.0006785714285714287,
23
+ "loss": 5.5441,
24
+ "step": 20
25
+ },
26
+ {
27
+ "epoch": 0.05504587155963303,
28
+ "grad_norm": 0.18718622624874115,
29
+ "learning_rate": 0.0009999907688176173,
30
+ "loss": 4.9249,
31
+ "step": 30
32
+ },
33
+ {
34
+ "epoch": 0.07339449541284404,
35
+ "grad_norm": 0.0829075425863266,
36
+ "learning_rate": 0.0009988834393115766,
37
+ "loss": 4.7619,
38
+ "step": 40
39
+ },
40
+ {
41
+ "epoch": 0.09174311926605505,
42
+ "grad_norm": 0.06917140632867813,
43
+ "learning_rate": 0.0009959345572940772,
44
+ "loss": 4.6938,
45
+ "step": 50
46
+ },
47
+ {
48
+ "epoch": 0.11009174311926606,
49
+ "grad_norm": 0.1796322762966156,
50
+ "learning_rate": 0.0009911550081155983,
51
+ "loss": 4.6561,
52
+ "step": 60
53
+ },
54
+ {
55
+ "epoch": 0.12844036697247707,
56
+ "grad_norm": 0.1438809633255005,
57
+ "learning_rate": 0.0009845624347566433,
58
+ "loss": 4.6147,
59
+ "step": 70
60
+ },
61
+ {
62
+ "epoch": 0.14678899082568808,
63
+ "grad_norm": 0.12461377680301666,
64
+ "learning_rate": 0.0009761811727013547,
65
+ "loss": 4.5976,
66
+ "step": 80
67
+ },
68
+ {
69
+ "epoch": 0.1651376146788991,
70
+ "grad_norm": 0.07040806114673615,
71
+ "learning_rate": 0.0009660421601067666,
72
+ "loss": 4.5606,
73
+ "step": 90
74
+ },
75
+ {
76
+ "epoch": 0.1834862385321101,
77
+ "grad_norm": 0.207647904753685,
78
+ "learning_rate": 0.0009541828235992926,
79
+ "loss": 4.5337,
80
+ "step": 100
81
+ },
82
+ {
83
+ "epoch": 0.2018348623853211,
84
+ "grad_norm": 0.08468043804168701,
85
+ "learning_rate": 0.0009406469401200151,
86
+ "loss": 4.511,
87
+ "step": 110
88
+ },
89
+ {
90
+ "epoch": 0.22018348623853212,
91
+ "grad_norm": 0.10016771405935287,
92
+ "learning_rate": 0.0009254844753287492,
93
+ "loss": 4.4823,
94
+ "step": 120
95
+ },
96
+ {
97
+ "epoch": 0.23853211009174313,
98
+ "grad_norm": 0.12923039495944977,
99
+ "learning_rate": 0.0009087513991633924,
100
+ "loss": 4.4389,
101
+ "step": 130
102
+ },
103
+ {
104
+ "epoch": 0.25688073394495414,
105
+ "grad_norm": 0.09340023994445801,
106
+ "learning_rate": 0.0008905094792353917,
107
+ "loss": 4.3993,
108
+ "step": 140
109
+ },
110
+ {
111
+ "epoch": 0.27522935779816515,
112
+ "grad_norm": 0.0897039845585823,
113
+ "learning_rate": 0.0008708260528239789,
114
+ "loss": 4.3666,
115
+ "step": 150
116
+ },
117
+ {
118
+ "epoch": 0.29357798165137616,
119
+ "grad_norm": 0.10029884427785873,
120
+ "learning_rate": 0.0008497737783108237,
121
+ "loss": 4.3173,
122
+ "step": 160
123
+ },
124
+ {
125
+ "epoch": 0.3119266055045872,
126
+ "grad_norm": 0.07765354216098785,
127
+ "learning_rate": 0.0008274303669726426,
128
+ "loss": 4.2795,
129
+ "step": 170
130
+ },
131
+ {
132
+ "epoch": 0.3302752293577982,
133
+ "grad_norm": 0.11448921263217926,
134
+ "learning_rate": 0.0008038782961218137,
135
+ "loss": 4.2496,
136
+ "step": 180
137
+ },
138
+ {
139
+ "epoch": 0.3486238532110092,
140
+ "grad_norm": 0.1024710014462471,
141
+ "learning_rate": 0.000779204504653894,
142
+ "loss": 4.2274,
143
+ "step": 190
144
+ },
145
+ {
146
+ "epoch": 0.3669724770642202,
147
+ "grad_norm": 0.10741103440523148,
148
+ "learning_rate": 0.0007535000721258791,
149
+ "loss": 4.1903,
150
+ "step": 200
151
+ },
152
+ {
153
+ "epoch": 0.3853211009174312,
154
+ "grad_norm": 0.10220352560281754,
155
+ "learning_rate": 0.00072685988254984,
156
+ "loss": 4.1698,
157
+ "step": 210
158
+ },
159
+ {
160
+ "epoch": 0.4036697247706422,
161
+ "grad_norm": 0.08578507602214813,
162
+ "learning_rate": 0.0006993822741429906,
163
+ "loss": 4.1672,
164
+ "step": 220
165
+ },
166
+ {
167
+ "epoch": 0.42201834862385323,
168
+ "grad_norm": 0.13511505722999573,
169
+ "learning_rate": 0.0006711686763270818,
170
+ "loss": 4.1497,
171
+ "step": 230
172
+ },
173
+ {
174
+ "epoch": 0.44036697247706424,
175
+ "grad_norm": 0.09442631900310516,
176
+ "learning_rate": 0.0006423232353170831,
177
+ "loss": 4.1286,
178
+ "step": 240
179
+ },
180
+ {
181
+ "epoch": 0.45871559633027525,
182
+ "grad_norm": 0.08586172014474869,
183
+ "learning_rate": 0.0006129524296812335,
184
+ "loss": 4.0957,
185
+ "step": 250
186
+ },
187
+ {
188
+ "epoch": 0.47706422018348627,
189
+ "grad_norm": 0.0794299840927124,
190
+ "learning_rate": 0.0005831646772915651,
191
+ "loss": 4.0955,
192
+ "step": 260
193
+ },
194
+ {
195
+ "epoch": 0.4954128440366973,
196
+ "grad_norm": 0.08928098529577255,
197
+ "learning_rate": 0.0005530699351157809,
198
+ "loss": 4.0625,
199
+ "step": 270
200
+ },
201
+ {
202
+ "epoch": 0.5137614678899083,
203
+ "grad_norm": 0.09517360478639603,
204
+ "learning_rate": 0.0005227792933277943,
205
+ "loss": 4.0425,
206
+ "step": 280
207
+ },
208
+ {
209
+ "epoch": 0.5321100917431193,
210
+ "grad_norm": 0.09574073553085327,
211
+ "learning_rate": 0.0004924045652352092,
212
+ "loss": 4.0095,
213
+ "step": 290
214
+ },
215
+ {
216
+ "epoch": 0.5504587155963303,
217
+ "grad_norm": 0.11322334408760071,
218
+ "learning_rate": 0.0004620578745374633,
219
+ "loss": 3.9839,
220
+ "step": 300
221
+ },
222
+ {
223
+ "epoch": 0.5688073394495413,
224
+ "grad_norm": 0.1056467741727829,
225
+ "learning_rate": 0.0004318512414382059,
226
+ "loss": 3.9702,
227
+ "step": 310
228
+ },
229
+ {
230
+ "epoch": 0.5871559633027523,
231
+ "grad_norm": 0.09862352162599564,
232
+ "learning_rate": 0.0004018961691397148,
233
+ "loss": 3.9438,
234
+ "step": 320
235
+ },
236
+ {
237
+ "epoch": 0.6055045871559633,
238
+ "grad_norm": 0.08425777405500412,
239
+ "learning_rate": 0.0003723032322457458,
240
+ "loss": 3.9065,
241
+ "step": 330
242
+ },
243
+ {
244
+ "epoch": 0.6238532110091743,
245
+ "grad_norm": 0.08464296907186508,
246
+ "learning_rate": 0.0003431816685921625,
247
+ "loss": 3.8965,
248
+ "step": 340
249
+ },
250
+ {
251
+ "epoch": 0.6422018348623854,
252
+ "grad_norm": 0.07609152793884277,
253
+ "learning_rate": 0.0003146389760120416,
254
+ "loss": 3.8889,
255
+ "step": 350
256
+ },
257
+ {
258
+ "epoch": 0.6605504587155964,
259
+ "grad_norm": 0.08307695388793945,
260
+ "learning_rate": 0.00028678051552373485,
261
+ "loss": 3.8905,
262
+ "step": 360
263
+ },
264
+ {
265
+ "epoch": 0.6788990825688074,
266
+ "grad_norm": 0.078050397336483,
267
+ "learning_rate": 0.0002597091224066581,
268
+ "loss": 3.8685,
269
+ "step": 370
270
+ },
271
+ {
272
+ "epoch": 0.6972477064220184,
273
+ "grad_norm": 0.08195412158966064,
274
+ "learning_rate": 0.0002335247266004629,
275
+ "loss": 3.8775,
276
+ "step": 380
277
+ },
278
+ {
279
+ "epoch": 0.7155963302752294,
280
+ "grad_norm": 0.07861229032278061,
281
+ "learning_rate": 0.000208323983828828,
282
+ "loss": 3.8598,
283
+ "step": 390
284
+ },
285
+ {
286
+ "epoch": 0.7339449541284404,
287
+ "grad_norm": 0.07933512330055237,
288
+ "learning_rate": 0.0001841999188095224,
289
+ "loss": 3.8534,
290
+ "step": 400
291
+ },
292
+ {
293
+ "epoch": 0.7522935779816514,
294
+ "grad_norm": 0.07791566103696823,
295
+ "learning_rate": 0.00016124158186777676,
296
+ "loss": 3.8567,
297
+ "step": 410
298
+ },
299
+ {
300
+ "epoch": 0.7706422018348624,
301
+ "grad_norm": 0.07676967233419418,
302
+ "learning_rate": 0.0001395337202205211,
303
+ "loss": 3.8478,
304
+ "step": 420
305
+ },
306
+ {
307
+ "epoch": 0.7889908256880734,
308
+ "grad_norm": 0.08134885877370834,
309
+ "learning_rate": 0.00011915646514489292,
310
+ "loss": 3.8472,
311
+ "step": 430
312
+ },
313
+ {
314
+ "epoch": 0.8073394495412844,
315
+ "grad_norm": 0.08101513236761093,
316
+ "learning_rate": 0.00010018503618578818,
317
+ "loss": 3.851,
318
+ "step": 440
319
+ },
320
+ {
321
+ "epoch": 0.8256880733944955,
322
+ "grad_norm": 0.07142005860805511,
323
+ "learning_rate": 8.268946349432583e-05,
324
+ "loss": 3.8417,
325
+ "step": 450
326
+ },
327
+ {
328
+ "epoch": 0.8440366972477065,
329
+ "grad_norm": 0.07557690143585205,
330
+ "learning_rate": 6.673432932217117e-05,
331
+ "loss": 3.8385,
332
+ "step": 460
333
+ },
334
+ {
335
+ "epoch": 0.8623853211009175,
336
+ "grad_norm": 0.07234933227300644,
337
+ "learning_rate": 5.2378529625954684e-05,
338
+ "loss": 3.8231,
339
+ "step": 470
340
+ },
341
+ {
342
+ "epoch": 0.8807339449541285,
343
+ "grad_norm": 0.07122330367565155,
344
+ "learning_rate": 3.9675056661785556e-05,
345
+ "loss": 3.8331,
346
+ "step": 480
347
+ },
348
+ {
349
+ "epoch": 0.8990825688073395,
350
+ "grad_norm": 0.06965400278568268,
351
+ "learning_rate": 2.8670803372379538e-05,
352
+ "loss": 3.832,
353
+ "step": 490
354
+ },
355
+ {
356
+ "epoch": 0.9174311926605505,
357
+ "grad_norm": 0.06935345381498337,
358
+ "learning_rate": 1.9406390288876586e-05,
359
+ "loss": 3.843,
360
+ "step": 500
361
+ },
362
+ {
363
+ "epoch": 0.9357798165137615,
364
+ "grad_norm": 0.06357274204492569,
365
+ "learning_rate": 1.191601558631461e-05,
366
+ "loss": 3.8355,
367
+ "step": 510
368
+ },
369
+ {
370
+ "epoch": 0.9541284403669725,
371
+ "grad_norm": 0.06666282564401627,
372
+ "learning_rate": 6.2273288462559305e-06,
373
+ "loss": 3.828,
374
+ "step": 520
375
+ },
376
+ {
377
+ "epoch": 0.9724770642201835,
378
+ "grad_norm": 0.06515973806381226,
379
+ "learning_rate": 2.361328992552314e-06,
380
+ "loss": 3.8305,
381
+ "step": 530
382
+ },
383
+ {
384
+ "epoch": 0.9908256880733946,
385
+ "grad_norm": 0.07123097032308578,
386
+ "learning_rate": 3.322867770044202e-07,
387
+ "loss": 3.8301,
388
+ "step": 540
389
+ },
390
+ {
391
+ "epoch": 1.0,
392
+ "step": 545,
393
+ "total_flos": 3.52940365088555e+17,
394
+ "train_loss": 4.233528980640096,
395
+ "train_runtime": 5763.1112,
396
+ "train_samples_per_second": 96.845,
397
+ "train_steps_per_second": 0.095
398
+ }
399
+ ],
400
+ "logging_steps": 10,
401
+ "max_steps": 545,
402
+ "num_input_tokens_seen": 0,
403
+ "num_train_epochs": 1,
404
+ "save_steps": 500,
405
+ "stateful_callbacks": {
406
+ "TrainerControl": {
407
+ "args": {
408
+ "should_epoch_stop": false,
409
+ "should_evaluate": false,
410
+ "should_log": false,
411
+ "should_save": false,
412
+ "should_training_stop": false
413
+ },
414
+ "attributes": {}
415
+ }
416
+ },
417
+ "total_flos": 3.52940365088555e+17,
418
+ "train_batch_size": 32,
419
+ "trial_name": null,
420
+ "trial_params": null
421
+ }
continuous/qwen3/siglip2_sm16_256/finetune/added_tokens.json ADDED
@@ -0,0 +1,28 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "</think>": 151668,
3
+ "</tool_call>": 151658,
4
+ "</tool_response>": 151666,
5
+ "<think>": 151667,
6
+ "<tool_call>": 151657,
7
+ "<tool_response>": 151665,
8
+ "<|box_end|>": 151649,
9
+ "<|box_start|>": 151648,
10
+ "<|endoftext|>": 151643,
11
+ "<|file_sep|>": 151664,
12
+ "<|fim_middle|>": 151660,
13
+ "<|fim_pad|>": 151662,
14
+ "<|fim_prefix|>": 151659,
15
+ "<|fim_suffix|>": 151661,
16
+ "<|im_end|>": 151645,
17
+ "<|im_start|>": 151644,
18
+ "<|image_pad|>": 151655,
19
+ "<|object_ref_end|>": 151647,
20
+ "<|object_ref_start|>": 151646,
21
+ "<|quad_end|>": 151651,
22
+ "<|quad_start|>": 151650,
23
+ "<|repo_name|>": 151663,
24
+ "<|video_pad|>": 151656,
25
+ "<|vision_end|>": 151653,
26
+ "<|vision_pad|>": 151654,
27
+ "<|vision_start|>": 151652
28
+ }
continuous/qwen3/siglip2_sm16_256/finetune/config.json ADDED
@@ -0,0 +1,59 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "add_faster_video": false,
3
+ "add_time_instruction": false,
4
+ "architectures": [
5
+ "LlavaQwen3ForCausalLM"
6
+ ],
7
+ "attention_bias": false,
8
+ "attention_dropout": 0.0,
9
+ "bos_token_id": 151643,
10
+ "eos_token_id": 151645,
11
+ "faster_token_stride": 10,
12
+ "force_sample": false,
13
+ "head_dim": 128,
14
+ "hidden_act": "silu",
15
+ "hidden_size": 2048,
16
+ "image_aspect_ratio": "square",
17
+ "image_crop_resolution": null,
18
+ "image_grid_pinpoints": null,
19
+ "image_split_resolution": null,
20
+ "initializer_range": 0.02,
21
+ "intermediate_size": 6144,
22
+ "max_position_embeddings": 40960,
23
+ "max_window_layers": 28,
24
+ "mm_hidden_size": 1152,
25
+ "mm_newline_position": "grid",
26
+ "mm_patch_merge_type": "flat",
27
+ "mm_projector_lr": 0.0001,
28
+ "mm_projector_type": "mlp2x_gelu",
29
+ "mm_resampler_type": null,
30
+ "mm_spatial_pool_mode": "bilinear",
31
+ "mm_spatial_pool_stride": null,
32
+ "mm_tunable_parts": "mm_mlp_adapter,mm_language_model",
33
+ "mm_use_im_patch_token": false,
34
+ "mm_use_im_start_end": false,
35
+ "mm_vision_select_feature": "patch",
36
+ "mm_vision_select_layer": -2,
37
+ "mm_vision_tower": "open_clip_hub:ViT-SO400M-16-SigLIP2-256",
38
+ "mm_vision_tower_lr": null,
39
+ "model_type": "qwen3",
40
+ "num_attention_heads": 16,
41
+ "num_hidden_layers": 28,
42
+ "num_key_value_heads": 8,
43
+ "pos_skipping_range": 4096,
44
+ "rms_norm_eps": 1e-06,
45
+ "rope_scaling": null,
46
+ "rope_theta": 1000000,
47
+ "sliding_window": null,
48
+ "tie_word_embeddings": true,
49
+ "tokenizer_model_max_length": 8192,
50
+ "tokenizer_padding_side": "right",
51
+ "torch_dtype": "bfloat16",
52
+ "transformers_version": "4.51.3",
53
+ "use_cache": true,
54
+ "use_mm_proj": true,
55
+ "use_pos_skipping": false,
56
+ "use_sliding_window": false,
57
+ "vision_tower_pretrained": "/cache/ckpt/download/tokenizer/continuous/siglip2/siglip2_sm16_256.npz",
58
+ "vocab_size": 151936
59
+ }
continuous/qwen3/siglip2_sm16_256/finetune/generation_config.json ADDED
@@ -0,0 +1,13 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "bos_token_id": 151643,
3
+ "do_sample": true,
4
+ "eos_token_id": [
5
+ 151645,
6
+ 151643
7
+ ],
8
+ "pad_token_id": 151643,
9
+ "temperature": 0.6,
10
+ "top_k": 20,
11
+ "top_p": 0.95,
12
+ "transformers_version": "4.51.3"
13
+ }
continuous/qwen3/siglip2_sm16_256/finetune/merges.txt ADDED
The diff for this file is too large to render. See raw diff
 
continuous/qwen3/siglip2_sm16_256/finetune/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:85521dc2d1ca9fe333ee699eabd5a5f5a87d5191a416636fe1a4f36e49437250
3
+ size 4945569424
continuous/qwen3/siglip2_sm16_256/finetune/special_tokens_map.json ADDED
@@ -0,0 +1,31 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "additional_special_tokens": [
3
+ "<|im_start|>",
4
+ "<|im_end|>",
5
+ "<|object_ref_start|>",
6
+ "<|object_ref_end|>",
7
+ "<|box_start|>",
8
+ "<|box_end|>",
9
+ "<|quad_start|>",
10
+ "<|quad_end|>",
11
+ "<|vision_start|>",
12
+ "<|vision_end|>",
13
+ "<|vision_pad|>",
14
+ "<|image_pad|>",
15
+ "<|video_pad|>"
16
+ ],
17
+ "eos_token": {
18
+ "content": "<|im_end|>",
19
+ "lstrip": false,
20
+ "normalized": false,
21
+ "rstrip": false,
22
+ "single_word": false
23
+ },
24
+ "pad_token": {
25
+ "content": "<|endoftext|>",
26
+ "lstrip": false,
27
+ "normalized": false,
28
+ "rstrip": false,
29
+ "single_word": false
30
+ }
31
+ }
continuous/qwen3/siglip2_sm16_256/finetune/tokenizer.json ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:aeb13307a71acd8fe81861d94ad54ab689df773318809eed3cbe794b4492dae4
3
+ size 11422654
continuous/qwen3/siglip2_sm16_256/finetune/tokenizer_config.json ADDED
@@ -0,0 +1,241 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "add_bos_token": false,
3
+ "add_prefix_space": false,
4
+ "added_tokens_decoder": {
5
+ "151643": {
6
+ "content": "<|endoftext|>",
7
+ "lstrip": false,
8
+ "normalized": false,
9
+ "rstrip": false,
10
+ "single_word": false,
11
+ "special": true
12
+ },
13
+ "151644": {
14
+ "content": "<|im_start|>",
15
+ "lstrip": false,
16
+ "normalized": false,
17
+ "rstrip": false,
18
+ "single_word": false,
19
+ "special": true
20
+ },
21
+ "151645": {
22
+ "content": "<|im_end|>",
23
+ "lstrip": false,
24
+ "normalized": false,
25
+ "rstrip": false,
26
+ "single_word": false,
27
+ "special": true
28
+ },
29
+ "151646": {
30
+ "content": "<|object_ref_start|>",
31
+ "lstrip": false,
32
+ "normalized": false,
33
+ "rstrip": false,
34
+ "single_word": false,
35
+ "special": true
36
+ },
37
+ "151647": {
38
+ "content": "<|object_ref_end|>",
39
+ "lstrip": false,
40
+ "normalized": false,
41
+ "rstrip": false,
42
+ "single_word": false,
43
+ "special": true
44
+ },
45
+ "151648": {
46
+ "content": "<|box_start|>",
47
+ "lstrip": false,
48
+ "normalized": false,
49
+ "rstrip": false,
50
+ "single_word": false,
51
+ "special": true
52
+ },
53
+ "151649": {
54
+ "content": "<|box_end|>",
55
+ "lstrip": false,
56
+ "normalized": false,
57
+ "rstrip": false,
58
+ "single_word": false,
59
+ "special": true
60
+ },
61
+ "151650": {
62
+ "content": "<|quad_start|>",
63
+ "lstrip": false,
64
+ "normalized": false,
65
+ "rstrip": false,
66
+ "single_word": false,
67
+ "special": true
68
+ },
69
+ "151651": {
70
+ "content": "<|quad_end|>",
71
+ "lstrip": false,
72
+ "normalized": false,
73
+ "rstrip": false,
74
+ "single_word": false,
75
+ "special": true
76
+ },
77
+ "151652": {
78
+ "content": "<|vision_start|>",
79
+ "lstrip": false,
80
+ "normalized": false,
81
+ "rstrip": false,
82
+ "single_word": false,
83
+ "special": true
84
+ },
85
+ "151653": {
86
+ "content": "<|vision_end|>",
87
+ "lstrip": false,
88
+ "normalized": false,
89
+ "rstrip": false,
90
+ "single_word": false,
91
+ "special": true
92
+ },
93
+ "151654": {
94
+ "content": "<|vision_pad|>",
95
+ "lstrip": false,
96
+ "normalized": false,
97
+ "rstrip": false,
98
+ "single_word": false,
99
+ "special": true
100
+ },
101
+ "151655": {
102
+ "content": "<|image_pad|>",
103
+ "lstrip": false,
104
+ "normalized": false,
105
+ "rstrip": false,
106
+ "single_word": false,
107
+ "special": true
108
+ },
109
+ "151656": {
110
+ "content": "<|video_pad|>",
111
+ "lstrip": false,
112
+ "normalized": false,
113
+ "rstrip": false,
114
+ "single_word": false,
115
+ "special": true
116
+ },
117
+ "151657": {
118
+ "content": "<tool_call>",
119
+ "lstrip": false,
120
+ "normalized": false,
121
+ "rstrip": false,
122
+ "single_word": false,
123
+ "special": false
124
+ },
125
+ "151658": {
126
+ "content": "</tool_call>",
127
+ "lstrip": false,
128
+ "normalized": false,
129
+ "rstrip": false,
130
+ "single_word": false,
131
+ "special": false
132
+ },
133
+ "151659": {
134
+ "content": "<|fim_prefix|>",
135
+ "lstrip": false,
136
+ "normalized": false,
137
+ "rstrip": false,
138
+ "single_word": false,
139
+ "special": false
140
+ },
141
+ "151660": {
142
+ "content": "<|fim_middle|>",
143
+ "lstrip": false,
144
+ "normalized": false,
145
+ "rstrip": false,
146
+ "single_word": false,
147
+ "special": false
148
+ },
149
+ "151661": {
150
+ "content": "<|fim_suffix|>",
151
+ "lstrip": false,
152
+ "normalized": false,
153
+ "rstrip": false,
154
+ "single_word": false,
155
+ "special": false
156
+ },
157
+ "151662": {
158
+ "content": "<|fim_pad|>",
159
+ "lstrip": false,
160
+ "normalized": false,
161
+ "rstrip": false,
162
+ "single_word": false,
163
+ "special": false
164
+ },
165
+ "151663": {
166
+ "content": "<|repo_name|>",
167
+ "lstrip": false,
168
+ "normalized": false,
169
+ "rstrip": false,
170
+ "single_word": false,
171
+ "special": false
172
+ },
173
+ "151664": {
174
+ "content": "<|file_sep|>",
175
+ "lstrip": false,
176
+ "normalized": false,
177
+ "rstrip": false,
178
+ "single_word": false,
179
+ "special": false
180
+ },
181
+ "151665": {
182
+ "content": "<tool_response>",
183
+ "lstrip": false,
184
+ "normalized": false,
185
+ "rstrip": false,
186
+ "single_word": false,
187
+ "special": false
188
+ },
189
+ "151666": {
190
+ "content": "</tool_response>",
191
+ "lstrip": false,
192
+ "normalized": false,
193
+ "rstrip": false,
194
+ "single_word": false,
195
+ "special": false
196
+ },
197
+ "151667": {
198
+ "content": "<think>",
199
+ "lstrip": false,
200
+ "normalized": false,
201
+ "rstrip": false,
202
+ "single_word": false,
203
+ "special": false
204
+ },
205
+ "151668": {
206
+ "content": "</think>",
207
+ "lstrip": false,
208
+ "normalized": false,
209
+ "rstrip": false,
210
+ "single_word": false,
211
+ "special": false
212
+ }
213
+ },
214
+ "additional_special_tokens": [
215
+ "<|im_start|>",
216
+ "<|im_end|>",
217
+ "<|object_ref_start|>",
218
+ "<|object_ref_end|>",
219
+ "<|box_start|>",
220
+ "<|box_end|>",
221
+ "<|quad_start|>",
222
+ "<|quad_end|>",
223
+ "<|vision_start|>",
224
+ "<|vision_end|>",
225
+ "<|vision_pad|>",
226
+ "<|image_pad|>",
227
+ "<|video_pad|>"
228
+ ],
229
+ "bos_token": null,
230
+ "chat_template": "{%- if tools %}\n {{- '<|im_start|>system\\n' }}\n {%- if messages[0].role == 'system' %}\n {{- messages[0].content + '\\n\\n' }}\n {%- endif %}\n {{- \"# Tools\\n\\nYou may call one or more functions to assist with the user query.\\n\\nYou are provided with function signatures within <tools></tools> XML tags:\\n<tools>\" }}\n {%- for tool in tools %}\n {{- \"\\n\" }}\n {{- tool | tojson }}\n {%- endfor %}\n {{- \"\\n</tools>\\n\\nFor each function call, return a json object with function name and arguments within <tool_call></tool_call> XML tags:\\n<tool_call>\\n{\\\"name\\\": <function-name>, \\\"arguments\\\": <args-json-object>}\\n</tool_call><|im_end|>\\n\" }}\n{%- else %}\n {%- if messages[0].role == 'system' %}\n {{- '<|im_start|>system\\n' + messages[0].content + '<|im_end|>\\n' }}\n {%- endif %}\n{%- endif %}\n{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}\n{%- for message in messages[::-1] %}\n {%- set index = (messages|length - 1) - loop.index0 %}\n {%- if ns.multi_step_tool and message.role == \"user\" and message.content is string and not(message.content.startswith('<tool_response>') and message.content.endswith('</tool_response>')) %}\n {%- set ns.multi_step_tool = false %}\n {%- set ns.last_query_index = index %}\n {%- endif %}\n{%- endfor %}\n{%- for message in messages %}\n {%- if message.content is string %}\n {%- set content = message.content %}\n {%- else %}\n {%- set content = '' %}\n {%- endif %}\n {%- if (message.role == \"user\") or (message.role == \"system\" and not loop.first) %}\n {{- '<|im_start|>' + message.role + '\\n' + content + '<|im_end|>' + '\\n' }}\n {%- elif message.role == \"assistant\" %}\n {%- set reasoning_content = '' %}\n {%- if message.reasoning_content is string %}\n {%- set reasoning_content = message.reasoning_content %}\n {%- else %}\n {%- if '</think>' in content %}\n {%- set reasoning_content = content.split('</think>')[0].rstrip('\\n').split('<think>')[-1].lstrip('\\n') %}\n {%- set content = content.split('</think>')[-1].lstrip('\\n') %}\n {%- endif %}\n {%- endif %}\n {%- if loop.index0 > ns.last_query_index %}\n {%- if loop.last or (not loop.last and reasoning_content) %}\n {{- '<|im_start|>' + message.role + '\\n<think>\\n' + reasoning_content.strip('\\n') + '\\n</think>\\n\\n' + content.lstrip('\\n') }}\n {%- else %}\n {{- '<|im_start|>' + message.role + '\\n' + content }}\n {%- endif %}\n {%- else %}\n {{- '<|im_start|>' + message.role + '\\n' + content }}\n {%- endif %}\n {%- if message.tool_calls %}\n {%- for tool_call in message.tool_calls %}\n {%- if (loop.first and content) or (not loop.first) %}\n {{- '\\n' }}\n {%- endif %}\n {%- if tool_call.function %}\n {%- set tool_call = tool_call.function %}\n {%- endif %}\n {{- '<tool_call>\\n{\"name\": \"' }}\n {{- tool_call.name }}\n {{- '\", \"arguments\": ' }}\n {%- if tool_call.arguments is string %}\n {{- tool_call.arguments }}\n {%- else %}\n {{- tool_call.arguments | tojson }}\n {%- endif %}\n {{- '}\\n</tool_call>' }}\n {%- endfor %}\n {%- endif %}\n {{- '<|im_end|>\\n' }}\n {%- elif message.role == \"tool\" %}\n {%- if loop.first or (messages[loop.index0 - 1].role != \"tool\") %}\n {{- '<|im_start|>user' }}\n {%- endif %}\n {{- '\\n<tool_response>\\n' }}\n {{- content }}\n {{- '\\n</tool_response>' }}\n {%- if loop.last or (messages[loop.index0 + 1].role != \"tool\") %}\n {{- '<|im_end|>\\n' }}\n {%- endif %}\n {%- endif %}\n{%- endfor %}\n{%- if add_generation_prompt %}\n {{- '<|im_start|>assistant\\n' }}\n {%- if enable_thinking is defined and enable_thinking is false %}\n {{- '<think>\\n\\n</think>\\n\\n' }}\n {%- endif %}\n{%- endif %}",
231
+ "clean_up_tokenization_spaces": false,
232
+ "eos_token": "<|im_end|>",
233
+ "errors": "replace",
234
+ "extra_special_tokens": {},
235
+ "model_max_length": 8192,
236
+ "pad_token": "<|endoftext|>",
237
+ "padding_side": "right",
238
+ "split_special_tokens": false,
239
+ "tokenizer_class": "Qwen2Tokenizer",
240
+ "unk_token": null
241
+ }
continuous/qwen3/siglip2_sm16_256/finetune/trainer_state.json ADDED
@@ -0,0 +1,1856 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_global_step": null,
3
+ "best_metric": null,
4
+ "best_model_checkpoint": null,
5
+ "epoch": 0.9999037443449803,
6
+ "eval_steps": 500,
7
+ "global_step": 2597,
8
+ "is_hyper_param_search": false,
9
+ "is_local_process_zero": true,
10
+ "is_world_process_zero": true,
11
+ "log_history": [
12
+ {
13
+ "epoch": 0.003850226200789296,
14
+ "grad_norm": 16.443357467651367,
15
+ "learning_rate": 1.153846153846154e-06,
16
+ "loss": 2.405,
17
+ "step": 10
18
+ },
19
+ {
20
+ "epoch": 0.007700452401578592,
21
+ "grad_norm": 6.338305473327637,
22
+ "learning_rate": 2.435897435897436e-06,
23
+ "loss": 1.8739,
24
+ "step": 20
25
+ },
26
+ {
27
+ "epoch": 0.01155067860236789,
28
+ "grad_norm": 4.257731914520264,
29
+ "learning_rate": 3.7179487179487184e-06,
30
+ "loss": 1.4438,
31
+ "step": 30
32
+ },
33
+ {
34
+ "epoch": 0.015400904803157185,
35
+ "grad_norm": 1.1406962871551514,
36
+ "learning_rate": 5e-06,
37
+ "loss": 1.113,
38
+ "step": 40
39
+ },
40
+ {
41
+ "epoch": 0.019251131003946483,
42
+ "grad_norm": 0.8033566474914551,
43
+ "learning_rate": 6.282051282051282e-06,
44
+ "loss": 1.0093,
45
+ "step": 50
46
+ },
47
+ {
48
+ "epoch": 0.02310135720473578,
49
+ "grad_norm": 0.39472508430480957,
50
+ "learning_rate": 7.564102564102564e-06,
51
+ "loss": 0.922,
52
+ "step": 60
53
+ },
54
+ {
55
+ "epoch": 0.026951583405525073,
56
+ "grad_norm": 0.30989113450050354,
57
+ "learning_rate": 8.846153846153847e-06,
58
+ "loss": 0.876,
59
+ "step": 70
60
+ },
61
+ {
62
+ "epoch": 0.03080180960631437,
63
+ "grad_norm": 0.2950158417224884,
64
+ "learning_rate": 9.999996111488723e-06,
65
+ "loss": 0.8726,
66
+ "step": 80
67
+ },
68
+ {
69
+ "epoch": 0.034652035807103666,
70
+ "grad_norm": 0.2915459871292114,
71
+ "learning_rate": 9.999529497453782e-06,
72
+ "loss": 0.8426,
73
+ "step": 90
74
+ },
75
+ {
76
+ "epoch": 0.038502262007892966,
77
+ "grad_norm": 0.24879306554794312,
78
+ "learning_rate": 9.998285264324193e-06,
79
+ "loss": 0.8155,
80
+ "step": 100
81
+ },
82
+ {
83
+ "epoch": 0.04235248820868226,
84
+ "grad_norm": 0.27146589756011963,
85
+ "learning_rate": 9.996263605626052e-06,
86
+ "loss": 0.8035,
87
+ "step": 110
88
+ },
89
+ {
90
+ "epoch": 0.04620271440947156,
91
+ "grad_norm": 0.26007670164108276,
92
+ "learning_rate": 9.99346483580503e-06,
93
+ "loss": 0.8137,
94
+ "step": 120
95
+ },
96
+ {
97
+ "epoch": 0.05005294061026085,
98
+ "grad_norm": 0.24125677347183228,
99
+ "learning_rate": 9.989889390177467e-06,
100
+ "loss": 0.8048,
101
+ "step": 130
102
+ },
103
+ {
104
+ "epoch": 0.053903166811050146,
105
+ "grad_norm": 0.279975026845932,
106
+ "learning_rate": 9.985537824862647e-06,
107
+ "loss": 0.8047,
108
+ "step": 140
109
+ },
110
+ {
111
+ "epoch": 0.057753393011839446,
112
+ "grad_norm": 0.2988622188568115,
113
+ "learning_rate": 9.980410816696317e-06,
114
+ "loss": 0.7845,
115
+ "step": 150
116
+ },
117
+ {
118
+ "epoch": 0.06160361921262874,
119
+ "grad_norm": 0.2616492211818695,
120
+ "learning_rate": 9.97450916312541e-06,
121
+ "loss": 0.7947,
122
+ "step": 160
123
+ },
124
+ {
125
+ "epoch": 0.06545384541341803,
126
+ "grad_norm": 0.22167420387268066,
127
+ "learning_rate": 9.967833782084004e-06,
128
+ "loss": 0.7968,
129
+ "step": 170
130
+ },
131
+ {
132
+ "epoch": 0.06930407161420733,
133
+ "grad_norm": 0.2548247277736664,
134
+ "learning_rate": 9.960385711850554e-06,
135
+ "loss": 0.8194,
136
+ "step": 180
137
+ },
138
+ {
139
+ "epoch": 0.07315429781499663,
140
+ "grad_norm": 0.26337409019470215,
141
+ "learning_rate": 9.952166110886394e-06,
142
+ "loss": 0.8113,
143
+ "step": 190
144
+ },
145
+ {
146
+ "epoch": 0.07700452401578593,
147
+ "grad_norm": 0.269176185131073,
148
+ "learning_rate": 9.943176257655567e-06,
149
+ "loss": 0.797,
150
+ "step": 200
151
+ },
152
+ {
153
+ "epoch": 0.08085475021657522,
154
+ "grad_norm": 0.26780539751052856,
155
+ "learning_rate": 9.933417550425952e-06,
156
+ "loss": 0.8004,
157
+ "step": 210
158
+ },
159
+ {
160
+ "epoch": 0.08470497641736452,
161
+ "grad_norm": 0.25500211119651794,
162
+ "learning_rate": 9.9228915070518e-06,
163
+ "loss": 0.7789,
164
+ "step": 220
165
+ },
166
+ {
167
+ "epoch": 0.08855520261815382,
168
+ "grad_norm": 0.24757502973079681,
169
+ "learning_rate": 9.911599764737632e-06,
170
+ "loss": 0.7804,
171
+ "step": 230
172
+ },
173
+ {
174
+ "epoch": 0.09240542881894312,
175
+ "grad_norm": 0.24312543869018555,
176
+ "learning_rate": 9.899544079783606e-06,
177
+ "loss": 0.7776,
178
+ "step": 240
179
+ },
180
+ {
181
+ "epoch": 0.0962556550197324,
182
+ "grad_norm": 0.28034642338752747,
183
+ "learning_rate": 9.886726327312338e-06,
184
+ "loss": 0.7949,
185
+ "step": 250
186
+ },
187
+ {
188
+ "epoch": 0.1001058812205217,
189
+ "grad_norm": 0.24145060777664185,
190
+ "learning_rate": 9.87314850097724e-06,
191
+ "loss": 0.7882,
192
+ "step": 260
193
+ },
194
+ {
195
+ "epoch": 0.103956107421311,
196
+ "grad_norm": 0.24968814849853516,
197
+ "learning_rate": 9.858812712652452e-06,
198
+ "loss": 0.7772,
199
+ "step": 270
200
+ },
201
+ {
202
+ "epoch": 0.10780633362210029,
203
+ "grad_norm": 0.24204115569591522,
204
+ "learning_rate": 9.843721192104341e-06,
205
+ "loss": 0.7748,
206
+ "step": 280
207
+ },
208
+ {
209
+ "epoch": 0.11165655982288959,
210
+ "grad_norm": 0.23302356898784637,
211
+ "learning_rate": 9.827876286644698e-06,
212
+ "loss": 0.795,
213
+ "step": 290
214
+ },
215
+ {
216
+ "epoch": 0.11550678602367889,
217
+ "grad_norm": 0.21392175555229187,
218
+ "learning_rate": 9.811280460765643e-06,
219
+ "loss": 0.7573,
220
+ "step": 300
221
+ },
222
+ {
223
+ "epoch": 0.11935701222446819,
224
+ "grad_norm": 0.2418786883354187,
225
+ "learning_rate": 9.793936295756292e-06,
226
+ "loss": 0.7663,
227
+ "step": 310
228
+ },
229
+ {
230
+ "epoch": 0.12320723842525748,
231
+ "grad_norm": 0.24791035056114197,
232
+ "learning_rate": 9.775846489301279e-06,
233
+ "loss": 0.772,
234
+ "step": 320
235
+ },
236
+ {
237
+ "epoch": 0.12705746462604678,
238
+ "grad_norm": 0.24030518531799316,
239
+ "learning_rate": 9.757013855061157e-06,
240
+ "loss": 0.7689,
241
+ "step": 330
242
+ },
243
+ {
244
+ "epoch": 0.13090769082683607,
245
+ "grad_norm": 0.26338300108909607,
246
+ "learning_rate": 9.737441322234758e-06,
247
+ "loss": 0.7793,
248
+ "step": 340
249
+ },
250
+ {
251
+ "epoch": 0.13475791702762538,
252
+ "grad_norm": 0.26197317242622375,
253
+ "learning_rate": 9.717131935103605e-06,
254
+ "loss": 0.7815,
255
+ "step": 350
256
+ },
257
+ {
258
+ "epoch": 0.13860814322841467,
259
+ "grad_norm": 0.24147701263427734,
260
+ "learning_rate": 9.696088852558398e-06,
261
+ "loss": 0.7676,
262
+ "step": 360
263
+ },
264
+ {
265
+ "epoch": 0.14245836942920398,
266
+ "grad_norm": 0.24064692854881287,
267
+ "learning_rate": 9.674315347607686e-06,
268
+ "loss": 0.7964,
269
+ "step": 370
270
+ },
271
+ {
272
+ "epoch": 0.14630859562999327,
273
+ "grad_norm": 0.23855337500572205,
274
+ "learning_rate": 9.651814806868796e-06,
275
+ "loss": 0.7564,
276
+ "step": 380
277
+ },
278
+ {
279
+ "epoch": 0.15015882183078255,
280
+ "grad_norm": 0.23768046498298645,
281
+ "learning_rate": 9.62859073004107e-06,
282
+ "loss": 0.7587,
283
+ "step": 390
284
+ },
285
+ {
286
+ "epoch": 0.15400904803157187,
287
+ "grad_norm": 0.23103675246238708,
288
+ "learning_rate": 9.604646729361547e-06,
289
+ "loss": 0.7573,
290
+ "step": 400
291
+ },
292
+ {
293
+ "epoch": 0.15785927423236115,
294
+ "grad_norm": 0.2336001843214035,
295
+ "learning_rate": 9.579986529043099e-06,
296
+ "loss": 0.7473,
297
+ "step": 410
298
+ },
299
+ {
300
+ "epoch": 0.16170950043315044,
301
+ "grad_norm": 0.24959221482276917,
302
+ "learning_rate": 9.554613964695189e-06,
303
+ "loss": 0.7699,
304
+ "step": 420
305
+ },
306
+ {
307
+ "epoch": 0.16555972663393975,
308
+ "grad_norm": 0.2449207603931427,
309
+ "learning_rate": 9.528532982727281e-06,
310
+ "loss": 0.7623,
311
+ "step": 430
312
+ },
313
+ {
314
+ "epoch": 0.16940995283472904,
315
+ "grad_norm": 0.24203158915042877,
316
+ "learning_rate": 9.501747639735018e-06,
317
+ "loss": 0.7547,
318
+ "step": 440
319
+ },
320
+ {
321
+ "epoch": 0.17326017903551832,
322
+ "grad_norm": 0.24314452707767487,
323
+ "learning_rate": 9.474262101869272e-06,
324
+ "loss": 0.7461,
325
+ "step": 450
326
+ },
327
+ {
328
+ "epoch": 0.17711040523630764,
329
+ "grad_norm": 0.23702624440193176,
330
+ "learning_rate": 9.446080644188141e-06,
331
+ "loss": 0.7538,
332
+ "step": 460
333
+ },
334
+ {
335
+ "epoch": 0.18096063143709692,
336
+ "grad_norm": 0.23404332995414734,
337
+ "learning_rate": 9.417207649992017e-06,
338
+ "loss": 0.7569,
339
+ "step": 470
340
+ },
341
+ {
342
+ "epoch": 0.18481085763788624,
343
+ "grad_norm": 0.25266382098197937,
344
+ "learning_rate": 9.387647610141816e-06,
345
+ "loss": 0.7546,
346
+ "step": 480
347
+ },
348
+ {
349
+ "epoch": 0.18866108383867553,
350
+ "grad_norm": 0.22808560729026794,
351
+ "learning_rate": 9.35740512236047e-06,
352
+ "loss": 0.7442,
353
+ "step": 490
354
+ },
355
+ {
356
+ "epoch": 0.1925113100394648,
357
+ "grad_norm": 0.22755993902683258,
358
+ "learning_rate": 9.326484890517809e-06,
359
+ "loss": 0.7545,
360
+ "step": 500
361
+ },
362
+ {
363
+ "epoch": 0.19636153624025413,
364
+ "grad_norm": 0.22686438262462616,
365
+ "learning_rate": 9.294891723898926e-06,
366
+ "loss": 0.7445,
367
+ "step": 510
368
+ },
369
+ {
370
+ "epoch": 0.2002117624410434,
371
+ "grad_norm": 0.2486639767885208,
372
+ "learning_rate": 9.262630536456153e-06,
373
+ "loss": 0.7738,
374
+ "step": 520
375
+ },
376
+ {
377
+ "epoch": 0.2040619886418327,
378
+ "grad_norm": 0.24081294238567352,
379
+ "learning_rate": 9.229706346044749e-06,
380
+ "loss": 0.7486,
381
+ "step": 530
382
+ },
383
+ {
384
+ "epoch": 0.207912214842622,
385
+ "grad_norm": 0.2541293799877167,
386
+ "learning_rate": 9.196124273642422e-06,
387
+ "loss": 0.7598,
388
+ "step": 540
389
+ },
390
+ {
391
+ "epoch": 0.2117624410434113,
392
+ "grad_norm": 0.228404238820076,
393
+ "learning_rate": 9.161889542552842e-06,
394
+ "loss": 0.7684,
395
+ "step": 550
396
+ },
397
+ {
398
+ "epoch": 0.21561266724420058,
399
+ "grad_norm": 0.22831076383590698,
400
+ "learning_rate": 9.127007477593193e-06,
401
+ "loss": 0.7656,
402
+ "step": 560
403
+ },
404
+ {
405
+ "epoch": 0.2194628934449899,
406
+ "grad_norm": 0.253280371427536,
407
+ "learning_rate": 9.091483504265974e-06,
408
+ "loss": 0.7697,
409
+ "step": 570
410
+ },
411
+ {
412
+ "epoch": 0.22331311964577918,
413
+ "grad_norm": 0.24939945340156555,
414
+ "learning_rate": 9.055323147915116e-06,
415
+ "loss": 0.7546,
416
+ "step": 580
417
+ },
418
+ {
419
+ "epoch": 0.2271633458465685,
420
+ "grad_norm": 0.250679612159729,
421
+ "learning_rate": 9.018532032866585e-06,
422
+ "loss": 0.7607,
423
+ "step": 590
424
+ },
425
+ {
426
+ "epoch": 0.23101357204735778,
427
+ "grad_norm": 0.22366148233413696,
428
+ "learning_rate": 8.981115881553579e-06,
429
+ "loss": 0.7541,
430
+ "step": 600
431
+ },
432
+ {
433
+ "epoch": 0.23486379824814707,
434
+ "grad_norm": 0.257566899061203,
435
+ "learning_rate": 8.943080513626476e-06,
436
+ "loss": 0.773,
437
+ "step": 610
438
+ },
439
+ {
440
+ "epoch": 0.23871402444893638,
441
+ "grad_norm": 0.2287045121192932,
442
+ "learning_rate": 8.904431845047641e-06,
443
+ "loss": 0.7599,
444
+ "step": 620
445
+ },
446
+ {
447
+ "epoch": 0.24256425064972567,
448
+ "grad_norm": 0.23457437753677368,
449
+ "learning_rate": 8.86517588717128e-06,
450
+ "loss": 0.7751,
451
+ "step": 630
452
+ },
453
+ {
454
+ "epoch": 0.24641447685051496,
455
+ "grad_norm": 0.23213575780391693,
456
+ "learning_rate": 8.82531874580844e-06,
457
+ "loss": 0.7715,
458
+ "step": 640
459
+ },
460
+ {
461
+ "epoch": 0.25026470305130427,
462
+ "grad_norm": 0.22003386914730072,
463
+ "learning_rate": 8.784866620277314e-06,
464
+ "loss": 0.7256,
465
+ "step": 650
466
+ },
467
+ {
468
+ "epoch": 0.25411492925209356,
469
+ "grad_norm": 0.2504154443740845,
470
+ "learning_rate": 8.743825802439018e-06,
471
+ "loss": 0.7542,
472
+ "step": 660
473
+ },
474
+ {
475
+ "epoch": 0.25796515545288284,
476
+ "grad_norm": 0.26454678177833557,
477
+ "learning_rate": 8.702202675718956e-06,
478
+ "loss": 0.7673,
479
+ "step": 670
480
+ },
481
+ {
482
+ "epoch": 0.26181538165367213,
483
+ "grad_norm": 0.22092369198799133,
484
+ "learning_rate": 8.660003714113959e-06,
485
+ "loss": 0.7316,
486
+ "step": 680
487
+ },
488
+ {
489
+ "epoch": 0.26566560785446147,
490
+ "grad_norm": 0.2307596057653427,
491
+ "learning_rate": 8.617235481185323e-06,
492
+ "loss": 0.7556,
493
+ "step": 690
494
+ },
495
+ {
496
+ "epoch": 0.26951583405525076,
497
+ "grad_norm": 0.24423690140247345,
498
+ "learning_rate": 8.573904629037931e-06,
499
+ "loss": 0.7638,
500
+ "step": 700
501
+ },
502
+ {
503
+ "epoch": 0.27336606025604004,
504
+ "grad_norm": 0.24005594849586487,
505
+ "learning_rate": 8.530017897285585e-06,
506
+ "loss": 0.7461,
507
+ "step": 710
508
+ },
509
+ {
510
+ "epoch": 0.27721628645682933,
511
+ "grad_norm": 0.22120490670204163,
512
+ "learning_rate": 8.485582112002746e-06,
513
+ "loss": 0.753,
514
+ "step": 720
515
+ },
516
+ {
517
+ "epoch": 0.2810665126576186,
518
+ "grad_norm": 0.24022072553634644,
519
+ "learning_rate": 8.440604184662811e-06,
520
+ "loss": 0.7519,
521
+ "step": 730
522
+ },
523
+ {
524
+ "epoch": 0.28491673885840796,
525
+ "grad_norm": 0.22296322882175446,
526
+ "learning_rate": 8.395091111063118e-06,
527
+ "loss": 0.7401,
528
+ "step": 740
529
+ },
530
+ {
531
+ "epoch": 0.28876696505919724,
532
+ "grad_norm": 0.22271454334259033,
533
+ "learning_rate": 8.349049970236822e-06,
534
+ "loss": 0.7427,
535
+ "step": 750
536
+ },
537
+ {
538
+ "epoch": 0.29261719125998653,
539
+ "grad_norm": 0.2359963059425354,
540
+ "learning_rate": 8.302487923351842e-06,
541
+ "loss": 0.7334,
542
+ "step": 760
543
+ },
544
+ {
545
+ "epoch": 0.2964674174607758,
546
+ "grad_norm": 0.22603408992290497,
547
+ "learning_rate": 8.255412212597021e-06,
548
+ "loss": 0.7541,
549
+ "step": 770
550
+ },
551
+ {
552
+ "epoch": 0.3003176436615651,
553
+ "grad_norm": 0.22889547049999237,
554
+ "learning_rate": 8.207830160055688e-06,
555
+ "loss": 0.7508,
556
+ "step": 780
557
+ },
558
+ {
559
+ "epoch": 0.3041678698623544,
560
+ "grad_norm": 0.23885801434516907,
561
+ "learning_rate": 8.159749166566786e-06,
562
+ "loss": 0.7322,
563
+ "step": 790
564
+ },
565
+ {
566
+ "epoch": 0.30801809606314373,
567
+ "grad_norm": 0.22267626225948334,
568
+ "learning_rate": 8.11117671057377e-06,
569
+ "loss": 0.7569,
570
+ "step": 800
571
+ },
572
+ {
573
+ "epoch": 0.311868322263933,
574
+ "grad_norm": 0.24027813971042633,
575
+ "learning_rate": 8.062120346961404e-06,
576
+ "loss": 0.7483,
577
+ "step": 810
578
+ },
579
+ {
580
+ "epoch": 0.3157185484647223,
581
+ "grad_norm": 0.231545552611351,
582
+ "learning_rate": 8.012587705880705e-06,
583
+ "loss": 0.7425,
584
+ "step": 820
585
+ },
586
+ {
587
+ "epoch": 0.3195687746655116,
588
+ "grad_norm": 0.24665294587612152,
589
+ "learning_rate": 7.962586491562146e-06,
590
+ "loss": 0.7302,
591
+ "step": 830
592
+ },
593
+ {
594
+ "epoch": 0.3234190008663009,
595
+ "grad_norm": 0.2389557957649231,
596
+ "learning_rate": 7.912124481117357e-06,
597
+ "loss": 0.7358,
598
+ "step": 840
599
+ },
600
+ {
601
+ "epoch": 0.3272692270670902,
602
+ "grad_norm": 0.2335866242647171,
603
+ "learning_rate": 7.861209523329494e-06,
604
+ "loss": 0.7515,
605
+ "step": 850
606
+ },
607
+ {
608
+ "epoch": 0.3311194532678795,
609
+ "grad_norm": 0.2313600480556488,
610
+ "learning_rate": 7.809849537432432e-06,
611
+ "loss": 0.7293,
612
+ "step": 860
613
+ },
614
+ {
615
+ "epoch": 0.3349696794686688,
616
+ "grad_norm": 0.23034703731536865,
617
+ "learning_rate": 7.758052511879035e-06,
618
+ "loss": 0.7424,
619
+ "step": 870
620
+ },
621
+ {
622
+ "epoch": 0.3388199056694581,
623
+ "grad_norm": 0.22788743674755096,
624
+ "learning_rate": 7.705826503098642e-06,
625
+ "loss": 0.7349,
626
+ "step": 880
627
+ },
628
+ {
629
+ "epoch": 0.34267013187024736,
630
+ "grad_norm": 0.2472834587097168,
631
+ "learning_rate": 7.653179634243979e-06,
632
+ "loss": 0.7397,
633
+ "step": 890
634
+ },
635
+ {
636
+ "epoch": 0.34652035807103665,
637
+ "grad_norm": 0.21457000076770782,
638
+ "learning_rate": 7.600120093927696e-06,
639
+ "loss": 0.7383,
640
+ "step": 900
641
+ },
642
+ {
643
+ "epoch": 0.350370584271826,
644
+ "grad_norm": 0.24413040280342102,
645
+ "learning_rate": 7.546656134948729e-06,
646
+ "loss": 0.7287,
647
+ "step": 910
648
+ },
649
+ {
650
+ "epoch": 0.3542208104726153,
651
+ "grad_norm": 0.24662959575653076,
652
+ "learning_rate": 7.492796073008661e-06,
653
+ "loss": 0.7305,
654
+ "step": 920
655
+ },
656
+ {
657
+ "epoch": 0.35807103667340456,
658
+ "grad_norm": 0.24123650789260864,
659
+ "learning_rate": 7.438548285418326e-06,
660
+ "loss": 0.7159,
661
+ "step": 930
662
+ },
663
+ {
664
+ "epoch": 0.36192126287419385,
665
+ "grad_norm": 0.2669900953769684,
666
+ "learning_rate": 7.3839212097948e-06,
667
+ "loss": 0.7515,
668
+ "step": 940
669
+ },
670
+ {
671
+ "epoch": 0.36577148907498314,
672
+ "grad_norm": 0.22983017563819885,
673
+ "learning_rate": 7.328923342749039e-06,
674
+ "loss": 0.7694,
675
+ "step": 950
676
+ },
677
+ {
678
+ "epoch": 0.3696217152757725,
679
+ "grad_norm": 0.2500629127025604,
680
+ "learning_rate": 7.273563238564319e-06,
681
+ "loss": 0.751,
682
+ "step": 960
683
+ },
684
+ {
685
+ "epoch": 0.37347194147656176,
686
+ "grad_norm": 0.23460133373737335,
687
+ "learning_rate": 7.217849507865724e-06,
688
+ "loss": 0.7266,
689
+ "step": 970
690
+ },
691
+ {
692
+ "epoch": 0.37732216767735105,
693
+ "grad_norm": 0.23354275524616241,
694
+ "learning_rate": 7.161790816280858e-06,
695
+ "loss": 0.7392,
696
+ "step": 980
697
+ },
698
+ {
699
+ "epoch": 0.38117239387814034,
700
+ "grad_norm": 0.24138562381267548,
701
+ "learning_rate": 7.1053958830920125e-06,
702
+ "loss": 0.7152,
703
+ "step": 990
704
+ },
705
+ {
706
+ "epoch": 0.3850226200789296,
707
+ "grad_norm": 0.23430147767066956,
708
+ "learning_rate": 7.0486734798799715e-06,
709
+ "loss": 0.7392,
710
+ "step": 1000
711
+ },
712
+ {
713
+ "epoch": 0.3888728462797189,
714
+ "grad_norm": 0.2430313676595688,
715
+ "learning_rate": 6.991632429159708e-06,
716
+ "loss": 0.7635,
717
+ "step": 1010
718
+ },
719
+ {
720
+ "epoch": 0.39272307248050825,
721
+ "grad_norm": 0.2547365427017212,
722
+ "learning_rate": 6.934281603008132e-06,
723
+ "loss": 0.7487,
724
+ "step": 1020
725
+ },
726
+ {
727
+ "epoch": 0.39657329868129754,
728
+ "grad_norm": 0.23893508315086365,
729
+ "learning_rate": 6.876629921684143e-06,
730
+ "loss": 0.7483,
731
+ "step": 1030
732
+ },
733
+ {
734
+ "epoch": 0.4004235248820868,
735
+ "grad_norm": 0.23543035984039307,
736
+ "learning_rate": 6.818686352241199e-06,
737
+ "loss": 0.7369,
738
+ "step": 1040
739
+ },
740
+ {
741
+ "epoch": 0.4042737510828761,
742
+ "grad_norm": 0.2135132998228073,
743
+ "learning_rate": 6.760459907132573e-06,
744
+ "loss": 0.7337,
745
+ "step": 1050
746
+ },
747
+ {
748
+ "epoch": 0.4081239772836654,
749
+ "grad_norm": 0.22162741422653198,
750
+ "learning_rate": 6.701959642809589e-06,
751
+ "loss": 0.7385,
752
+ "step": 1060
753
+ },
754
+ {
755
+ "epoch": 0.41197420348445474,
756
+ "grad_norm": 0.23742549121379852,
757
+ "learning_rate": 6.643194658312987e-06,
758
+ "loss": 0.7297,
759
+ "step": 1070
760
+ },
761
+ {
762
+ "epoch": 0.415824429685244,
763
+ "grad_norm": 0.21888022124767303,
764
+ "learning_rate": 6.584174093857676e-06,
765
+ "loss": 0.757,
766
+ "step": 1080
767
+ },
768
+ {
769
+ "epoch": 0.4196746558860333,
770
+ "grad_norm": 0.2600443959236145,
771
+ "learning_rate": 6.5249071294110735e-06,
772
+ "loss": 0.7376,
773
+ "step": 1090
774
+ },
775
+ {
776
+ "epoch": 0.4235248820868226,
777
+ "grad_norm": 0.24497975409030914,
778
+ "learning_rate": 6.465402983265274e-06,
779
+ "loss": 0.7506,
780
+ "step": 1100
781
+ },
782
+ {
783
+ "epoch": 0.4273751082876119,
784
+ "grad_norm": 0.23106662929058075,
785
+ "learning_rate": 6.405670910603243e-06,
786
+ "loss": 0.7455,
787
+ "step": 1110
788
+ },
789
+ {
790
+ "epoch": 0.43122533448840117,
791
+ "grad_norm": 0.2482849657535553,
792
+ "learning_rate": 6.345720202059291e-06,
793
+ "loss": 0.742,
794
+ "step": 1120
795
+ },
796
+ {
797
+ "epoch": 0.4350755606891905,
798
+ "grad_norm": 0.24108658730983734,
799
+ "learning_rate": 6.285560182274008e-06,
800
+ "loss": 0.721,
801
+ "step": 1130
802
+ },
803
+ {
804
+ "epoch": 0.4389257868899798,
805
+ "grad_norm": 0.23850306868553162,
806
+ "learning_rate": 6.225200208443936e-06,
807
+ "loss": 0.7458,
808
+ "step": 1140
809
+ },
810
+ {
811
+ "epoch": 0.4427760130907691,
812
+ "grad_norm": 0.22096353769302368,
813
+ "learning_rate": 6.16464966886616e-06,
814
+ "loss": 0.7303,
815
+ "step": 1150
816
+ },
817
+ {
818
+ "epoch": 0.44662623929155837,
819
+ "grad_norm": 0.26680853962898254,
820
+ "learning_rate": 6.103917981478065e-06,
821
+ "loss": 0.7572,
822
+ "step": 1160
823
+ },
824
+ {
825
+ "epoch": 0.45047646549234766,
826
+ "grad_norm": 0.23085016012191772,
827
+ "learning_rate": 6.043014592392489e-06,
828
+ "loss": 0.7672,
829
+ "step": 1170
830
+ },
831
+ {
832
+ "epoch": 0.454326691693137,
833
+ "grad_norm": 0.23073603212833405,
834
+ "learning_rate": 5.981948974428485e-06,
835
+ "loss": 0.7434,
836
+ "step": 1180
837
+ },
838
+ {
839
+ "epoch": 0.4581769178939263,
840
+ "grad_norm": 0.24071775376796722,
841
+ "learning_rate": 5.920730625637934e-06,
842
+ "loss": 0.7182,
843
+ "step": 1190
844
+ },
845
+ {
846
+ "epoch": 0.46202714409471557,
847
+ "grad_norm": 0.24516922235488892,
848
+ "learning_rate": 5.8593690678282355e-06,
849
+ "loss": 0.7342,
850
+ "step": 1200
851
+ },
852
+ {
853
+ "epoch": 0.46587737029550486,
854
+ "grad_norm": 0.2363998293876648,
855
+ "learning_rate": 5.797873845081301e-06,
856
+ "loss": 0.7374,
857
+ "step": 1210
858
+ },
859
+ {
860
+ "epoch": 0.46972759649629414,
861
+ "grad_norm": 0.22996000945568085,
862
+ "learning_rate": 5.736254522269075e-06,
863
+ "loss": 0.7523,
864
+ "step": 1220
865
+ },
866
+ {
867
+ "epoch": 0.47357782269708343,
868
+ "grad_norm": 0.23353925347328186,
869
+ "learning_rate": 5.6745206835658386e-06,
870
+ "loss": 0.7561,
871
+ "step": 1230
872
+ },
873
+ {
874
+ "epoch": 0.47742804889787277,
875
+ "grad_norm": 0.22446739673614502,
876
+ "learning_rate": 5.612681930957495e-06,
877
+ "loss": 0.7267,
878
+ "step": 1240
879
+ },
880
+ {
881
+ "epoch": 0.48127827509866206,
882
+ "grad_norm": 0.22682546079158783,
883
+ "learning_rate": 5.550747882748095e-06,
884
+ "loss": 0.7223,
885
+ "step": 1250
886
+ },
887
+ {
888
+ "epoch": 0.48512850129945134,
889
+ "grad_norm": 0.25293636322021484,
890
+ "learning_rate": 5.4887281720638195e-06,
891
+ "loss": 0.7328,
892
+ "step": 1260
893
+ },
894
+ {
895
+ "epoch": 0.48897872750024063,
896
+ "grad_norm": 0.2580350637435913,
897
+ "learning_rate": 5.42663244535466e-06,
898
+ "loss": 0.7148,
899
+ "step": 1270
900
+ },
901
+ {
902
+ "epoch": 0.4928289537010299,
903
+ "grad_norm": 0.24116931855678558,
904
+ "learning_rate": 5.3644703608940205e-06,
905
+ "loss": 0.725,
906
+ "step": 1280
907
+ },
908
+ {
909
+ "epoch": 0.49667917990181926,
910
+ "grad_norm": 0.2382720410823822,
911
+ "learning_rate": 5.302251587276491e-06,
912
+ "loss": 0.7342,
913
+ "step": 1290
914
+ },
915
+ {
916
+ "epoch": 0.5005294061026085,
917
+ "grad_norm": 0.2204049676656723,
918
+ "learning_rate": 5.2399858019140005e-06,
919
+ "loss": 0.7348,
920
+ "step": 1300
921
+ },
922
+ {
923
+ "epoch": 0.5043796323033978,
924
+ "grad_norm": 0.23010540008544922,
925
+ "learning_rate": 5.177682689530615e-06,
926
+ "loss": 0.7547,
927
+ "step": 1310
928
+ },
929
+ {
930
+ "epoch": 0.5082298585041871,
931
+ "grad_norm": 0.23810261487960815,
932
+ "learning_rate": 5.11535194065619e-06,
933
+ "loss": 0.7161,
934
+ "step": 1320
935
+ },
936
+ {
937
+ "epoch": 0.5120800847049765,
938
+ "grad_norm": 0.24805285036563873,
939
+ "learning_rate": 5.053003250119113e-06,
940
+ "loss": 0.7252,
941
+ "step": 1330
942
+ },
943
+ {
944
+ "epoch": 0.5159303109057657,
945
+ "grad_norm": 0.23955266177654266,
946
+ "learning_rate": 4.9906463155383925e-06,
947
+ "loss": 0.7366,
948
+ "step": 1340
949
+ },
950
+ {
951
+ "epoch": 0.519780537106555,
952
+ "grad_norm": 0.22466902434825897,
953
+ "learning_rate": 4.9282908358153046e-06,
954
+ "loss": 0.7319,
955
+ "step": 1350
956
+ },
957
+ {
958
+ "epoch": 0.5236307633073443,
959
+ "grad_norm": 0.2528676986694336,
960
+ "learning_rate": 4.8659465096248345e-06,
961
+ "loss": 0.7429,
962
+ "step": 1360
963
+ },
964
+ {
965
+ "epoch": 0.5274809895081336,
966
+ "grad_norm": 0.2451244294643402,
967
+ "learning_rate": 4.803623033907169e-06,
968
+ "loss": 0.7391,
969
+ "step": 1370
970
+ },
971
+ {
972
+ "epoch": 0.5313312157089229,
973
+ "grad_norm": 0.2219330072402954,
974
+ "learning_rate": 4.741330102359441e-06,
975
+ "loss": 0.7496,
976
+ "step": 1380
977
+ },
978
+ {
979
+ "epoch": 0.5351814419097122,
980
+ "grad_norm": 0.25091353058815,
981
+ "learning_rate": 4.679077403927996e-06,
982
+ "loss": 0.7292,
983
+ "step": 1390
984
+ },
985
+ {
986
+ "epoch": 0.5390316681105015,
987
+ "grad_norm": 0.23269177973270416,
988
+ "learning_rate": 4.616874621301373e-06,
989
+ "loss": 0.7302,
990
+ "step": 1400
991
+ },
992
+ {
993
+ "epoch": 0.5428818943112907,
994
+ "grad_norm": 0.23282548785209656,
995
+ "learning_rate": 4.554731429404293e-06,
996
+ "loss": 0.7452,
997
+ "step": 1410
998
+ },
999
+ {
1000
+ "epoch": 0.5467321205120801,
1001
+ "grad_norm": 0.2501433491706848,
1002
+ "learning_rate": 4.492657493892822e-06,
1003
+ "loss": 0.7253,
1004
+ "step": 1420
1005
+ },
1006
+ {
1007
+ "epoch": 0.5505823467128694,
1008
+ "grad_norm": 0.26108691096305847,
1009
+ "learning_rate": 4.430662469650999e-06,
1010
+ "loss": 0.7421,
1011
+ "step": 1430
1012
+ },
1013
+ {
1014
+ "epoch": 0.5544325729136587,
1015
+ "grad_norm": 0.25390198826789856,
1016
+ "learning_rate": 4.368755999289118e-06,
1017
+ "loss": 0.7492,
1018
+ "step": 1440
1019
+ },
1020
+ {
1021
+ "epoch": 0.558282799114448,
1022
+ "grad_norm": 0.24770860373973846,
1023
+ "learning_rate": 4.306947711643949e-06,
1024
+ "loss": 0.73,
1025
+ "step": 1450
1026
+ },
1027
+ {
1028
+ "epoch": 0.5621330253152372,
1029
+ "grad_norm": 0.24091008305549622,
1030
+ "learning_rate": 4.245247220281068e-06,
1031
+ "loss": 0.7336,
1032
+ "step": 1460
1033
+ },
1034
+ {
1035
+ "epoch": 0.5659832515160266,
1036
+ "grad_norm": 0.22514736652374268,
1037
+ "learning_rate": 4.183664121999587e-06,
1038
+ "loss": 0.7275,
1039
+ "step": 1470
1040
+ },
1041
+ {
1042
+ "epoch": 0.5698334777168159,
1043
+ "grad_norm": 0.23206207156181335,
1044
+ "learning_rate": 4.122207995339474e-06,
1045
+ "loss": 0.7288,
1046
+ "step": 1480
1047
+ },
1048
+ {
1049
+ "epoch": 0.5736837039176051,
1050
+ "grad_norm": 0.22928830981254578,
1051
+ "learning_rate": 4.060888399091734e-06,
1052
+ "loss": 0.7309,
1053
+ "step": 1490
1054
+ },
1055
+ {
1056
+ "epoch": 0.5775339301183945,
1057
+ "grad_norm": 0.2306356132030487,
1058
+ "learning_rate": 3.999714870811639e-06,
1059
+ "loss": 0.7439,
1060
+ "step": 1500
1061
+ },
1062
+ {
1063
+ "epoch": 0.5813841563191837,
1064
+ "grad_norm": 0.23400349915027618,
1065
+ "learning_rate": 3.938696925335275e-06,
1066
+ "loss": 0.7361,
1067
+ "step": 1510
1068
+ },
1069
+ {
1070
+ "epoch": 0.5852343825199731,
1071
+ "grad_norm": 0.23420268297195435,
1072
+ "learning_rate": 3.87784405329962e-06,
1073
+ "loss": 0.7427,
1074
+ "step": 1520
1075
+ },
1076
+ {
1077
+ "epoch": 0.5890846087207623,
1078
+ "grad_norm": 0.23095855116844177,
1079
+ "learning_rate": 3.817165719666382e-06,
1080
+ "loss": 0.743,
1081
+ "step": 1530
1082
+ },
1083
+ {
1084
+ "epoch": 0.5929348349215516,
1085
+ "grad_norm": 0.22891034185886383,
1086
+ "learning_rate": 3.756671362249841e-06,
1087
+ "loss": 0.7522,
1088
+ "step": 1540
1089
+ },
1090
+ {
1091
+ "epoch": 0.596785061122341,
1092
+ "grad_norm": 0.23149672150611877,
1093
+ "learning_rate": 3.6963703902488907e-06,
1094
+ "loss": 0.7355,
1095
+ "step": 1550
1096
+ },
1097
+ {
1098
+ "epoch": 0.6006352873231302,
1099
+ "grad_norm": 0.22326141595840454,
1100
+ "learning_rate": 3.6362721827835655e-06,
1101
+ "loss": 0.7469,
1102
+ "step": 1560
1103
+ },
1104
+ {
1105
+ "epoch": 0.6044855135239195,
1106
+ "grad_norm": 0.24373599886894226,
1107
+ "learning_rate": 3.576386087436212e-06,
1108
+ "loss": 0.7469,
1109
+ "step": 1570
1110
+ },
1111
+ {
1112
+ "epoch": 0.6083357397247088,
1113
+ "grad_norm": 0.2461036741733551,
1114
+ "learning_rate": 3.5167214187975913e-06,
1115
+ "loss": 0.7362,
1116
+ "step": 1580
1117
+ },
1118
+ {
1119
+ "epoch": 0.6121859659254981,
1120
+ "grad_norm": 0.21031983196735382,
1121
+ "learning_rate": 3.457287457018086e-06,
1122
+ "loss": 0.739,
1123
+ "step": 1590
1124
+ },
1125
+ {
1126
+ "epoch": 0.6160361921262875,
1127
+ "grad_norm": 0.2215377688407898,
1128
+ "learning_rate": 3.3980934463643003e-06,
1129
+ "loss": 0.7287,
1130
+ "step": 1600
1131
+ },
1132
+ {
1133
+ "epoch": 0.6198864183270767,
1134
+ "grad_norm": 0.22038224339485168,
1135
+ "learning_rate": 3.3391485937812053e-06,
1136
+ "loss": 0.7091,
1137
+ "step": 1610
1138
+ },
1139
+ {
1140
+ "epoch": 0.623736644527866,
1141
+ "grad_norm": 0.2480737417936325,
1142
+ "learning_rate": 3.280462067460115e-06,
1143
+ "loss": 0.7092,
1144
+ "step": 1620
1145
+ },
1146
+ {
1147
+ "epoch": 0.6275868707286553,
1148
+ "grad_norm": 0.23266732692718506,
1149
+ "learning_rate": 3.222042995412669e-06,
1150
+ "loss": 0.7399,
1151
+ "step": 1630
1152
+ },
1153
+ {
1154
+ "epoch": 0.6314370969294446,
1155
+ "grad_norm": 0.21849381923675537,
1156
+ "learning_rate": 3.163900464051089e-06,
1157
+ "loss": 0.7273,
1158
+ "step": 1640
1159
+ },
1160
+ {
1161
+ "epoch": 0.635287323130234,
1162
+ "grad_norm": 0.23824460804462433,
1163
+ "learning_rate": 3.1060435167748826e-06,
1164
+ "loss": 0.7387,
1165
+ "step": 1650
1166
+ },
1167
+ {
1168
+ "epoch": 0.6391375493310232,
1169
+ "grad_norm": 0.22563143074512482,
1170
+ "learning_rate": 3.048481152564253e-06,
1171
+ "loss": 0.7201,
1172
+ "step": 1660
1173
+ },
1174
+ {
1175
+ "epoch": 0.6429877755318125,
1176
+ "grad_norm": 0.24190707504749298,
1177
+ "learning_rate": 2.9912223245804017e-06,
1178
+ "loss": 0.745,
1179
+ "step": 1670
1180
+ },
1181
+ {
1182
+ "epoch": 0.6468380017326018,
1183
+ "grad_norm": 0.2444942742586136,
1184
+ "learning_rate": 2.934275938772981e-06,
1185
+ "loss": 0.7348,
1186
+ "step": 1680
1187
+ },
1188
+ {
1189
+ "epoch": 0.6506882279333911,
1190
+ "grad_norm": 0.2451329082250595,
1191
+ "learning_rate": 2.8776508524948676e-06,
1192
+ "loss": 0.7264,
1193
+ "step": 1690
1194
+ },
1195
+ {
1196
+ "epoch": 0.6545384541341804,
1197
+ "grad_norm": 0.21279406547546387,
1198
+ "learning_rate": 2.821355873124503e-06,
1199
+ "loss": 0.7226,
1200
+ "step": 1700
1201
+ },
1202
+ {
1203
+ "epoch": 0.6583886803349697,
1204
+ "grad_norm": 0.24221539497375488,
1205
+ "learning_rate": 2.7653997566960144e-06,
1206
+ "loss": 0.7413,
1207
+ "step": 1710
1208
+ },
1209
+ {
1210
+ "epoch": 0.662238906535759,
1211
+ "grad_norm": 0.24025501310825348,
1212
+ "learning_rate": 2.709791206537309e-06,
1213
+ "loss": 0.7311,
1214
+ "step": 1720
1215
+ },
1216
+ {
1217
+ "epoch": 0.6660891327365482,
1218
+ "grad_norm": 0.2363608032464981,
1219
+ "learning_rate": 2.6545388719163747e-06,
1220
+ "loss": 0.7297,
1221
+ "step": 1730
1222
+ },
1223
+ {
1224
+ "epoch": 0.6699393589373376,
1225
+ "grad_norm": 0.22837482392787933,
1226
+ "learning_rate": 2.599651346695979e-06,
1227
+ "loss": 0.7323,
1228
+ "step": 1740
1229
+ },
1230
+ {
1231
+ "epoch": 0.6737895851381268,
1232
+ "grad_norm": 0.23686936497688293,
1233
+ "learning_rate": 2.545137167997007e-06,
1234
+ "loss": 0.7312,
1235
+ "step": 1750
1236
+ },
1237
+ {
1238
+ "epoch": 0.6776398113389162,
1239
+ "grad_norm": 0.2284570187330246,
1240
+ "learning_rate": 2.4910048148705972e-06,
1241
+ "loss": 0.7145,
1242
+ "step": 1760
1243
+ },
1244
+ {
1245
+ "epoch": 0.6814900375397055,
1246
+ "grad_norm": 0.25094443559646606,
1247
+ "learning_rate": 2.4372627069793327e-06,
1248
+ "loss": 0.7624,
1249
+ "step": 1770
1250
+ },
1251
+ {
1252
+ "epoch": 0.6853402637404947,
1253
+ "grad_norm": 0.23770809173583984,
1254
+ "learning_rate": 2.3839192032876486e-06,
1255
+ "loss": 0.7461,
1256
+ "step": 1780
1257
+ },
1258
+ {
1259
+ "epoch": 0.6891904899412841,
1260
+ "grad_norm": 0.2248828262090683,
1261
+ "learning_rate": 2.3309826007617047e-06,
1262
+ "loss": 0.7573,
1263
+ "step": 1790
1264
+ },
1265
+ {
1266
+ "epoch": 0.6930407161420733,
1267
+ "grad_norm": 0.24861162900924683,
1268
+ "learning_rate": 2.278461133078881e-06,
1269
+ "loss": 0.7217,
1270
+ "step": 1800
1271
+ },
1272
+ {
1273
+ "epoch": 0.6968909423428626,
1274
+ "grad_norm": 0.23656992614269257,
1275
+ "learning_rate": 2.2263629693471258e-06,
1276
+ "loss": 0.7117,
1277
+ "step": 1810
1278
+ },
1279
+ {
1280
+ "epoch": 0.700741168543652,
1281
+ "grad_norm": 0.2230072170495987,
1282
+ "learning_rate": 2.1746962128343414e-06,
1283
+ "loss": 0.7348,
1284
+ "step": 1820
1285
+ },
1286
+ {
1287
+ "epoch": 0.7045913947444412,
1288
+ "grad_norm": 0.22149774432182312,
1289
+ "learning_rate": 2.1234688997080183e-06,
1290
+ "loss": 0.7261,
1291
+ "step": 1830
1292
+ },
1293
+ {
1294
+ "epoch": 0.7084416209452306,
1295
+ "grad_norm": 0.24225665628910065,
1296
+ "learning_rate": 2.0726889977853053e-06,
1297
+ "loss": 0.729,
1298
+ "step": 1840
1299
+ },
1300
+ {
1301
+ "epoch": 0.7122918471460198,
1302
+ "grad_norm": 0.24979867041110992,
1303
+ "learning_rate": 2.022364405293703e-06,
1304
+ "loss": 0.7447,
1305
+ "step": 1850
1306
+ },
1307
+ {
1308
+ "epoch": 0.7161420733468091,
1309
+ "grad_norm": 0.22596682608127594,
1310
+ "learning_rate": 1.9725029496425814e-06,
1311
+ "loss": 0.7351,
1312
+ "step": 1860
1313
+ },
1314
+ {
1315
+ "epoch": 0.7199922995475985,
1316
+ "grad_norm": 0.24373310804367065,
1317
+ "learning_rate": 1.923112386205727e-06,
1318
+ "loss": 0.7328,
1319
+ "step": 1870
1320
+ },
1321
+ {
1322
+ "epoch": 0.7238425257483877,
1323
+ "grad_norm": 0.22788499295711517,
1324
+ "learning_rate": 1.874200397115079e-06,
1325
+ "loss": 0.7406,
1326
+ "step": 1880
1327
+ },
1328
+ {
1329
+ "epoch": 0.727692751949177,
1330
+ "grad_norm": 0.23872101306915283,
1331
+ "learning_rate": 1.825774590065857e-06,
1332
+ "loss": 0.724,
1333
+ "step": 1890
1334
+ },
1335
+ {
1336
+ "epoch": 0.7315429781499663,
1337
+ "grad_norm": 0.23401899635791779,
1338
+ "learning_rate": 1.7778424971332813e-06,
1339
+ "loss": 0.7311,
1340
+ "step": 1900
1341
+ },
1342
+ {
1343
+ "epoch": 0.7353932043507556,
1344
+ "grad_norm": 0.22593490779399872,
1345
+ "learning_rate": 1.7304115736010407e-06,
1346
+ "loss": 0.7284,
1347
+ "step": 1910
1348
+ },
1349
+ {
1350
+ "epoch": 0.739243430551545,
1351
+ "grad_norm": 0.23915517330169678,
1352
+ "learning_rate": 1.6834891968017108e-06,
1353
+ "loss": 0.7232,
1354
+ "step": 1920
1355
+ },
1356
+ {
1357
+ "epoch": 0.7430936567523342,
1358
+ "grad_norm": 0.24575449526309967,
1359
+ "learning_rate": 1.6370826649692884e-06,
1360
+ "loss": 0.736,
1361
+ "step": 1930
1362
+ },
1363
+ {
1364
+ "epoch": 0.7469438829531235,
1365
+ "grad_norm": 0.23716682195663452,
1366
+ "learning_rate": 1.5911991961040463e-06,
1367
+ "loss": 0.7506,
1368
+ "step": 1940
1369
+ },
1370
+ {
1371
+ "epoch": 0.7507941091539128,
1372
+ "grad_norm": 0.23953062295913696,
1373
+ "learning_rate": 1.545845926849851e-06,
1374
+ "loss": 0.7382,
1375
+ "step": 1950
1376
+ },
1377
+ {
1378
+ "epoch": 0.7546443353547021,
1379
+ "grad_norm": 0.24875980615615845,
1380
+ "learning_rate": 1.5010299113841397e-06,
1381
+ "loss": 0.7539,
1382
+ "step": 1960
1383
+ },
1384
+ {
1385
+ "epoch": 0.7584945615554913,
1386
+ "grad_norm": 0.23102392256259918,
1387
+ "learning_rate": 1.4567581203207238e-06,
1388
+ "loss": 0.7429,
1389
+ "step": 1970
1390
+ },
1391
+ {
1392
+ "epoch": 0.7623447877562807,
1393
+ "grad_norm": 0.2493913620710373,
1394
+ "learning_rate": 1.413037439625593e-06,
1395
+ "loss": 0.7394,
1396
+ "step": 1980
1397
+ },
1398
+ {
1399
+ "epoch": 0.76619501395707,
1400
+ "grad_norm": 0.2564362585544586,
1401
+ "learning_rate": 1.3698746695458847e-06,
1402
+ "loss": 0.7371,
1403
+ "step": 1990
1404
+ },
1405
+ {
1406
+ "epoch": 0.7700452401578592,
1407
+ "grad_norm": 0.23687854409217834,
1408
+ "learning_rate": 1.3272765235521745e-06,
1409
+ "loss": 0.7281,
1410
+ "step": 2000
1411
+ },
1412
+ {
1413
+ "epoch": 0.7738954663586486,
1414
+ "grad_norm": 0.21927976608276367,
1415
+ "learning_rate": 1.2852496272942877e-06,
1416
+ "loss": 0.7369,
1417
+ "step": 2010
1418
+ },
1419
+ {
1420
+ "epoch": 0.7777456925594378,
1421
+ "grad_norm": 0.2209339588880539,
1422
+ "learning_rate": 1.2438005175707396e-06,
1423
+ "loss": 0.7199,
1424
+ "step": 2020
1425
+ },
1426
+ {
1427
+ "epoch": 0.7815959187602272,
1428
+ "grad_norm": 0.2315627485513687,
1429
+ "learning_rate": 1.2029356413120297e-06,
1430
+ "loss": 0.7061,
1431
+ "step": 2030
1432
+ },
1433
+ {
1434
+ "epoch": 0.7854461449610165,
1435
+ "grad_norm": 0.2243843972682953,
1436
+ "learning_rate": 1.1626613545778786e-06,
1437
+ "loss": 0.7237,
1438
+ "step": 2040
1439
+ },
1440
+ {
1441
+ "epoch": 0.7892963711618057,
1442
+ "grad_norm": 0.2291395366191864,
1443
+ "learning_rate": 1.1229839215686294e-06,
1444
+ "loss": 0.7286,
1445
+ "step": 2050
1446
+ },
1447
+ {
1448
+ "epoch": 0.7931465973625951,
1449
+ "grad_norm": 0.2181369960308075,
1450
+ "learning_rate": 1.0839095136509188e-06,
1451
+ "loss": 0.7112,
1452
+ "step": 2060
1453
+ },
1454
+ {
1455
+ "epoch": 0.7969968235633843,
1456
+ "grad_norm": 0.2442452758550644,
1457
+ "learning_rate": 1.045444208397791e-06,
1458
+ "loss": 0.7321,
1459
+ "step": 2070
1460
+ },
1461
+ {
1462
+ "epoch": 0.8008470497641736,
1463
+ "grad_norm": 0.22868116199970245,
1464
+ "learning_rate": 1.007593988643402e-06,
1465
+ "loss": 0.715,
1466
+ "step": 2080
1467
+ },
1468
+ {
1469
+ "epoch": 0.804697275964963,
1470
+ "grad_norm": 0.223911851644516,
1471
+ "learning_rate": 9.703647415524613e-07,
1472
+ "loss": 0.7388,
1473
+ "step": 2090
1474
+ },
1475
+ {
1476
+ "epoch": 0.8085475021657522,
1477
+ "grad_norm": 0.22634689509868622,
1478
+ "learning_rate": 9.337622577045485e-07,
1479
+ "loss": 0.7248,
1480
+ "step": 2100
1481
+ },
1482
+ {
1483
+ "epoch": 0.8123977283665416,
1484
+ "grad_norm": 0.22778797149658203,
1485
+ "learning_rate": 8.977922301934588e-07,
1486
+ "loss": 0.7393,
1487
+ "step": 2110
1488
+ },
1489
+ {
1490
+ "epoch": 0.8162479545673308,
1491
+ "grad_norm": 0.2325425148010254,
1492
+ "learning_rate": 8.624602537417009e-07,
1493
+ "loss": 0.7418,
1494
+ "step": 2120
1495
+ },
1496
+ {
1497
+ "epoch": 0.8200981807681201,
1498
+ "grad_norm": 0.24321742355823517,
1499
+ "learning_rate": 8.277718238303123e-07,
1500
+ "loss": 0.727,
1501
+ "step": 2130
1502
+ },
1503
+ {
1504
+ "epoch": 0.8239484069689095,
1505
+ "grad_norm": 0.232570618391037,
1506
+ "learning_rate": 7.937323358440935e-07,
1507
+ "loss": 0.7303,
1508
+ "step": 2140
1509
+ },
1510
+ {
1511
+ "epoch": 0.8277986331696987,
1512
+ "grad_norm": 0.2506219744682312,
1513
+ "learning_rate": 7.603470842324223e-07,
1514
+ "loss": 0.7217,
1515
+ "step": 2150
1516
+ },
1517
+ {
1518
+ "epoch": 0.831648859370488,
1519
+ "grad_norm": 0.22258798778057098,
1520
+ "learning_rate": 7.276212616857575e-07,
1521
+ "loss": 0.7395,
1522
+ "step": 2160
1523
+ },
1524
+ {
1525
+ "epoch": 0.8354990855712773,
1526
+ "grad_norm": 0.22617006301879883,
1527
+ "learning_rate": 6.955599583279837e-07,
1528
+ "loss": 0.7366,
1529
+ "step": 2170
1530
+ },
1531
+ {
1532
+ "epoch": 0.8393493117720666,
1533
+ "grad_norm": 0.23766928911209106,
1534
+ "learning_rate": 6.641681609246981e-07,
1535
+ "loss": 0.7317,
1536
+ "step": 2180
1537
+ },
1538
+ {
1539
+ "epoch": 0.8431995379728559,
1540
+ "grad_norm": 0.22384490072727203,
1541
+ "learning_rate": 6.334507521075711e-07,
1542
+ "loss": 0.7338,
1543
+ "step": 2190
1544
+ },
1545
+ {
1546
+ "epoch": 0.8470497641736452,
1547
+ "grad_norm": 0.247166246175766,
1548
+ "learning_rate": 6.034125096149179e-07,
1549
+ "loss": 0.7323,
1550
+ "step": 2200
1551
+ },
1552
+ {
1553
+ "epoch": 0.8508999903744345,
1554
+ "grad_norm": 0.23910857737064362,
1555
+ "learning_rate": 5.740581055485722e-07,
1556
+ "loss": 0.7036,
1557
+ "step": 2210
1558
+ },
1559
+ {
1560
+ "epoch": 0.8547502165752238,
1561
+ "grad_norm": 0.21904686093330383,
1562
+ "learning_rate": 5.453921056471961e-07,
1563
+ "loss": 0.7366,
1564
+ "step": 2220
1565
+ },
1566
+ {
1567
+ "epoch": 0.8586004427760131,
1568
+ "grad_norm": 0.23360924422740936,
1569
+ "learning_rate": 5.174189685761288e-07,
1570
+ "loss": 0.7312,
1571
+ "step": 2230
1572
+ },
1573
+ {
1574
+ "epoch": 0.8624506689768023,
1575
+ "grad_norm": 0.2273528128862381,
1576
+ "learning_rate": 4.901430452338979e-07,
1577
+ "loss": 0.7447,
1578
+ "step": 2240
1579
+ },
1580
+ {
1581
+ "epoch": 0.8663008951775917,
1582
+ "grad_norm": 0.22678826749324799,
1583
+ "learning_rate": 4.6356857807548286e-07,
1584
+ "loss": 0.7242,
1585
+ "step": 2250
1586
+ },
1587
+ {
1588
+ "epoch": 0.870151121378381,
1589
+ "grad_norm": 0.2365577667951584,
1590
+ "learning_rate": 4.376997004524541e-07,
1591
+ "loss": 0.7146,
1592
+ "step": 2260
1593
+ },
1594
+ {
1595
+ "epoch": 0.8740013475791703,
1596
+ "grad_norm": 0.23097039759159088,
1597
+ "learning_rate": 4.12540435970073e-07,
1598
+ "loss": 0.7208,
1599
+ "step": 2270
1600
+ },
1601
+ {
1602
+ "epoch": 0.8778515737799596,
1603
+ "grad_norm": 0.2248639017343521,
1604
+ "learning_rate": 3.880946978614703e-07,
1605
+ "loss": 0.7403,
1606
+ "step": 2280
1607
+ },
1608
+ {
1609
+ "epoch": 0.8817017999807488,
1610
+ "grad_norm": 0.22823628783226013,
1611
+ "learning_rate": 3.643662883789878e-07,
1612
+ "loss": 0.7355,
1613
+ "step": 2290
1614
+ },
1615
+ {
1616
+ "epoch": 0.8855520261815382,
1617
+ "grad_norm": 0.21748699247837067,
1618
+ "learning_rate": 3.41358898202781e-07,
1619
+ "loss": 0.7154,
1620
+ "step": 2300
1621
+ },
1622
+ {
1623
+ "epoch": 0.8894022523823275,
1624
+ "grad_norm": 0.2191355675458908,
1625
+ "learning_rate": 3.190761058667741e-07,
1626
+ "loss": 0.7271,
1627
+ "step": 2310
1628
+ },
1629
+ {
1630
+ "epoch": 0.8932524785831167,
1631
+ "grad_norm": 0.2308504581451416,
1632
+ "learning_rate": 2.9752137720206463e-07,
1633
+ "loss": 0.7312,
1634
+ "step": 2320
1635
+ },
1636
+ {
1637
+ "epoch": 0.8971027047839061,
1638
+ "grad_norm": 0.22700758278369904,
1639
+ "learning_rate": 2.7669806479785067e-07,
1640
+ "loss": 0.7387,
1641
+ "step": 2330
1642
+ },
1643
+ {
1644
+ "epoch": 0.9009529309846953,
1645
+ "grad_norm": 0.23735877871513367,
1646
+ "learning_rate": 2.566094074799763e-07,
1647
+ "loss": 0.7298,
1648
+ "step": 2340
1649
+ },
1650
+ {
1651
+ "epoch": 0.9048031571854847,
1652
+ "grad_norm": 0.22306914627552032,
1653
+ "learning_rate": 2.3725852980716557e-07,
1654
+ "loss": 0.716,
1655
+ "step": 2350
1656
+ },
1657
+ {
1658
+ "epoch": 0.908653383386274,
1659
+ "grad_norm": 0.2244417369365692,
1660
+ "learning_rate": 2.1864844158503983e-07,
1661
+ "loss": 0.7251,
1662
+ "step": 2360
1663
+ },
1664
+ {
1665
+ "epoch": 0.9125036095870632,
1666
+ "grad_norm": 0.2512446939945221,
1667
+ "learning_rate": 2.007820373979702e-07,
1668
+ "loss": 0.7176,
1669
+ "step": 2370
1670
+ },
1671
+ {
1672
+ "epoch": 0.9163538357878526,
1673
+ "grad_norm": 0.2312445342540741,
1674
+ "learning_rate": 1.8366209615886532e-07,
1675
+ "loss": 0.7283,
1676
+ "step": 2380
1677
+ },
1678
+ {
1679
+ "epoch": 0.9202040619886418,
1680
+ "grad_norm": 0.22906425595283508,
1681
+ "learning_rate": 1.6729128067693613e-07,
1682
+ "loss": 0.7432,
1683
+ "step": 2390
1684
+ },
1685
+ {
1686
+ "epoch": 0.9240542881894311,
1687
+ "grad_norm": 0.24329553544521332,
1688
+ "learning_rate": 1.5167213724353426e-07,
1689
+ "loss": 0.7421,
1690
+ "step": 2400
1691
+ },
1692
+ {
1693
+ "epoch": 0.9279045143902205,
1694
+ "grad_norm": 0.22913627326488495,
1695
+ "learning_rate": 1.3680709523610058e-07,
1696
+ "loss": 0.7231,
1697
+ "step": 2410
1698
+ },
1699
+ {
1700
+ "epoch": 0.9317547405910097,
1701
+ "grad_norm": 0.22058475017547607,
1702
+ "learning_rate": 1.2269846674030418e-07,
1703
+ "loss": 0.7433,
1704
+ "step": 2420
1705
+ },
1706
+ {
1707
+ "epoch": 0.935604966791799,
1708
+ "grad_norm": 0.23765850067138672,
1709
+ "learning_rate": 1.0934844619042595e-07,
1710
+ "loss": 0.7432,
1711
+ "step": 2430
1712
+ },
1713
+ {
1714
+ "epoch": 0.9394551929925883,
1715
+ "grad_norm": 0.23876994848251343,
1716
+ "learning_rate": 9.675911002803961e-08,
1717
+ "loss": 0.728,
1718
+ "step": 2440
1719
+ },
1720
+ {
1721
+ "epoch": 0.9433054191933776,
1722
+ "grad_norm": 0.24049784243106842,
1723
+ "learning_rate": 8.493241637904381e-08,
1724
+ "loss": 0.7244,
1725
+ "step": 2450
1726
+ },
1727
+ {
1728
+ "epoch": 0.9471556453941669,
1729
+ "grad_norm": 0.2411442995071411,
1730
+ "learning_rate": 7.387020474909689e-08,
1731
+ "loss": 0.7478,
1732
+ "step": 2460
1733
+ },
1734
+ {
1735
+ "epoch": 0.9510058715949562,
1736
+ "grad_norm": 0.2427770048379898,
1737
+ "learning_rate": 6.357419573750856e-08,
1738
+ "loss": 0.742,
1739
+ "step": 2470
1740
+ },
1741
+ {
1742
+ "epoch": 0.9548560977957455,
1743
+ "grad_norm": 0.24098002910614014,
1744
+ "learning_rate": 5.4045990769613365e-08,
1745
+ "loss": 0.7246,
1746
+ "step": 2480
1747
+ },
1748
+ {
1749
+ "epoch": 0.9587063239965348,
1750
+ "grad_norm": 0.27965712547302246,
1751
+ "learning_rate": 4.5287071847694434e-08,
1752
+ "loss": 0.7282,
1753
+ "step": 2490
1754
+ },
1755
+ {
1756
+ "epoch": 0.9625565501973241,
1757
+ "grad_norm": 0.22558003664016724,
1758
+ "learning_rate": 3.7298801320467817e-08,
1759
+ "loss": 0.7256,
1760
+ "step": 2500
1761
+ },
1762
+ {
1763
+ "epoch": 0.9664067763981133,
1764
+ "grad_norm": 0.23487718403339386,
1765
+ "learning_rate": 3.0082421671192576e-08,
1766
+ "loss": 0.7194,
1767
+ "step": 2510
1768
+ },
1769
+ {
1770
+ "epoch": 0.9702570025989027,
1771
+ "grad_norm": 0.23993195593357086,
1772
+ "learning_rate": 2.3639055324410355e-08,
1773
+ "loss": 0.7434,
1774
+ "step": 2520
1775
+ },
1776
+ {
1777
+ "epoch": 0.974107228799692,
1778
+ "grad_norm": 0.22560152411460876,
1779
+ "learning_rate": 1.7969704471369477e-08,
1780
+ "loss": 0.7227,
1781
+ "step": 2530
1782
+ },
1783
+ {
1784
+ "epoch": 0.9779574550004813,
1785
+ "grad_norm": 0.21987169981002808,
1786
+ "learning_rate": 1.3075250914142413e-08,
1787
+ "loss": 0.7293,
1788
+ "step": 2540
1789
+ },
1790
+ {
1791
+ "epoch": 0.9818076812012706,
1792
+ "grad_norm": 0.2212374359369278,
1793
+ "learning_rate": 8.956455928476049e-09,
1794
+ "loss": 0.7351,
1795
+ "step": 2550
1796
+ },
1797
+ {
1798
+ "epoch": 0.9856579074020598,
1799
+ "grad_norm": 0.2350045144557953,
1800
+ "learning_rate": 5.6139601453786275e-09,
1801
+ "loss": 0.726,
1802
+ "step": 2560
1803
+ },
1804
+ {
1805
+ "epoch": 0.9895081336028492,
1806
+ "grad_norm": 0.22632461786270142,
1807
+ "learning_rate": 3.0482834514805602e-09,
1808
+ "loss": 0.7169,
1809
+ "step": 2570
1810
+ },
1811
+ {
1812
+ "epoch": 0.9933583598036385,
1813
+ "grad_norm": 0.20718833804130554,
1814
+ "learning_rate": 1.2598249081713366e-09,
1815
+ "loss": 0.7073,
1816
+ "step": 2580
1817
+ },
1818
+ {
1819
+ "epoch": 0.9972085860044277,
1820
+ "grad_norm": 0.24415473639965057,
1821
+ "learning_rate": 2.488626895302826e-10,
1822
+ "loss": 0.7449,
1823
+ "step": 2590
1824
+ },
1825
+ {
1826
+ "epoch": 0.9999037443449803,
1827
+ "step": 2597,
1828
+ "total_flos": 4.633853246585176e+18,
1829
+ "train_loss": 0.7616396600299493,
1830
+ "train_runtime": 10817.1045,
1831
+ "train_samples_per_second": 61.468,
1832
+ "train_steps_per_second": 0.24
1833
+ }
1834
+ ],
1835
+ "logging_steps": 10,
1836
+ "max_steps": 2597,
1837
+ "num_input_tokens_seen": 0,
1838
+ "num_train_epochs": 1,
1839
+ "save_steps": 500,
1840
+ "stateful_callbacks": {
1841
+ "TrainerControl": {
1842
+ "args": {
1843
+ "should_epoch_stop": false,
1844
+ "should_evaluate": false,
1845
+ "should_log": false,
1846
+ "should_save": true,
1847
+ "should_training_stop": true
1848
+ },
1849
+ "attributes": {}
1850
+ }
1851
+ },
1852
+ "total_flos": 4.633853246585176e+18,
1853
+ "train_batch_size": 8,
1854
+ "trial_name": null,
1855
+ "trial_params": null
1856
+ }
continuous/qwen3/siglip2_sm16_256/finetune/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5859966ecd858ce3af8ccc3a6449d86c067de547185d82aaca2af1c255157b49
3
+ size 6136
continuous/qwen3/siglip2_sm16_256/finetune/vocab.json ADDED
The diff for this file is too large to render. See raw diff
 
continuous/qwen3/siglip2_sm16_256/pretrain/config.json ADDED
@@ -0,0 +1,60 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_attn_implementation_autoset": true,
3
+ "add_faster_video": false,
4
+ "add_time_instruction": false,
5
+ "architectures": [
6
+ "Qwen3ForCausalLM"
7
+ ],
8
+ "attention_bias": false,
9
+ "attention_dropout": 0.0,
10
+ "bos_token_id": 151643,
11
+ "eos_token_id": 151645,
12
+ "faster_token_stride": 10,
13
+ "force_sample": false,
14
+ "head_dim": 128,
15
+ "hidden_act": "silu",
16
+ "hidden_size": 2048,
17
+ "image_aspect_ratio": "square",
18
+ "image_crop_resolution": null,
19
+ "image_grid_pinpoints": null,
20
+ "image_split_resolution": null,
21
+ "initializer_range": 0.02,
22
+ "intermediate_size": 6144,
23
+ "max_position_embeddings": 40960,
24
+ "max_window_layers": 28,
25
+ "mm_hidden_size": 1152,
26
+ "mm_newline_position": "grid",
27
+ "mm_patch_merge_type": "flat",
28
+ "mm_projector_lr": null,
29
+ "mm_projector_type": "mlp2x_gelu",
30
+ "mm_resampler_type": null,
31
+ "mm_spatial_pool_mode": "bilinear",
32
+ "mm_spatial_pool_stride": null,
33
+ "mm_tunable_parts": "mm_mlp_adapter",
34
+ "mm_use_im_patch_token": false,
35
+ "mm_use_im_start_end": false,
36
+ "mm_vision_select_feature": "patch",
37
+ "mm_vision_select_layer": -2,
38
+ "mm_vision_tower": "open_clip_hub:ViT-SO400M-16-SigLIP2-256",
39
+ "mm_vision_tower_lr": null,
40
+ "model_type": "qwen3",
41
+ "num_attention_heads": 16,
42
+ "num_hidden_layers": 28,
43
+ "num_key_value_heads": 8,
44
+ "pos_skipping_range": 4096,
45
+ "rms_norm_eps": 1e-06,
46
+ "rope_scaling": null,
47
+ "rope_theta": 1000000,
48
+ "sliding_window": null,
49
+ "tie_word_embeddings": true,
50
+ "tokenizer_model_max_length": 8192,
51
+ "tokenizer_padding_side": "right",
52
+ "torch_dtype": "bfloat16",
53
+ "transformers_version": "4.51.3",
54
+ "use_cache": true,
55
+ "use_mm_proj": true,
56
+ "use_pos_skipping": false,
57
+ "use_sliding_window": false,
58
+ "vision_tower_pretrained": "/cache/ckpt/download/tokenizer/continuous/siglip2/siglip2_sm16_256.npz",
59
+ "vocab_size": 151936
60
+ }
continuous/qwen3/siglip2_sm16_256/pretrain/mm_projector.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:db55a18e088214db19f46cbd9e27b65e7d967ea4c10478e5a4c8b84d1016138b
3
+ size 26232888
continuous/qwen3/siglip2_sm16_256/pretrain/trainer_state.json ADDED
@@ -0,0 +1,421 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_global_step": null,
3
+ "best_metric": null,
4
+ "best_model_checkpoint": null,
5
+ "epoch": 1.0,
6
+ "eval_steps": 500,
7
+ "global_step": 545,
8
+ "is_hyper_param_search": false,
9
+ "is_local_process_zero": true,
10
+ "is_world_process_zero": true,
11
+ "log_history": [
12
+ {
13
+ "epoch": 0.01834862385321101,
14
+ "grad_norm": 14.884354591369629,
15
+ "learning_rate": 0.00032142857142857147,
16
+ "loss": 9.2651,
17
+ "step": 10
18
+ },
19
+ {
20
+ "epoch": 0.03669724770642202,
21
+ "grad_norm": 2.299124240875244,
22
+ "learning_rate": 0.0006785714285714287,
23
+ "loss": 5.9244,
24
+ "step": 20
25
+ },
26
+ {
27
+ "epoch": 0.05504587155963303,
28
+ "grad_norm": 0.3154951333999634,
29
+ "learning_rate": 0.0009999907688176173,
30
+ "loss": 5.1095,
31
+ "step": 30
32
+ },
33
+ {
34
+ "epoch": 0.07339449541284404,
35
+ "grad_norm": 0.12303392589092255,
36
+ "learning_rate": 0.0009988834393115766,
37
+ "loss": 4.7747,
38
+ "step": 40
39
+ },
40
+ {
41
+ "epoch": 0.09174311926605505,
42
+ "grad_norm": 0.13689632713794708,
43
+ "learning_rate": 0.0009959345572940772,
44
+ "loss": 4.6847,
45
+ "step": 50
46
+ },
47
+ {
48
+ "epoch": 0.11009174311926606,
49
+ "grad_norm": 0.09773959219455719,
50
+ "learning_rate": 0.0009911550081155983,
51
+ "loss": 4.6384,
52
+ "step": 60
53
+ },
54
+ {
55
+ "epoch": 0.12844036697247707,
56
+ "grad_norm": 0.11991074681282043,
57
+ "learning_rate": 0.0009845624347566433,
58
+ "loss": 4.5878,
59
+ "step": 70
60
+ },
61
+ {
62
+ "epoch": 0.14678899082568808,
63
+ "grad_norm": 0.13685007393360138,
64
+ "learning_rate": 0.0009761811727013547,
65
+ "loss": 4.5464,
66
+ "step": 80
67
+ },
68
+ {
69
+ "epoch": 0.1651376146788991,
70
+ "grad_norm": 0.15269911289215088,
71
+ "learning_rate": 0.0009660421601067666,
72
+ "loss": 4.4825,
73
+ "step": 90
74
+ },
75
+ {
76
+ "epoch": 0.1834862385321101,
77
+ "grad_norm": 0.1835877001285553,
78
+ "learning_rate": 0.0009541828235992926,
79
+ "loss": 4.4293,
80
+ "step": 100
81
+ },
82
+ {
83
+ "epoch": 0.2018348623853211,
84
+ "grad_norm": 0.15100613236427307,
85
+ "learning_rate": 0.0009406469401200151,
86
+ "loss": 4.3719,
87
+ "step": 110
88
+ },
89
+ {
90
+ "epoch": 0.22018348623853212,
91
+ "grad_norm": 0.16482654213905334,
92
+ "learning_rate": 0.0009254844753287492,
93
+ "loss": 4.3334,
94
+ "step": 120
95
+ },
96
+ {
97
+ "epoch": 0.23853211009174313,
98
+ "grad_norm": 0.16256624460220337,
99
+ "learning_rate": 0.0009087513991633924,
100
+ "loss": 4.2837,
101
+ "step": 130
102
+ },
103
+ {
104
+ "epoch": 0.25688073394495414,
105
+ "grad_norm": 0.1534961611032486,
106
+ "learning_rate": 0.0008905094792353917,
107
+ "loss": 4.251,
108
+ "step": 140
109
+ },
110
+ {
111
+ "epoch": 0.27522935779816515,
112
+ "grad_norm": 0.15952779352664948,
113
+ "learning_rate": 0.0008708260528239789,
114
+ "loss": 4.2295,
115
+ "step": 150
116
+ },
117
+ {
118
+ "epoch": 0.29357798165137616,
119
+ "grad_norm": 0.15001919865608215,
120
+ "learning_rate": 0.0008497737783108237,
121
+ "loss": 4.1996,
122
+ "step": 160
123
+ },
124
+ {
125
+ "epoch": 0.3119266055045872,
126
+ "grad_norm": 0.14762714505195618,
127
+ "learning_rate": 0.0008274303669726426,
128
+ "loss": 4.1795,
129
+ "step": 170
130
+ },
131
+ {
132
+ "epoch": 0.3302752293577982,
133
+ "grad_norm": 0.1487431675195694,
134
+ "learning_rate": 0.0008038782961218137,
135
+ "loss": 4.1533,
136
+ "step": 180
137
+ },
138
+ {
139
+ "epoch": 0.3486238532110092,
140
+ "grad_norm": 0.15354697406291962,
141
+ "learning_rate": 0.000779204504653894,
142
+ "loss": 4.1371,
143
+ "step": 190
144
+ },
145
+ {
146
+ "epoch": 0.3669724770642202,
147
+ "grad_norm": 0.13474558293819427,
148
+ "learning_rate": 0.0007535000721258791,
149
+ "loss": 4.1039,
150
+ "step": 200
151
+ },
152
+ {
153
+ "epoch": 0.3853211009174312,
154
+ "grad_norm": 0.14867264032363892,
155
+ "learning_rate": 0.00072685988254984,
156
+ "loss": 4.0849,
157
+ "step": 210
158
+ },
159
+ {
160
+ "epoch": 0.4036697247706422,
161
+ "grad_norm": 0.1431742012500763,
162
+ "learning_rate": 0.0006993822741429906,
163
+ "loss": 4.0913,
164
+ "step": 220
165
+ },
166
+ {
167
+ "epoch": 0.42201834862385323,
168
+ "grad_norm": 0.12665586173534393,
169
+ "learning_rate": 0.0006711686763270818,
170
+ "loss": 4.067,
171
+ "step": 230
172
+ },
173
+ {
174
+ "epoch": 0.44036697247706424,
175
+ "grad_norm": 0.14106588065624237,
176
+ "learning_rate": 0.0006423232353170831,
177
+ "loss": 4.0413,
178
+ "step": 240
179
+ },
180
+ {
181
+ "epoch": 0.45871559633027525,
182
+ "grad_norm": 0.14046552777290344,
183
+ "learning_rate": 0.0006129524296812335,
184
+ "loss": 4.005,
185
+ "step": 250
186
+ },
187
+ {
188
+ "epoch": 0.47706422018348627,
189
+ "grad_norm": 0.15286125242710114,
190
+ "learning_rate": 0.0005831646772915651,
191
+ "loss": 3.9893,
192
+ "step": 260
193
+ },
194
+ {
195
+ "epoch": 0.4954128440366973,
196
+ "grad_norm": 0.17392466962337494,
197
+ "learning_rate": 0.0005530699351157809,
198
+ "loss": 3.9431,
199
+ "step": 270
200
+ },
201
+ {
202
+ "epoch": 0.5137614678899083,
203
+ "grad_norm": 0.14311498403549194,
204
+ "learning_rate": 0.0005227792933277943,
205
+ "loss": 3.9188,
206
+ "step": 280
207
+ },
208
+ {
209
+ "epoch": 0.5321100917431193,
210
+ "grad_norm": 0.13869452476501465,
211
+ "learning_rate": 0.0004924045652352092,
212
+ "loss": 3.8944,
213
+ "step": 290
214
+ },
215
+ {
216
+ "epoch": 0.5504587155963303,
217
+ "grad_norm": 0.13892978429794312,
218
+ "learning_rate": 0.0004620578745374633,
219
+ "loss": 3.8824,
220
+ "step": 300
221
+ },
222
+ {
223
+ "epoch": 0.5688073394495413,
224
+ "grad_norm": 0.14064829051494598,
225
+ "learning_rate": 0.0004318512414382059,
226
+ "loss": 3.8877,
227
+ "step": 310
228
+ },
229
+ {
230
+ "epoch": 0.5871559633027523,
231
+ "grad_norm": 0.13492688536643982,
232
+ "learning_rate": 0.0004018961691397148,
233
+ "loss": 3.8763,
234
+ "step": 320
235
+ },
236
+ {
237
+ "epoch": 0.6055045871559633,
238
+ "grad_norm": 0.13445539772510529,
239
+ "learning_rate": 0.0003723032322457458,
240
+ "loss": 3.8546,
241
+ "step": 330
242
+ },
243
+ {
244
+ "epoch": 0.6238532110091743,
245
+ "grad_norm": 0.13182802498340607,
246
+ "learning_rate": 0.0003431816685921625,
247
+ "loss": 3.8501,
248
+ "step": 340
249
+ },
250
+ {
251
+ "epoch": 0.6422018348623854,
252
+ "grad_norm": 0.13592582941055298,
253
+ "learning_rate": 0.0003146389760120416,
254
+ "loss": 3.8469,
255
+ "step": 350
256
+ },
257
+ {
258
+ "epoch": 0.6605504587155964,
259
+ "grad_norm": 0.1220010295510292,
260
+ "learning_rate": 0.00028678051552373485,
261
+ "loss": 3.8526,
262
+ "step": 360
263
+ },
264
+ {
265
+ "epoch": 0.6788990825688074,
266
+ "grad_norm": 0.1198761984705925,
267
+ "learning_rate": 0.0002597091224066581,
268
+ "loss": 3.8353,
269
+ "step": 370
270
+ },
271
+ {
272
+ "epoch": 0.6972477064220184,
273
+ "grad_norm": 0.1179768368601799,
274
+ "learning_rate": 0.0002335247266004629,
275
+ "loss": 3.847,
276
+ "step": 380
277
+ },
278
+ {
279
+ "epoch": 0.7155963302752294,
280
+ "grad_norm": 0.13183355331420898,
281
+ "learning_rate": 0.000208323983828828,
282
+ "loss": 3.8294,
283
+ "step": 390
284
+ },
285
+ {
286
+ "epoch": 0.7339449541284404,
287
+ "grad_norm": 0.12682953476905823,
288
+ "learning_rate": 0.0001841999188095224,
289
+ "loss": 3.8253,
290
+ "step": 400
291
+ },
292
+ {
293
+ "epoch": 0.7522935779816514,
294
+ "grad_norm": 0.11767513304948807,
295
+ "learning_rate": 0.00016124158186777676,
296
+ "loss": 3.8303,
297
+ "step": 410
298
+ },
299
+ {
300
+ "epoch": 0.7706422018348624,
301
+ "grad_norm": 0.11222029477357864,
302
+ "learning_rate": 0.0001395337202205211,
303
+ "loss": 3.8216,
304
+ "step": 420
305
+ },
306
+ {
307
+ "epoch": 0.7889908256880734,
308
+ "grad_norm": 0.11920074373483658,
309
+ "learning_rate": 0.00011915646514489292,
310
+ "loss": 3.8229,
311
+ "step": 430
312
+ },
313
+ {
314
+ "epoch": 0.8073394495412844,
315
+ "grad_norm": 0.12086892127990723,
316
+ "learning_rate": 0.00010018503618578818,
317
+ "loss": 3.8264,
318
+ "step": 440
319
+ },
320
+ {
321
+ "epoch": 0.8256880733944955,
322
+ "grad_norm": 0.12085786461830139,
323
+ "learning_rate": 8.268946349432583e-05,
324
+ "loss": 3.8171,
325
+ "step": 450
326
+ },
327
+ {
328
+ "epoch": 0.8440366972477065,
329
+ "grad_norm": 0.11471433192491531,
330
+ "learning_rate": 6.673432932217117e-05,
331
+ "loss": 3.814,
332
+ "step": 460
333
+ },
334
+ {
335
+ "epoch": 0.8623853211009175,
336
+ "grad_norm": 0.11380089819431305,
337
+ "learning_rate": 5.2378529625954684e-05,
338
+ "loss": 3.8005,
339
+ "step": 470
340
+ },
341
+ {
342
+ "epoch": 0.8807339449541285,
343
+ "grad_norm": 0.11350731551647186,
344
+ "learning_rate": 3.9675056661785556e-05,
345
+ "loss": 3.8094,
346
+ "step": 480
347
+ },
348
+ {
349
+ "epoch": 0.8990825688073395,
350
+ "grad_norm": 0.11172082275152206,
351
+ "learning_rate": 2.8670803372379538e-05,
352
+ "loss": 3.8075,
353
+ "step": 490
354
+ },
355
+ {
356
+ "epoch": 0.9174311926605505,
357
+ "grad_norm": 0.11159991472959518,
358
+ "learning_rate": 1.9406390288876586e-05,
359
+ "loss": 3.8193,
360
+ "step": 500
361
+ },
362
+ {
363
+ "epoch": 0.9357798165137615,
364
+ "grad_norm": 0.11386371403932571,
365
+ "learning_rate": 1.191601558631461e-05,
366
+ "loss": 3.8119,
367
+ "step": 510
368
+ },
369
+ {
370
+ "epoch": 0.9541284403669725,
371
+ "grad_norm": 0.11355827748775482,
372
+ "learning_rate": 6.2273288462559305e-06,
373
+ "loss": 3.8034,
374
+ "step": 520
375
+ },
376
+ {
377
+ "epoch": 0.9724770642201835,
378
+ "grad_norm": 0.11273953318595886,
379
+ "learning_rate": 2.361328992552314e-06,
380
+ "loss": 3.808,
381
+ "step": 530
382
+ },
383
+ {
384
+ "epoch": 0.9908256880733946,
385
+ "grad_norm": 0.11326994001865387,
386
+ "learning_rate": 3.322867770044202e-07,
387
+ "loss": 3.806,
388
+ "step": 540
389
+ },
390
+ {
391
+ "epoch": 1.0,
392
+ "step": 545,
393
+ "total_flos": 3.666645035861934e+17,
394
+ "train_loss": 4.191423409138251,
395
+ "train_runtime": 5594.4151,
396
+ "train_samples_per_second": 99.765,
397
+ "train_steps_per_second": 0.097
398
+ }
399
+ ],
400
+ "logging_steps": 10,
401
+ "max_steps": 545,
402
+ "num_input_tokens_seen": 0,
403
+ "num_train_epochs": 1,
404
+ "save_steps": 500,
405
+ "stateful_callbacks": {
406
+ "TrainerControl": {
407
+ "args": {
408
+ "should_epoch_stop": false,
409
+ "should_evaluate": false,
410
+ "should_log": false,
411
+ "should_save": false,
412
+ "should_training_stop": false
413
+ },
414
+ "attributes": {}
415
+ }
416
+ },
417
+ "total_flos": 3.666645035861934e+17,
418
+ "train_batch_size": 64,
419
+ "trial_name": null,
420
+ "trial_params": null
421
+ }
continuous/qwen3/siglip2_sm16_384/finetune/added_tokens.json ADDED
@@ -0,0 +1,28 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "</think>": 151668,
3
+ "</tool_call>": 151658,
4
+ "</tool_response>": 151666,
5
+ "<think>": 151667,
6
+ "<tool_call>": 151657,
7
+ "<tool_response>": 151665,
8
+ "<|box_end|>": 151649,
9
+ "<|box_start|>": 151648,
10
+ "<|endoftext|>": 151643,
11
+ "<|file_sep|>": 151664,
12
+ "<|fim_middle|>": 151660,
13
+ "<|fim_pad|>": 151662,
14
+ "<|fim_prefix|>": 151659,
15
+ "<|fim_suffix|>": 151661,
16
+ "<|im_end|>": 151645,
17
+ "<|im_start|>": 151644,
18
+ "<|image_pad|>": 151655,
19
+ "<|object_ref_end|>": 151647,
20
+ "<|object_ref_start|>": 151646,
21
+ "<|quad_end|>": 151651,
22
+ "<|quad_start|>": 151650,
23
+ "<|repo_name|>": 151663,
24
+ "<|video_pad|>": 151656,
25
+ "<|vision_end|>": 151653,
26
+ "<|vision_pad|>": 151654,
27
+ "<|vision_start|>": 151652
28
+ }
continuous/qwen3/siglip2_sm16_384/finetune/config.json ADDED
@@ -0,0 +1,59 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "add_faster_video": false,
3
+ "add_time_instruction": false,
4
+ "architectures": [
5
+ "LlavaQwen3ForCausalLM"
6
+ ],
7
+ "attention_bias": false,
8
+ "attention_dropout": 0.0,
9
+ "bos_token_id": 151643,
10
+ "eos_token_id": 151645,
11
+ "faster_token_stride": 10,
12
+ "force_sample": false,
13
+ "head_dim": 128,
14
+ "hidden_act": "silu",
15
+ "hidden_size": 2048,
16
+ "image_aspect_ratio": "square",
17
+ "image_crop_resolution": null,
18
+ "image_grid_pinpoints": null,
19
+ "image_split_resolution": null,
20
+ "initializer_range": 0.02,
21
+ "intermediate_size": 6144,
22
+ "max_position_embeddings": 40960,
23
+ "max_window_layers": 28,
24
+ "mm_hidden_size": 1152,
25
+ "mm_newline_position": "grid",
26
+ "mm_patch_merge_type": "flat",
27
+ "mm_projector_lr": 0.0001,
28
+ "mm_projector_type": "mlp2x_gelu",
29
+ "mm_resampler_type": null,
30
+ "mm_spatial_pool_mode": "bilinear",
31
+ "mm_spatial_pool_stride": null,
32
+ "mm_tunable_parts": "mm_mlp_adapter,mm_language_model",
33
+ "mm_use_im_patch_token": false,
34
+ "mm_use_im_start_end": false,
35
+ "mm_vision_select_feature": "patch",
36
+ "mm_vision_select_layer": -2,
37
+ "mm_vision_tower": "open_clip_hub:ViT-SO400M-16-SigLIP2-384",
38
+ "mm_vision_tower_lr": null,
39
+ "model_type": "qwen3",
40
+ "num_attention_heads": 16,
41
+ "num_hidden_layers": 28,
42
+ "num_key_value_heads": 8,
43
+ "pos_skipping_range": 4096,
44
+ "rms_norm_eps": 1e-06,
45
+ "rope_scaling": null,
46
+ "rope_theta": 1000000,
47
+ "sliding_window": null,
48
+ "tie_word_embeddings": true,
49
+ "tokenizer_model_max_length": 8192,
50
+ "tokenizer_padding_side": "right",
51
+ "torch_dtype": "bfloat16",
52
+ "transformers_version": "4.51.3",
53
+ "use_cache": true,
54
+ "use_mm_proj": true,
55
+ "use_pos_skipping": false,
56
+ "use_sliding_window": false,
57
+ "vision_tower_pretrained": "/cache/ckpt/download/tokenizer/continuous/siglip2/siglip2_sm16_384.npz",
58
+ "vocab_size": 151936
59
+ }
continuous/qwen3/siglip2_sm16_384/finetune/generation_config.json ADDED
@@ -0,0 +1,13 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "bos_token_id": 151643,
3
+ "do_sample": true,
4
+ "eos_token_id": [
5
+ 151645,
6
+ 151643
7
+ ],
8
+ "pad_token_id": 151643,
9
+ "temperature": 0.6,
10
+ "top_k": 20,
11
+ "top_p": 0.95,
12
+ "transformers_version": "4.51.3"
13
+ }
continuous/qwen3/siglip2_sm16_384/finetune/merges.txt ADDED
The diff for this file is too large to render. See raw diff