PhaaNe commited on
Commit
76b8779
·
verified ·
1 Parent(s): d3130e6

Add files using upload-large-folder tool

Browse files
This view is limited to 50 files because it contains too many changes.   See raw diff
Files changed (50) hide show
  1. ood-viclickbait2025-xlmr-source-source-04-s44-20260901T140530Z-6863e5c6/trainer/checkpoint-720/trainer_state.json +338 -0
  2. ood-viclickbait2025-xlmr-source-source-05-s44-20260901T141022Z-6863e5c6/trainer/checkpoint-1350/special_tokens_map.json +15 -0
  3. ood-viclickbait2025-xlmr-source-source-05-s44-20260901T141022Z-6863e5c6/trainer/checkpoint-1350/tokenizer_config.json +55 -0
  4. ood-viclickbait2025-xlmr-source-source-05-s44-20260901T141022Z-6863e5c6/trainer/checkpoint-1350/trainer_state.json +586 -0
  5. ood-viclickbait2025-xlmr-source-source-05-s44-20260901T141022Z-6863e5c6/trainer/checkpoint-1620/special_tokens_map.json +15 -0
  6. ood-viclickbait2025-xlmr-source-source-05-s44-20260901T141022Z-6863e5c6/trainer/checkpoint-1620/tokenizer_config.json +55 -0
  7. ood-viclickbait2025-xlmr-source-source-05-s44-20260901T141022Z-6863e5c6/trainer/checkpoint-1620/trainer_state.json +689 -0
  8. ood-viclickbait2025-xlmr-source-source-06-s44-20260901T142033Z-6863e5c6/trainer/checkpoint-516/special_tokens_map.json +15 -0
  9. ood-viclickbait2025-xlmr-source-source-06-s44-20260901T142033Z-6863e5c6/trainer/checkpoint-516/tokenizer_config.json +55 -0
  10. ood-viclickbait2025-xlmr-source-source-06-s44-20260901T142033Z-6863e5c6/trainer/checkpoint-516/trainer_state.json +249 -0
  11. ood-viclickbait2025-xlmr-source-source-06-s44-20260901T142033Z-6863e5c6/trainer/checkpoint-774/special_tokens_map.json +15 -0
  12. ood-viclickbait2025-xlmr-source-source-06-s44-20260901T142033Z-6863e5c6/trainer/checkpoint-774/tokenizer_config.json +55 -0
  13. ood-viclickbait2025-xlmr-source-source-06-s44-20260901T142033Z-6863e5c6/trainer/checkpoint-774/trainer_state.json +352 -0
  14. ood-viclickbait2025-xlmr-source-source-07-s44-20260901T142547Z-6863e5c6/trainer/checkpoint-711/special_tokens_map.json +15 -0
  15. ood-viclickbait2025-xlmr-source-source-07-s44-20260901T142547Z-6863e5c6/trainer/checkpoint-711/tokenizer_config.json +55 -0
  16. ood-viclickbait2025-xlmr-source-source-07-s44-20260901T142547Z-6863e5c6/trainer/checkpoint-711/trainer_state.json +338 -0
  17. ood-viclickbait2025-xlmr-source-source-07-s44-20260901T142547Z-6863e5c6/trainer/checkpoint-948/special_tokens_map.json +15 -0
  18. ood-viclickbait2025-xlmr-source-source-07-s44-20260901T142547Z-6863e5c6/trainer/checkpoint-948/tokenizer_config.json +55 -0
  19. ood-viclickbait2025-xlmr-source-source-07-s44-20260901T142547Z-6863e5c6/trainer/checkpoint-948/trainer_state.json +434 -0
  20. ood-vifactcheck-xlmr-source-source-00-s42-20260901T150602Z-27bda108/trainer/checkpoint-2548/special_tokens_map.json +15 -0
  21. ood-vifactcheck-xlmr-source-source-00-s42-20260901T150602Z-27bda108/trainer/checkpoint-2548/tokenizer_config.json +55 -0
  22. ood-vifactcheck-xlmr-source-source-00-s42-20260901T150602Z-27bda108/trainer/checkpoint-2548/trainer_state.json +904 -0
  23. ood-vifactcheck-xlmr-source-source-00-s42-20260901T150602Z-27bda108/trainer/checkpoint-3094/special_tokens_map.json +15 -0
  24. ood-vifactcheck-xlmr-source-source-00-s42-20260901T150602Z-27bda108/trainer/checkpoint-3094/tokenizer_config.json +55 -0
  25. ood-vifactcheck-xlmr-source-source-00-s42-20260901T150602Z-27bda108/trainer/checkpoint-3094/trainer_state.json +1091 -0
  26. ood-vifactcheck-xlmr-source-source-00-s43-20260901T164831Z-fbd63439/trainer/checkpoint-2548/special_tokens_map.json +15 -0
  27. ood-vifactcheck-xlmr-source-source-00-s43-20260901T164831Z-fbd63439/trainer/checkpoint-2548/tokenizer_config.json +55 -0
  28. ood-vifactcheck-xlmr-source-source-00-s43-20260901T164831Z-fbd63439/trainer/checkpoint-2548/trainer_state.json +904 -0
  29. ood-vifactcheck-xlmr-source-source-00-s43-20260901T164831Z-fbd63439/trainer/checkpoint-3094/special_tokens_map.json +15 -0
  30. ood-vifactcheck-xlmr-source-source-00-s43-20260901T164831Z-fbd63439/trainer/checkpoint-3094/tokenizer_config.json +55 -0
  31. ood-vifactcheck-xlmr-source-source-00-s43-20260901T164831Z-fbd63439/trainer/checkpoint-3094/trainer_state.json +1091 -0
  32. ood-vifactcheck-xlmr-source-source-01-s42-20260901T152349Z-27bda108/trainer/checkpoint-1197/special_tokens_map.json +15 -0
  33. ood-vifactcheck-xlmr-source-source-01-s42-20260901T152349Z-27bda108/trainer/checkpoint-1197/tokenizer_config.json +55 -0
  34. ood-vifactcheck-xlmr-source-source-01-s42-20260901T152349Z-27bda108/trainer/checkpoint-1197/trainer_state.json +449 -0
  35. ood-vifactcheck-xlmr-source-source-01-s42-20260901T152349Z-27bda108/trainer/checkpoint-1710/special_tokens_map.json +15 -0
  36. ood-vifactcheck-xlmr-source-source-01-s42-20260901T152349Z-27bda108/trainer/checkpoint-1710/tokenizer_config.json +55 -0
  37. ood-vifactcheck-xlmr-source-source-01-s42-20260901T152349Z-27bda108/trainer/checkpoint-1710/trainer_state.json +629 -0
  38. ood-vifactcheck-xlmr-source-source-01-s43-20260901T170609Z-fbd63439/trainer/checkpoint-1197/special_tokens_map.json +15 -0
  39. ood-vifactcheck-xlmr-source-source-01-s43-20260901T170609Z-fbd63439/trainer/checkpoint-1197/tokenizer_config.json +55 -0
  40. ood-vifactcheck-xlmr-source-source-01-s43-20260901T170609Z-fbd63439/trainer/checkpoint-1197/trainer_state.json +449 -0
  41. ood-vifactcheck-xlmr-source-source-01-s43-20260901T170609Z-fbd63439/trainer/checkpoint-1710/special_tokens_map.json +15 -0
  42. ood-vifactcheck-xlmr-source-source-01-s43-20260901T170609Z-fbd63439/trainer/checkpoint-1710/tokenizer_config.json +55 -0
  43. ood-vifactcheck-xlmr-source-source-01-s43-20260901T170609Z-fbd63439/trainer/checkpoint-1710/trainer_state.json +629 -0
  44. ood-vifactcheck-xlmr-source-source-02-s42-20260901T153352Z-27bda108/trainer/checkpoint-1448/special_tokens_map.json +15 -0
  45. ood-vifactcheck-xlmr-source-source-02-s42-20260901T153352Z-27bda108/trainer/checkpoint-1448/tokenizer_config.json +55 -0
  46. ood-vifactcheck-xlmr-source-source-02-s42-20260901T153352Z-27bda108/trainer/checkpoint-1448/trainer_state.json +530 -0
  47. ood-vifactcheck-xlmr-source-source-02-s42-20260901T153352Z-27bda108/trainer/checkpoint-905/special_tokens_map.json +15 -0
  48. ood-vifactcheck-xlmr-source-source-02-s42-20260901T153352Z-27bda108/trainer/checkpoint-905/tokenizer_config.json +55 -0
  49. ood-vifactcheck-xlmr-source-source-02-s42-20260901T153352Z-27bda108/trainer/checkpoint-905/trainer_state.json +350 -0
  50. ood-vifactcheck-xlmr-source-source-02-s43-20260901T171619Z-fbd63439/trainer/checkpoint-1267/special_tokens_map.json +15 -0
ood-viclickbait2025-xlmr-source-source-04-s44-20260901T140530Z-6863e5c6/trainer/checkpoint-720/trainer_state.json ADDED
@@ -0,0 +1,338 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_global_step": 480,
3
+ "best_metric": 0.8749119097956307,
4
+ "best_model_checkpoint": "/root/VNEncoder/results/checkpoints/ood-viclickbait2025-xlmr-source-source-04-s44-20260901T140530Z-6863e5c6/trainer/checkpoint-480",
5
+ "epoch": 9.0,
6
+ "eval_steps": 500,
7
+ "global_step": 720,
8
+ "is_hyper_param_search": false,
9
+ "is_local_process_zero": true,
10
+ "is_world_process_zero": true,
11
+ "log_history": [
12
+ {
13
+ "epoch": 0.3125,
14
+ "grad_norm": 36.113670349121094,
15
+ "learning_rate": 3e-06,
16
+ "loss": 2.9318,
17
+ "step": 25
18
+ },
19
+ {
20
+ "epoch": 0.625,
21
+ "grad_norm": 120.8009262084961,
22
+ "learning_rate": 6.125000000000001e-06,
23
+ "loss": 2.6839,
24
+ "step": 50
25
+ },
26
+ {
27
+ "epoch": 0.9375,
28
+ "grad_norm": 27.118284225463867,
29
+ "learning_rate": 9.250000000000001e-06,
30
+ "loss": 2.4275,
31
+ "step": 75
32
+ },
33
+ {
34
+ "epoch": 1.0,
35
+ "eval_accuracy": 0.676056338028169,
36
+ "eval_loss": 0.5310995578765869,
37
+ "eval_macro_f1": 0.40336134453781514,
38
+ "eval_runtime": 0.6358,
39
+ "eval_samples_per_second": 446.665,
40
+ "eval_steps_per_second": 28.31,
41
+ "eval_weighted_f1": 0.5453899869807078,
42
+ "step": 80
43
+ },
44
+ {
45
+ "epoch": 1.25,
46
+ "grad_norm": 33.2912712097168,
47
+ "learning_rate": 1.2375000000000001e-05,
48
+ "loss": 2.1637,
49
+ "step": 100
50
+ },
51
+ {
52
+ "epoch": 1.5625,
53
+ "grad_norm": 93.63400268554688,
54
+ "learning_rate": 1.55e-05,
55
+ "loss": 2.1012,
56
+ "step": 125
57
+ },
58
+ {
59
+ "epoch": 1.875,
60
+ "grad_norm": 59.89738845825195,
61
+ "learning_rate": 1.8625000000000002e-05,
62
+ "loss": 2.019,
63
+ "step": 150
64
+ },
65
+ {
66
+ "epoch": 2.0,
67
+ "eval_accuracy": 0.8345070422535211,
68
+ "eval_loss": 0.39797243475914,
69
+ "eval_macro_f1": 0.8198528915581348,
70
+ "eval_runtime": 0.639,
71
+ "eval_samples_per_second": 444.427,
72
+ "eval_steps_per_second": 28.168,
73
+ "eval_weighted_f1": 0.837944435626513,
74
+ "step": 160
75
+ },
76
+ {
77
+ "epoch": 2.1875,
78
+ "grad_norm": 48.018943786621094,
79
+ "learning_rate": 1.9805555555555557e-05,
80
+ "loss": 1.8596,
81
+ "step": 175
82
+ },
83
+ {
84
+ "epoch": 2.5,
85
+ "grad_norm": 97.55175018310547,
86
+ "learning_rate": 1.9458333333333333e-05,
87
+ "loss": 1.7207,
88
+ "step": 200
89
+ },
90
+ {
91
+ "epoch": 2.8125,
92
+ "grad_norm": 115.1563491821289,
93
+ "learning_rate": 1.9111111111111113e-05,
94
+ "loss": 1.6247,
95
+ "step": 225
96
+ },
97
+ {
98
+ "epoch": 3.0,
99
+ "eval_accuracy": 0.8908450704225352,
100
+ "eval_loss": 0.2916692793369293,
101
+ "eval_macro_f1": 0.8743129612974145,
102
+ "eval_runtime": 0.6489,
103
+ "eval_samples_per_second": 437.665,
104
+ "eval_steps_per_second": 27.739,
105
+ "eval_weighted_f1": 0.8903635526810268,
106
+ "step": 240
107
+ },
108
+ {
109
+ "epoch": 3.125,
110
+ "grad_norm": 30.12644386291504,
111
+ "learning_rate": 1.876388888888889e-05,
112
+ "loss": 1.4328,
113
+ "step": 250
114
+ },
115
+ {
116
+ "epoch": 3.4375,
117
+ "grad_norm": 69.03034210205078,
118
+ "learning_rate": 1.8416666666666666e-05,
119
+ "loss": 1.1091,
120
+ "step": 275
121
+ },
122
+ {
123
+ "epoch": 3.75,
124
+ "grad_norm": 72.46318054199219,
125
+ "learning_rate": 1.8069444444444446e-05,
126
+ "loss": 1.4007,
127
+ "step": 300
128
+ },
129
+ {
130
+ "epoch": 4.0,
131
+ "eval_accuracy": 0.8591549295774648,
132
+ "eval_loss": 0.32971417903900146,
133
+ "eval_macro_f1": 0.8409854423292273,
134
+ "eval_runtime": 0.6502,
135
+ "eval_samples_per_second": 436.762,
136
+ "eval_steps_per_second": 27.682,
137
+ "eval_weighted_f1": 0.8599119915461413,
138
+ "step": 320
139
+ },
140
+ {
141
+ "epoch": 4.0625,
142
+ "grad_norm": 28.667325973510742,
143
+ "learning_rate": 1.7722222222222222e-05,
144
+ "loss": 1.2847,
145
+ "step": 325
146
+ },
147
+ {
148
+ "epoch": 4.375,
149
+ "grad_norm": 44.14274597167969,
150
+ "learning_rate": 1.7375000000000002e-05,
151
+ "loss": 0.8504,
152
+ "step": 350
153
+ },
154
+ {
155
+ "epoch": 4.6875,
156
+ "grad_norm": 68.7794189453125,
157
+ "learning_rate": 1.702777777777778e-05,
158
+ "loss": 1.1538,
159
+ "step": 375
160
+ },
161
+ {
162
+ "epoch": 5.0,
163
+ "grad_norm": 70.4148178100586,
164
+ "learning_rate": 1.668055555555556e-05,
165
+ "loss": 1.0259,
166
+ "step": 400
167
+ },
168
+ {
169
+ "epoch": 5.0,
170
+ "eval_accuracy": 0.8802816901408451,
171
+ "eval_loss": 0.32677745819091797,
172
+ "eval_macro_f1": 0.8648376259798433,
173
+ "eval_runtime": 0.6628,
174
+ "eval_samples_per_second": 428.517,
175
+ "eval_steps_per_second": 27.16,
176
+ "eval_weighted_f1": 0.8809251928142202,
177
+ "step": 400
178
+ },
179
+ {
180
+ "epoch": 5.3125,
181
+ "grad_norm": 61.19955062866211,
182
+ "learning_rate": 1.6333333333333335e-05,
183
+ "loss": 0.7135,
184
+ "step": 425
185
+ },
186
+ {
187
+ "epoch": 5.625,
188
+ "grad_norm": 22.916311264038086,
189
+ "learning_rate": 1.5986111111111115e-05,
190
+ "loss": 0.7698,
191
+ "step": 450
192
+ },
193
+ {
194
+ "epoch": 5.9375,
195
+ "grad_norm": 21.670488357543945,
196
+ "learning_rate": 1.563888888888889e-05,
197
+ "loss": 0.5699,
198
+ "step": 475
199
+ },
200
+ {
201
+ "epoch": 6.0,
202
+ "eval_accuracy": 0.8943661971830986,
203
+ "eval_loss": 0.382527619600296,
204
+ "eval_macro_f1": 0.8749119097956307,
205
+ "eval_runtime": 0.6631,
206
+ "eval_samples_per_second": 428.31,
207
+ "eval_steps_per_second": 27.146,
208
+ "eval_weighted_f1": 0.892281809248727,
209
+ "step": 480
210
+ },
211
+ {
212
+ "epoch": 6.25,
213
+ "grad_norm": 108.23480224609375,
214
+ "learning_rate": 1.5291666666666668e-05,
215
+ "loss": 0.6613,
216
+ "step": 500
217
+ },
218
+ {
219
+ "epoch": 6.5625,
220
+ "grad_norm": 176.19325256347656,
221
+ "learning_rate": 1.4944444444444446e-05,
222
+ "loss": 0.4539,
223
+ "step": 525
224
+ },
225
+ {
226
+ "epoch": 6.875,
227
+ "grad_norm": 49.169189453125,
228
+ "learning_rate": 1.4597222222222223e-05,
229
+ "loss": 0.3864,
230
+ "step": 550
231
+ },
232
+ {
233
+ "epoch": 7.0,
234
+ "eval_accuracy": 0.8838028169014085,
235
+ "eval_loss": 0.37622955441474915,
236
+ "eval_macro_f1": 0.8684576192681797,
237
+ "eval_runtime": 0.6716,
238
+ "eval_samples_per_second": 422.896,
239
+ "eval_steps_per_second": 26.803,
240
+ "eval_weighted_f1": 0.8842774106426424,
241
+ "step": 560
242
+ },
243
+ {
244
+ "epoch": 7.1875,
245
+ "grad_norm": 55.59708786010742,
246
+ "learning_rate": 1.425e-05,
247
+ "loss": 0.4167,
248
+ "step": 575
249
+ },
250
+ {
251
+ "epoch": 7.5,
252
+ "grad_norm": 123.0556869506836,
253
+ "learning_rate": 1.3902777777777779e-05,
254
+ "loss": 0.3474,
255
+ "step": 600
256
+ },
257
+ {
258
+ "epoch": 7.8125,
259
+ "grad_norm": 126.29974365234375,
260
+ "learning_rate": 1.3555555555555557e-05,
261
+ "loss": 0.5173,
262
+ "step": 625
263
+ },
264
+ {
265
+ "epoch": 8.0,
266
+ "eval_accuracy": 0.8591549295774648,
267
+ "eval_loss": 0.5031241774559021,
268
+ "eval_macro_f1": 0.8441957428132543,
269
+ "eval_runtime": 0.6342,
270
+ "eval_samples_per_second": 447.795,
271
+ "eval_steps_per_second": 28.381,
272
+ "eval_weighted_f1": 0.8611948186816752,
273
+ "step": 640
274
+ },
275
+ {
276
+ "epoch": 8.125,
277
+ "grad_norm": 1.4242525100708008,
278
+ "learning_rate": 1.3208333333333334e-05,
279
+ "loss": 0.3358,
280
+ "step": 650
281
+ },
282
+ {
283
+ "epoch": 8.4375,
284
+ "grad_norm": 91.3457260131836,
285
+ "learning_rate": 1.2861111111111114e-05,
286
+ "loss": 0.4218,
287
+ "step": 675
288
+ },
289
+ {
290
+ "epoch": 8.75,
291
+ "grad_norm": 113.89402770996094,
292
+ "learning_rate": 1.251388888888889e-05,
293
+ "loss": 0.2973,
294
+ "step": 700
295
+ },
296
+ {
297
+ "epoch": 9.0,
298
+ "eval_accuracy": 0.8732394366197183,
299
+ "eval_loss": 0.7847744822502136,
300
+ "eval_macro_f1": 0.8544667767465695,
301
+ "eval_runtime": 0.6354,
302
+ "eval_samples_per_second": 446.983,
303
+ "eval_steps_per_second": 28.33,
304
+ "eval_weighted_f1": 0.8728713452496566,
305
+ "step": 720
306
+ }
307
+ ],
308
+ "logging_steps": 25,
309
+ "max_steps": 1600,
310
+ "num_input_tokens_seen": 0,
311
+ "num_train_epochs": 20,
312
+ "save_steps": 500,
313
+ "stateful_callbacks": {
314
+ "EarlyStoppingCallback": {
315
+ "args": {
316
+ "early_stopping_patience": 3,
317
+ "early_stopping_threshold": 0.0
318
+ },
319
+ "attributes": {
320
+ "early_stopping_patience_counter": 3
321
+ }
322
+ },
323
+ "TrainerControl": {
324
+ "args": {
325
+ "should_epoch_stop": false,
326
+ "should_evaluate": false,
327
+ "should_log": false,
328
+ "should_save": true,
329
+ "should_training_stop": true
330
+ },
331
+ "attributes": {}
332
+ }
333
+ },
334
+ "total_flos": 0.0,
335
+ "train_batch_size": 8,
336
+ "trial_name": null,
337
+ "trial_params": null
338
+ }
ood-viclickbait2025-xlmr-source-source-05-s44-20260901T141022Z-6863e5c6/trainer/checkpoint-1350/special_tokens_map.json ADDED
@@ -0,0 +1,15 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "bos_token": "<s>",
3
+ "cls_token": "<s>",
4
+ "eos_token": "</s>",
5
+ "mask_token": {
6
+ "content": "<mask>",
7
+ "lstrip": true,
8
+ "normalized": false,
9
+ "rstrip": false,
10
+ "single_word": false
11
+ },
12
+ "pad_token": "<pad>",
13
+ "sep_token": "</s>",
14
+ "unk_token": "<unk>"
15
+ }
ood-viclickbait2025-xlmr-source-source-05-s44-20260901T141022Z-6863e5c6/trainer/checkpoint-1350/tokenizer_config.json ADDED
@@ -0,0 +1,55 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "added_tokens_decoder": {
3
+ "0": {
4
+ "content": "<s>",
5
+ "lstrip": false,
6
+ "normalized": false,
7
+ "rstrip": false,
8
+ "single_word": false,
9
+ "special": true
10
+ },
11
+ "1": {
12
+ "content": "<pad>",
13
+ "lstrip": false,
14
+ "normalized": false,
15
+ "rstrip": false,
16
+ "single_word": false,
17
+ "special": true
18
+ },
19
+ "2": {
20
+ "content": "</s>",
21
+ "lstrip": false,
22
+ "normalized": false,
23
+ "rstrip": false,
24
+ "single_word": false,
25
+ "special": true
26
+ },
27
+ "3": {
28
+ "content": "<unk>",
29
+ "lstrip": false,
30
+ "normalized": false,
31
+ "rstrip": false,
32
+ "single_word": false,
33
+ "special": true
34
+ },
35
+ "250001": {
36
+ "content": "<mask>",
37
+ "lstrip": true,
38
+ "normalized": false,
39
+ "rstrip": false,
40
+ "single_word": false,
41
+ "special": true
42
+ }
43
+ },
44
+ "bos_token": "<s>",
45
+ "clean_up_tokenization_spaces": false,
46
+ "cls_token": "<s>",
47
+ "eos_token": "</s>",
48
+ "extra_special_tokens": {},
49
+ "mask_token": "<mask>",
50
+ "model_max_length": 512,
51
+ "pad_token": "<pad>",
52
+ "sep_token": "</s>",
53
+ "tokenizer_class": "XLMRobertaTokenizer",
54
+ "unk_token": "<unk>"
55
+ }
ood-viclickbait2025-xlmr-source-source-05-s44-20260901T141022Z-6863e5c6/trainer/checkpoint-1350/trainer_state.json ADDED
@@ -0,0 +1,586 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_global_step": 1350,
3
+ "best_metric": 0.8314726209463051,
4
+ "best_model_checkpoint": "/root/VNEncoder/results/checkpoints/ood-viclickbait2025-xlmr-source-source-05-s44-20260901T141022Z-6863e5c6/trainer/checkpoint-1350",
5
+ "epoch": 15.0,
6
+ "eval_steps": 500,
7
+ "global_step": 1350,
8
+ "is_hyper_param_search": false,
9
+ "is_local_process_zero": true,
10
+ "is_world_process_zero": true,
11
+ "log_history": [
12
+ {
13
+ "epoch": 0.2801120448179272,
14
+ "grad_norm": 21.286026000976562,
15
+ "learning_rate": 2.666666666666667e-06,
16
+ "loss": 2.9303,
17
+ "step": 25
18
+ },
19
+ {
20
+ "epoch": 0.5602240896358543,
21
+ "grad_norm": 18.28496742248535,
22
+ "learning_rate": 5.444444444444445e-06,
23
+ "loss": 2.7658,
24
+ "step": 50
25
+ },
26
+ {
27
+ "epoch": 0.8403361344537815,
28
+ "grad_norm": 23.925424575805664,
29
+ "learning_rate": 8.222222222222222e-06,
30
+ "loss": 2.427,
31
+ "step": 75
32
+ },
33
+ {
34
+ "epoch": 1.0,
35
+ "eval_accuracy": 0.7665615141955836,
36
+ "eval_loss": 0.5131620168685913,
37
+ "eval_macro_f1": 0.7311835350201688,
38
+ "eval_runtime": 0.7127,
39
+ "eval_samples_per_second": 444.81,
40
+ "eval_steps_per_second": 28.064,
41
+ "eval_weighted_f1": 0.759793552962026,
42
+ "step": 90
43
+ },
44
+ {
45
+ "epoch": 1.112044817927171,
46
+ "grad_norm": 42.43369674682617,
47
+ "learning_rate": 1.1000000000000001e-05,
48
+ "loss": 2.151,
49
+ "step": 100
50
+ },
51
+ {
52
+ "epoch": 1.392156862745098,
53
+ "grad_norm": 75.90241241455078,
54
+ "learning_rate": 1.377777777777778e-05,
55
+ "loss": 1.8566,
56
+ "step": 125
57
+ },
58
+ {
59
+ "epoch": 1.6722689075630253,
60
+ "grad_norm": 39.13203048706055,
61
+ "learning_rate": 1.6555555555555556e-05,
62
+ "loss": 2.0029,
63
+ "step": 150
64
+ },
65
+ {
66
+ "epoch": 1.9523809523809523,
67
+ "grad_norm": 113.73194122314453,
68
+ "learning_rate": 1.9333333333333333e-05,
69
+ "loss": 1.7688,
70
+ "step": 175
71
+ },
72
+ {
73
+ "epoch": 2.0,
74
+ "eval_accuracy": 0.8107255520504731,
75
+ "eval_loss": 0.535222053527832,
76
+ "eval_macro_f1": 0.7784456248252726,
77
+ "eval_runtime": 0.6496,
78
+ "eval_samples_per_second": 488.029,
79
+ "eval_steps_per_second": 30.791,
80
+ "eval_weighted_f1": 0.803255816824807,
81
+ "step": 180
82
+ },
83
+ {
84
+ "epoch": 2.224089635854342,
85
+ "grad_norm": 58.54833984375,
86
+ "learning_rate": 1.9765432098765434e-05,
87
+ "loss": 1.5445,
88
+ "step": 200
89
+ },
90
+ {
91
+ "epoch": 2.504201680672269,
92
+ "grad_norm": 121.798095703125,
93
+ "learning_rate": 1.9456790123456793e-05,
94
+ "loss": 1.6574,
95
+ "step": 225
96
+ },
97
+ {
98
+ "epoch": 2.784313725490196,
99
+ "grad_norm": 34.078330993652344,
100
+ "learning_rate": 1.9148148148148148e-05,
101
+ "loss": 1.455,
102
+ "step": 250
103
+ },
104
+ {
105
+ "epoch": 3.0,
106
+ "eval_accuracy": 0.8107255520504731,
107
+ "eval_loss": 0.5058051943778992,
108
+ "eval_macro_f1": 0.803933696190005,
109
+ "eval_runtime": 0.6798,
110
+ "eval_samples_per_second": 466.328,
111
+ "eval_steps_per_second": 29.421,
112
+ "eval_weighted_f1": 0.8146395028853194,
113
+ "step": 270
114
+ },
115
+ {
116
+ "epoch": 3.0560224089635852,
117
+ "grad_norm": 53.804298400878906,
118
+ "learning_rate": 1.8839506172839507e-05,
119
+ "loss": 1.3164,
120
+ "step": 275
121
+ },
122
+ {
123
+ "epoch": 3.3361344537815127,
124
+ "grad_norm": 46.20854949951172,
125
+ "learning_rate": 1.8530864197530866e-05,
126
+ "loss": 1.3639,
127
+ "step": 300
128
+ },
129
+ {
130
+ "epoch": 3.61624649859944,
131
+ "grad_norm": 45.22013473510742,
132
+ "learning_rate": 1.8222222222222224e-05,
133
+ "loss": 1.1226,
134
+ "step": 325
135
+ },
136
+ {
137
+ "epoch": 3.896358543417367,
138
+ "grad_norm": 48.710594177246094,
139
+ "learning_rate": 1.7913580246913583e-05,
140
+ "loss": 1.1293,
141
+ "step": 350
142
+ },
143
+ {
144
+ "epoch": 4.0,
145
+ "eval_accuracy": 0.8170347003154574,
146
+ "eval_loss": 0.3787948191165924,
147
+ "eval_macro_f1": 0.7998040069686412,
148
+ "eval_runtime": 0.7137,
149
+ "eval_samples_per_second": 444.141,
150
+ "eval_steps_per_second": 28.022,
151
+ "eval_weighted_f1": 0.8170347003154574,
152
+ "step": 360
153
+ },
154
+ {
155
+ "epoch": 4.168067226890757,
156
+ "grad_norm": 83.72383880615234,
157
+ "learning_rate": 1.760493827160494e-05,
158
+ "loss": 0.8385,
159
+ "step": 375
160
+ },
161
+ {
162
+ "epoch": 4.448179271708684,
163
+ "grad_norm": 62.41585922241211,
164
+ "learning_rate": 1.7296296296296297e-05,
165
+ "loss": 0.9109,
166
+ "step": 400
167
+ },
168
+ {
169
+ "epoch": 4.728291316526611,
170
+ "grad_norm": 37.59474563598633,
171
+ "learning_rate": 1.6987654320987656e-05,
172
+ "loss": 0.9558,
173
+ "step": 425
174
+ },
175
+ {
176
+ "epoch": 5.0,
177
+ "grad_norm": 4.466841697692871,
178
+ "learning_rate": 1.6679012345679015e-05,
179
+ "loss": 0.8452,
180
+ "step": 450
181
+ },
182
+ {
183
+ "epoch": 5.0,
184
+ "eval_accuracy": 0.8138801261829653,
185
+ "eval_loss": 0.561353325843811,
186
+ "eval_macro_f1": 0.8031532527127867,
187
+ "eval_runtime": 0.6452,
188
+ "eval_samples_per_second": 491.315,
189
+ "eval_steps_per_second": 30.998,
190
+ "eval_weighted_f1": 0.8166343234253083,
191
+ "step": 450
192
+ },
193
+ {
194
+ "epoch": 5.280112044817927,
195
+ "grad_norm": 46.44953918457031,
196
+ "learning_rate": 1.6370370370370374e-05,
197
+ "loss": 0.6156,
198
+ "step": 475
199
+ },
200
+ {
201
+ "epoch": 5.560224089635854,
202
+ "grad_norm": 44.48462677001953,
203
+ "learning_rate": 1.606172839506173e-05,
204
+ "loss": 0.615,
205
+ "step": 500
206
+ },
207
+ {
208
+ "epoch": 5.840336134453781,
209
+ "grad_norm": 77.2662353515625,
210
+ "learning_rate": 1.5753086419753088e-05,
211
+ "loss": 0.7461,
212
+ "step": 525
213
+ },
214
+ {
215
+ "epoch": 6.0,
216
+ "eval_accuracy": 0.8422712933753943,
217
+ "eval_loss": 0.6332255601882935,
218
+ "eval_macro_f1": 0.8163979241960894,
219
+ "eval_runtime": 0.666,
220
+ "eval_samples_per_second": 476.01,
221
+ "eval_steps_per_second": 30.032,
222
+ "eval_weighted_f1": 0.8366182883446218,
223
+ "step": 540
224
+ },
225
+ {
226
+ "epoch": 6.1120448179271705,
227
+ "grad_norm": 5.867457389831543,
228
+ "learning_rate": 1.5444444444444446e-05,
229
+ "loss": 0.6319,
230
+ "step": 550
231
+ },
232
+ {
233
+ "epoch": 6.392156862745098,
234
+ "grad_norm": 3.1487138271331787,
235
+ "learning_rate": 1.5135802469135803e-05,
236
+ "loss": 0.3849,
237
+ "step": 575
238
+ },
239
+ {
240
+ "epoch": 6.6722689075630255,
241
+ "grad_norm": 39.26576232910156,
242
+ "learning_rate": 1.4827160493827162e-05,
243
+ "loss": 0.5494,
244
+ "step": 600
245
+ },
246
+ {
247
+ "epoch": 6.9523809523809526,
248
+ "grad_norm": 3.9887986183166504,
249
+ "learning_rate": 1.4518518518518521e-05,
250
+ "loss": 0.6336,
251
+ "step": 625
252
+ },
253
+ {
254
+ "epoch": 7.0,
255
+ "eval_accuracy": 0.8391167192429022,
256
+ "eval_loss": 0.7031358480453491,
257
+ "eval_macro_f1": 0.8197116188819378,
258
+ "eval_runtime": 0.6708,
259
+ "eval_samples_per_second": 472.594,
260
+ "eval_steps_per_second": 29.817,
261
+ "eval_weighted_f1": 0.8370642567047232,
262
+ "step": 630
263
+ },
264
+ {
265
+ "epoch": 7.224089635854342,
266
+ "grad_norm": 11.982047080993652,
267
+ "learning_rate": 1.4209876543209878e-05,
268
+ "loss": 0.3573,
269
+ "step": 650
270
+ },
271
+ {
272
+ "epoch": 7.504201680672269,
273
+ "grad_norm": 40.24326705932617,
274
+ "learning_rate": 1.3901234567901237e-05,
275
+ "loss": 0.3807,
276
+ "step": 675
277
+ },
278
+ {
279
+ "epoch": 7.784313725490196,
280
+ "grad_norm": 1.3944097757339478,
281
+ "learning_rate": 1.3592592592592592e-05,
282
+ "loss": 0.4012,
283
+ "step": 700
284
+ },
285
+ {
286
+ "epoch": 8.0,
287
+ "eval_accuracy": 0.8170347003154574,
288
+ "eval_loss": 1.148803472518921,
289
+ "eval_macro_f1": 0.805562605752961,
290
+ "eval_runtime": 0.665,
291
+ "eval_samples_per_second": 476.714,
292
+ "eval_steps_per_second": 30.077,
293
+ "eval_weighted_f1": 0.8194185121725995,
294
+ "step": 720
295
+ },
296
+ {
297
+ "epoch": 8.056022408963585,
298
+ "grad_norm": 212.5960235595703,
299
+ "learning_rate": 1.3283950617283951e-05,
300
+ "loss": 0.4906,
301
+ "step": 725
302
+ },
303
+ {
304
+ "epoch": 8.336134453781513,
305
+ "grad_norm": 318.5392761230469,
306
+ "learning_rate": 1.297530864197531e-05,
307
+ "loss": 0.3477,
308
+ "step": 750
309
+ },
310
+ {
311
+ "epoch": 8.61624649859944,
312
+ "grad_norm": 46.817447662353516,
313
+ "learning_rate": 1.2666666666666667e-05,
314
+ "loss": 0.3973,
315
+ "step": 775
316
+ },
317
+ {
318
+ "epoch": 8.896358543417367,
319
+ "grad_norm": 1.5035258531570435,
320
+ "learning_rate": 1.2358024691358026e-05,
321
+ "loss": 0.2958,
322
+ "step": 800
323
+ },
324
+ {
325
+ "epoch": 9.0,
326
+ "eval_accuracy": 0.8454258675078864,
327
+ "eval_loss": 0.9726073741912842,
328
+ "eval_macro_f1": 0.8290917093029653,
329
+ "eval_runtime": 0.6293,
330
+ "eval_samples_per_second": 503.774,
331
+ "eval_steps_per_second": 31.784,
332
+ "eval_weighted_f1": 0.8445924920892679,
333
+ "step": 810
334
+ },
335
+ {
336
+ "epoch": 9.168067226890756,
337
+ "grad_norm": 0.4589764177799225,
338
+ "learning_rate": 1.2049382716049384e-05,
339
+ "loss": 0.1497,
340
+ "step": 825
341
+ },
342
+ {
343
+ "epoch": 9.448179271708684,
344
+ "grad_norm": 8.860508918762207,
345
+ "learning_rate": 1.1740740740740741e-05,
346
+ "loss": 0.243,
347
+ "step": 850
348
+ },
349
+ {
350
+ "epoch": 9.72829131652661,
351
+ "grad_norm": 130.12161254882812,
352
+ "learning_rate": 1.14320987654321e-05,
353
+ "loss": 0.2929,
354
+ "step": 875
355
+ },
356
+ {
357
+ "epoch": 10.0,
358
+ "grad_norm": 0.031857386231422424,
359
+ "learning_rate": 1.1123456790123457e-05,
360
+ "loss": 0.2966,
361
+ "step": 900
362
+ },
363
+ {
364
+ "epoch": 10.0,
365
+ "eval_accuracy": 0.8422712933753943,
366
+ "eval_loss": 1.1647114753723145,
367
+ "eval_macro_f1": 0.8259771629336847,
368
+ "eval_runtime": 0.6312,
369
+ "eval_samples_per_second": 502.21,
370
+ "eval_steps_per_second": 31.685,
371
+ "eval_weighted_f1": 0.8415993704705815,
372
+ "step": 900
373
+ },
374
+ {
375
+ "epoch": 10.280112044817928,
376
+ "grad_norm": 1010.2288208007812,
377
+ "learning_rate": 1.0814814814814816e-05,
378
+ "loss": 0.1748,
379
+ "step": 925
380
+ },
381
+ {
382
+ "epoch": 10.560224089635854,
383
+ "grad_norm": 0.4934237003326416,
384
+ "learning_rate": 1.0506172839506175e-05,
385
+ "loss": 0.2865,
386
+ "step": 950
387
+ },
388
+ {
389
+ "epoch": 10.840336134453782,
390
+ "grad_norm": 6.401635646820068,
391
+ "learning_rate": 1.0197530864197532e-05,
392
+ "loss": 0.2486,
393
+ "step": 975
394
+ },
395
+ {
396
+ "epoch": 11.0,
397
+ "eval_accuracy": 0.8359621451104101,
398
+ "eval_loss": 1.4146658182144165,
399
+ "eval_macro_f1": 0.8219255034137067,
400
+ "eval_runtime": 0.6608,
401
+ "eval_samples_per_second": 479.755,
402
+ "eval_steps_per_second": 30.268,
403
+ "eval_weighted_f1": 0.8365930054113856,
404
+ "step": 990
405
+ },
406
+ {
407
+ "epoch": 11.11204481792717,
408
+ "grad_norm": 438.1229248046875,
409
+ "learning_rate": 9.88888888888889e-06,
410
+ "loss": 0.1684,
411
+ "step": 1000
412
+ },
413
+ {
414
+ "epoch": 11.392156862745098,
415
+ "grad_norm": 416.82550048828125,
416
+ "learning_rate": 9.580246913580248e-06,
417
+ "loss": 0.1978,
418
+ "step": 1025
419
+ },
420
+ {
421
+ "epoch": 11.672268907563025,
422
+ "grad_norm": 204.48353576660156,
423
+ "learning_rate": 9.271604938271606e-06,
424
+ "loss": 0.2504,
425
+ "step": 1050
426
+ },
427
+ {
428
+ "epoch": 11.952380952380953,
429
+ "grad_norm": 3.435577869415283,
430
+ "learning_rate": 8.962962962962963e-06,
431
+ "loss": 0.2117,
432
+ "step": 1075
433
+ },
434
+ {
435
+ "epoch": 12.0,
436
+ "eval_accuracy": 0.8485804416403786,
437
+ "eval_loss": 1.2211965322494507,
438
+ "eval_macro_f1": 0.8299204149154968,
439
+ "eval_runtime": 0.6933,
440
+ "eval_samples_per_second": 457.219,
441
+ "eval_steps_per_second": 28.847,
442
+ "eval_weighted_f1": 0.8464478671575351,
443
+ "step": 1080
444
+ },
445
+ {
446
+ "epoch": 12.224089635854341,
447
+ "grad_norm": 0.07730665057897568,
448
+ "learning_rate": 8.654320987654322e-06,
449
+ "loss": 0.1066,
450
+ "step": 1100
451
+ },
452
+ {
453
+ "epoch": 12.504201680672269,
454
+ "grad_norm": 0.006811291445046663,
455
+ "learning_rate": 8.34567901234568e-06,
456
+ "loss": 0.1689,
457
+ "step": 1125
458
+ },
459
+ {
460
+ "epoch": 12.784313725490197,
461
+ "grad_norm": 26.207834243774414,
462
+ "learning_rate": 8.037037037037038e-06,
463
+ "loss": 0.102,
464
+ "step": 1150
465
+ },
466
+ {
467
+ "epoch": 13.0,
468
+ "eval_accuracy": 0.8454258675078864,
469
+ "eval_loss": 1.3052610158920288,
470
+ "eval_macro_f1": 0.8290917093029653,
471
+ "eval_runtime": 0.6859,
472
+ "eval_samples_per_second": 462.174,
473
+ "eval_steps_per_second": 29.159,
474
+ "eval_weighted_f1": 0.8445924920892679,
475
+ "step": 1170
476
+ },
477
+ {
478
+ "epoch": 13.056022408963585,
479
+ "grad_norm": 161.05384826660156,
480
+ "learning_rate": 7.728395061728395e-06,
481
+ "loss": 0.199,
482
+ "step": 1175
483
+ },
484
+ {
485
+ "epoch": 13.336134453781513,
486
+ "grad_norm": 158.7040557861328,
487
+ "learning_rate": 7.419753086419754e-06,
488
+ "loss": 0.1567,
489
+ "step": 1200
490
+ },
491
+ {
492
+ "epoch": 13.61624649859944,
493
+ "grad_norm": 0.009637095965445042,
494
+ "learning_rate": 7.111111111111112e-06,
495
+ "loss": 0.0224,
496
+ "step": 1225
497
+ },
498
+ {
499
+ "epoch": 13.896358543417367,
500
+ "grad_norm": 4.374759197235107,
501
+ "learning_rate": 6.80246913580247e-06,
502
+ "loss": 0.1211,
503
+ "step": 1250
504
+ },
505
+ {
506
+ "epoch": 14.0,
507
+ "eval_accuracy": 0.8391167192429022,
508
+ "eval_loss": 1.3063347339630127,
509
+ "eval_macro_f1": 0.8205361602930565,
510
+ "eval_runtime": 0.6592,
511
+ "eval_samples_per_second": 480.858,
512
+ "eval_steps_per_second": 30.338,
513
+ "eval_weighted_f1": 0.8374772581590924,
514
+ "step": 1260
515
+ },
516
+ {
517
+ "epoch": 14.168067226890756,
518
+ "grad_norm": 0.043752290308475494,
519
+ "learning_rate": 6.493827160493828e-06,
520
+ "loss": 0.1534,
521
+ "step": 1275
522
+ },
523
+ {
524
+ "epoch": 14.448179271708684,
525
+ "grad_norm": 0.2848133146762848,
526
+ "learning_rate": 6.1851851851851856e-06,
527
+ "loss": 0.1551,
528
+ "step": 1300
529
+ },
530
+ {
531
+ "epoch": 14.72829131652661,
532
+ "grad_norm": 0.20860472321510315,
533
+ "learning_rate": 5.876543209876544e-06,
534
+ "loss": 0.0696,
535
+ "step": 1325
536
+ },
537
+ {
538
+ "epoch": 15.0,
539
+ "grad_norm": 0.0009489697404205799,
540
+ "learning_rate": 5.567901234567901e-06,
541
+ "loss": 0.1178,
542
+ "step": 1350
543
+ },
544
+ {
545
+ "epoch": 15.0,
546
+ "eval_accuracy": 0.8485804416403786,
547
+ "eval_loss": 1.4265798330307007,
548
+ "eval_macro_f1": 0.8314726209463051,
549
+ "eval_runtime": 0.6825,
550
+ "eval_samples_per_second": 464.443,
551
+ "eval_steps_per_second": 29.302,
552
+ "eval_weighted_f1": 0.8472253667339172,
553
+ "step": 1350
554
+ }
555
+ ],
556
+ "logging_steps": 25,
557
+ "max_steps": 1800,
558
+ "num_input_tokens_seen": 0,
559
+ "num_train_epochs": 20,
560
+ "save_steps": 500,
561
+ "stateful_callbacks": {
562
+ "EarlyStoppingCallback": {
563
+ "args": {
564
+ "early_stopping_patience": 3,
565
+ "early_stopping_threshold": 0.0
566
+ },
567
+ "attributes": {
568
+ "early_stopping_patience_counter": 0
569
+ }
570
+ },
571
+ "TrainerControl": {
572
+ "args": {
573
+ "should_epoch_stop": false,
574
+ "should_evaluate": false,
575
+ "should_log": false,
576
+ "should_save": true,
577
+ "should_training_stop": false
578
+ },
579
+ "attributes": {}
580
+ }
581
+ },
582
+ "total_flos": 0.0,
583
+ "train_batch_size": 8,
584
+ "trial_name": null,
585
+ "trial_params": null
586
+ }
ood-viclickbait2025-xlmr-source-source-05-s44-20260901T141022Z-6863e5c6/trainer/checkpoint-1620/special_tokens_map.json ADDED
@@ -0,0 +1,15 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "bos_token": "<s>",
3
+ "cls_token": "<s>",
4
+ "eos_token": "</s>",
5
+ "mask_token": {
6
+ "content": "<mask>",
7
+ "lstrip": true,
8
+ "normalized": false,
9
+ "rstrip": false,
10
+ "single_word": false
11
+ },
12
+ "pad_token": "<pad>",
13
+ "sep_token": "</s>",
14
+ "unk_token": "<unk>"
15
+ }
ood-viclickbait2025-xlmr-source-source-05-s44-20260901T141022Z-6863e5c6/trainer/checkpoint-1620/tokenizer_config.json ADDED
@@ -0,0 +1,55 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "added_tokens_decoder": {
3
+ "0": {
4
+ "content": "<s>",
5
+ "lstrip": false,
6
+ "normalized": false,
7
+ "rstrip": false,
8
+ "single_word": false,
9
+ "special": true
10
+ },
11
+ "1": {
12
+ "content": "<pad>",
13
+ "lstrip": false,
14
+ "normalized": false,
15
+ "rstrip": false,
16
+ "single_word": false,
17
+ "special": true
18
+ },
19
+ "2": {
20
+ "content": "</s>",
21
+ "lstrip": false,
22
+ "normalized": false,
23
+ "rstrip": false,
24
+ "single_word": false,
25
+ "special": true
26
+ },
27
+ "3": {
28
+ "content": "<unk>",
29
+ "lstrip": false,
30
+ "normalized": false,
31
+ "rstrip": false,
32
+ "single_word": false,
33
+ "special": true
34
+ },
35
+ "250001": {
36
+ "content": "<mask>",
37
+ "lstrip": true,
38
+ "normalized": false,
39
+ "rstrip": false,
40
+ "single_word": false,
41
+ "special": true
42
+ }
43
+ },
44
+ "bos_token": "<s>",
45
+ "clean_up_tokenization_spaces": false,
46
+ "cls_token": "<s>",
47
+ "eos_token": "</s>",
48
+ "extra_special_tokens": {},
49
+ "mask_token": "<mask>",
50
+ "model_max_length": 512,
51
+ "pad_token": "<pad>",
52
+ "sep_token": "</s>",
53
+ "tokenizer_class": "XLMRobertaTokenizer",
54
+ "unk_token": "<unk>"
55
+ }
ood-viclickbait2025-xlmr-source-source-05-s44-20260901T141022Z-6863e5c6/trainer/checkpoint-1620/trainer_state.json ADDED
@@ -0,0 +1,689 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_global_step": 1350,
3
+ "best_metric": 0.8314726209463051,
4
+ "best_model_checkpoint": "/root/VNEncoder/results/checkpoints/ood-viclickbait2025-xlmr-source-source-05-s44-20260901T141022Z-6863e5c6/trainer/checkpoint-1350",
5
+ "epoch": 18.0,
6
+ "eval_steps": 500,
7
+ "global_step": 1620,
8
+ "is_hyper_param_search": false,
9
+ "is_local_process_zero": true,
10
+ "is_world_process_zero": true,
11
+ "log_history": [
12
+ {
13
+ "epoch": 0.2801120448179272,
14
+ "grad_norm": 21.286026000976562,
15
+ "learning_rate": 2.666666666666667e-06,
16
+ "loss": 2.9303,
17
+ "step": 25
18
+ },
19
+ {
20
+ "epoch": 0.5602240896358543,
21
+ "grad_norm": 18.28496742248535,
22
+ "learning_rate": 5.444444444444445e-06,
23
+ "loss": 2.7658,
24
+ "step": 50
25
+ },
26
+ {
27
+ "epoch": 0.8403361344537815,
28
+ "grad_norm": 23.925424575805664,
29
+ "learning_rate": 8.222222222222222e-06,
30
+ "loss": 2.427,
31
+ "step": 75
32
+ },
33
+ {
34
+ "epoch": 1.0,
35
+ "eval_accuracy": 0.7665615141955836,
36
+ "eval_loss": 0.5131620168685913,
37
+ "eval_macro_f1": 0.7311835350201688,
38
+ "eval_runtime": 0.7127,
39
+ "eval_samples_per_second": 444.81,
40
+ "eval_steps_per_second": 28.064,
41
+ "eval_weighted_f1": 0.759793552962026,
42
+ "step": 90
43
+ },
44
+ {
45
+ "epoch": 1.112044817927171,
46
+ "grad_norm": 42.43369674682617,
47
+ "learning_rate": 1.1000000000000001e-05,
48
+ "loss": 2.151,
49
+ "step": 100
50
+ },
51
+ {
52
+ "epoch": 1.392156862745098,
53
+ "grad_norm": 75.90241241455078,
54
+ "learning_rate": 1.377777777777778e-05,
55
+ "loss": 1.8566,
56
+ "step": 125
57
+ },
58
+ {
59
+ "epoch": 1.6722689075630253,
60
+ "grad_norm": 39.13203048706055,
61
+ "learning_rate": 1.6555555555555556e-05,
62
+ "loss": 2.0029,
63
+ "step": 150
64
+ },
65
+ {
66
+ "epoch": 1.9523809523809523,
67
+ "grad_norm": 113.73194122314453,
68
+ "learning_rate": 1.9333333333333333e-05,
69
+ "loss": 1.7688,
70
+ "step": 175
71
+ },
72
+ {
73
+ "epoch": 2.0,
74
+ "eval_accuracy": 0.8107255520504731,
75
+ "eval_loss": 0.535222053527832,
76
+ "eval_macro_f1": 0.7784456248252726,
77
+ "eval_runtime": 0.6496,
78
+ "eval_samples_per_second": 488.029,
79
+ "eval_steps_per_second": 30.791,
80
+ "eval_weighted_f1": 0.803255816824807,
81
+ "step": 180
82
+ },
83
+ {
84
+ "epoch": 2.224089635854342,
85
+ "grad_norm": 58.54833984375,
86
+ "learning_rate": 1.9765432098765434e-05,
87
+ "loss": 1.5445,
88
+ "step": 200
89
+ },
90
+ {
91
+ "epoch": 2.504201680672269,
92
+ "grad_norm": 121.798095703125,
93
+ "learning_rate": 1.9456790123456793e-05,
94
+ "loss": 1.6574,
95
+ "step": 225
96
+ },
97
+ {
98
+ "epoch": 2.784313725490196,
99
+ "grad_norm": 34.078330993652344,
100
+ "learning_rate": 1.9148148148148148e-05,
101
+ "loss": 1.455,
102
+ "step": 250
103
+ },
104
+ {
105
+ "epoch": 3.0,
106
+ "eval_accuracy": 0.8107255520504731,
107
+ "eval_loss": 0.5058051943778992,
108
+ "eval_macro_f1": 0.803933696190005,
109
+ "eval_runtime": 0.6798,
110
+ "eval_samples_per_second": 466.328,
111
+ "eval_steps_per_second": 29.421,
112
+ "eval_weighted_f1": 0.8146395028853194,
113
+ "step": 270
114
+ },
115
+ {
116
+ "epoch": 3.0560224089635852,
117
+ "grad_norm": 53.804298400878906,
118
+ "learning_rate": 1.8839506172839507e-05,
119
+ "loss": 1.3164,
120
+ "step": 275
121
+ },
122
+ {
123
+ "epoch": 3.3361344537815127,
124
+ "grad_norm": 46.20854949951172,
125
+ "learning_rate": 1.8530864197530866e-05,
126
+ "loss": 1.3639,
127
+ "step": 300
128
+ },
129
+ {
130
+ "epoch": 3.61624649859944,
131
+ "grad_norm": 45.22013473510742,
132
+ "learning_rate": 1.8222222222222224e-05,
133
+ "loss": 1.1226,
134
+ "step": 325
135
+ },
136
+ {
137
+ "epoch": 3.896358543417367,
138
+ "grad_norm": 48.710594177246094,
139
+ "learning_rate": 1.7913580246913583e-05,
140
+ "loss": 1.1293,
141
+ "step": 350
142
+ },
143
+ {
144
+ "epoch": 4.0,
145
+ "eval_accuracy": 0.8170347003154574,
146
+ "eval_loss": 0.3787948191165924,
147
+ "eval_macro_f1": 0.7998040069686412,
148
+ "eval_runtime": 0.7137,
149
+ "eval_samples_per_second": 444.141,
150
+ "eval_steps_per_second": 28.022,
151
+ "eval_weighted_f1": 0.8170347003154574,
152
+ "step": 360
153
+ },
154
+ {
155
+ "epoch": 4.168067226890757,
156
+ "grad_norm": 83.72383880615234,
157
+ "learning_rate": 1.760493827160494e-05,
158
+ "loss": 0.8385,
159
+ "step": 375
160
+ },
161
+ {
162
+ "epoch": 4.448179271708684,
163
+ "grad_norm": 62.41585922241211,
164
+ "learning_rate": 1.7296296296296297e-05,
165
+ "loss": 0.9109,
166
+ "step": 400
167
+ },
168
+ {
169
+ "epoch": 4.728291316526611,
170
+ "grad_norm": 37.59474563598633,
171
+ "learning_rate": 1.6987654320987656e-05,
172
+ "loss": 0.9558,
173
+ "step": 425
174
+ },
175
+ {
176
+ "epoch": 5.0,
177
+ "grad_norm": 4.466841697692871,
178
+ "learning_rate": 1.6679012345679015e-05,
179
+ "loss": 0.8452,
180
+ "step": 450
181
+ },
182
+ {
183
+ "epoch": 5.0,
184
+ "eval_accuracy": 0.8138801261829653,
185
+ "eval_loss": 0.561353325843811,
186
+ "eval_macro_f1": 0.8031532527127867,
187
+ "eval_runtime": 0.6452,
188
+ "eval_samples_per_second": 491.315,
189
+ "eval_steps_per_second": 30.998,
190
+ "eval_weighted_f1": 0.8166343234253083,
191
+ "step": 450
192
+ },
193
+ {
194
+ "epoch": 5.280112044817927,
195
+ "grad_norm": 46.44953918457031,
196
+ "learning_rate": 1.6370370370370374e-05,
197
+ "loss": 0.6156,
198
+ "step": 475
199
+ },
200
+ {
201
+ "epoch": 5.560224089635854,
202
+ "grad_norm": 44.48462677001953,
203
+ "learning_rate": 1.606172839506173e-05,
204
+ "loss": 0.615,
205
+ "step": 500
206
+ },
207
+ {
208
+ "epoch": 5.840336134453781,
209
+ "grad_norm": 77.2662353515625,
210
+ "learning_rate": 1.5753086419753088e-05,
211
+ "loss": 0.7461,
212
+ "step": 525
213
+ },
214
+ {
215
+ "epoch": 6.0,
216
+ "eval_accuracy": 0.8422712933753943,
217
+ "eval_loss": 0.6332255601882935,
218
+ "eval_macro_f1": 0.8163979241960894,
219
+ "eval_runtime": 0.666,
220
+ "eval_samples_per_second": 476.01,
221
+ "eval_steps_per_second": 30.032,
222
+ "eval_weighted_f1": 0.8366182883446218,
223
+ "step": 540
224
+ },
225
+ {
226
+ "epoch": 6.1120448179271705,
227
+ "grad_norm": 5.867457389831543,
228
+ "learning_rate": 1.5444444444444446e-05,
229
+ "loss": 0.6319,
230
+ "step": 550
231
+ },
232
+ {
233
+ "epoch": 6.392156862745098,
234
+ "grad_norm": 3.1487138271331787,
235
+ "learning_rate": 1.5135802469135803e-05,
236
+ "loss": 0.3849,
237
+ "step": 575
238
+ },
239
+ {
240
+ "epoch": 6.6722689075630255,
241
+ "grad_norm": 39.26576232910156,
242
+ "learning_rate": 1.4827160493827162e-05,
243
+ "loss": 0.5494,
244
+ "step": 600
245
+ },
246
+ {
247
+ "epoch": 6.9523809523809526,
248
+ "grad_norm": 3.9887986183166504,
249
+ "learning_rate": 1.4518518518518521e-05,
250
+ "loss": 0.6336,
251
+ "step": 625
252
+ },
253
+ {
254
+ "epoch": 7.0,
255
+ "eval_accuracy": 0.8391167192429022,
256
+ "eval_loss": 0.7031358480453491,
257
+ "eval_macro_f1": 0.8197116188819378,
258
+ "eval_runtime": 0.6708,
259
+ "eval_samples_per_second": 472.594,
260
+ "eval_steps_per_second": 29.817,
261
+ "eval_weighted_f1": 0.8370642567047232,
262
+ "step": 630
263
+ },
264
+ {
265
+ "epoch": 7.224089635854342,
266
+ "grad_norm": 11.982047080993652,
267
+ "learning_rate": 1.4209876543209878e-05,
268
+ "loss": 0.3573,
269
+ "step": 650
270
+ },
271
+ {
272
+ "epoch": 7.504201680672269,
273
+ "grad_norm": 40.24326705932617,
274
+ "learning_rate": 1.3901234567901237e-05,
275
+ "loss": 0.3807,
276
+ "step": 675
277
+ },
278
+ {
279
+ "epoch": 7.784313725490196,
280
+ "grad_norm": 1.3944097757339478,
281
+ "learning_rate": 1.3592592592592592e-05,
282
+ "loss": 0.4012,
283
+ "step": 700
284
+ },
285
+ {
286
+ "epoch": 8.0,
287
+ "eval_accuracy": 0.8170347003154574,
288
+ "eval_loss": 1.148803472518921,
289
+ "eval_macro_f1": 0.805562605752961,
290
+ "eval_runtime": 0.665,
291
+ "eval_samples_per_second": 476.714,
292
+ "eval_steps_per_second": 30.077,
293
+ "eval_weighted_f1": 0.8194185121725995,
294
+ "step": 720
295
+ },
296
+ {
297
+ "epoch": 8.056022408963585,
298
+ "grad_norm": 212.5960235595703,
299
+ "learning_rate": 1.3283950617283951e-05,
300
+ "loss": 0.4906,
301
+ "step": 725
302
+ },
303
+ {
304
+ "epoch": 8.336134453781513,
305
+ "grad_norm": 318.5392761230469,
306
+ "learning_rate": 1.297530864197531e-05,
307
+ "loss": 0.3477,
308
+ "step": 750
309
+ },
310
+ {
311
+ "epoch": 8.61624649859944,
312
+ "grad_norm": 46.817447662353516,
313
+ "learning_rate": 1.2666666666666667e-05,
314
+ "loss": 0.3973,
315
+ "step": 775
316
+ },
317
+ {
318
+ "epoch": 8.896358543417367,
319
+ "grad_norm": 1.5035258531570435,
320
+ "learning_rate": 1.2358024691358026e-05,
321
+ "loss": 0.2958,
322
+ "step": 800
323
+ },
324
+ {
325
+ "epoch": 9.0,
326
+ "eval_accuracy": 0.8454258675078864,
327
+ "eval_loss": 0.9726073741912842,
328
+ "eval_macro_f1": 0.8290917093029653,
329
+ "eval_runtime": 0.6293,
330
+ "eval_samples_per_second": 503.774,
331
+ "eval_steps_per_second": 31.784,
332
+ "eval_weighted_f1": 0.8445924920892679,
333
+ "step": 810
334
+ },
335
+ {
336
+ "epoch": 9.168067226890756,
337
+ "grad_norm": 0.4589764177799225,
338
+ "learning_rate": 1.2049382716049384e-05,
339
+ "loss": 0.1497,
340
+ "step": 825
341
+ },
342
+ {
343
+ "epoch": 9.448179271708684,
344
+ "grad_norm": 8.860508918762207,
345
+ "learning_rate": 1.1740740740740741e-05,
346
+ "loss": 0.243,
347
+ "step": 850
348
+ },
349
+ {
350
+ "epoch": 9.72829131652661,
351
+ "grad_norm": 130.12161254882812,
352
+ "learning_rate": 1.14320987654321e-05,
353
+ "loss": 0.2929,
354
+ "step": 875
355
+ },
356
+ {
357
+ "epoch": 10.0,
358
+ "grad_norm": 0.031857386231422424,
359
+ "learning_rate": 1.1123456790123457e-05,
360
+ "loss": 0.2966,
361
+ "step": 900
362
+ },
363
+ {
364
+ "epoch": 10.0,
365
+ "eval_accuracy": 0.8422712933753943,
366
+ "eval_loss": 1.1647114753723145,
367
+ "eval_macro_f1": 0.8259771629336847,
368
+ "eval_runtime": 0.6312,
369
+ "eval_samples_per_second": 502.21,
370
+ "eval_steps_per_second": 31.685,
371
+ "eval_weighted_f1": 0.8415993704705815,
372
+ "step": 900
373
+ },
374
+ {
375
+ "epoch": 10.280112044817928,
376
+ "grad_norm": 1010.2288208007812,
377
+ "learning_rate": 1.0814814814814816e-05,
378
+ "loss": 0.1748,
379
+ "step": 925
380
+ },
381
+ {
382
+ "epoch": 10.560224089635854,
383
+ "grad_norm": 0.4934237003326416,
384
+ "learning_rate": 1.0506172839506175e-05,
385
+ "loss": 0.2865,
386
+ "step": 950
387
+ },
388
+ {
389
+ "epoch": 10.840336134453782,
390
+ "grad_norm": 6.401635646820068,
391
+ "learning_rate": 1.0197530864197532e-05,
392
+ "loss": 0.2486,
393
+ "step": 975
394
+ },
395
+ {
396
+ "epoch": 11.0,
397
+ "eval_accuracy": 0.8359621451104101,
398
+ "eval_loss": 1.4146658182144165,
399
+ "eval_macro_f1": 0.8219255034137067,
400
+ "eval_runtime": 0.6608,
401
+ "eval_samples_per_second": 479.755,
402
+ "eval_steps_per_second": 30.268,
403
+ "eval_weighted_f1": 0.8365930054113856,
404
+ "step": 990
405
+ },
406
+ {
407
+ "epoch": 11.11204481792717,
408
+ "grad_norm": 438.1229248046875,
409
+ "learning_rate": 9.88888888888889e-06,
410
+ "loss": 0.1684,
411
+ "step": 1000
412
+ },
413
+ {
414
+ "epoch": 11.392156862745098,
415
+ "grad_norm": 416.82550048828125,
416
+ "learning_rate": 9.580246913580248e-06,
417
+ "loss": 0.1978,
418
+ "step": 1025
419
+ },
420
+ {
421
+ "epoch": 11.672268907563025,
422
+ "grad_norm": 204.48353576660156,
423
+ "learning_rate": 9.271604938271606e-06,
424
+ "loss": 0.2504,
425
+ "step": 1050
426
+ },
427
+ {
428
+ "epoch": 11.952380952380953,
429
+ "grad_norm": 3.435577869415283,
430
+ "learning_rate": 8.962962962962963e-06,
431
+ "loss": 0.2117,
432
+ "step": 1075
433
+ },
434
+ {
435
+ "epoch": 12.0,
436
+ "eval_accuracy": 0.8485804416403786,
437
+ "eval_loss": 1.2211965322494507,
438
+ "eval_macro_f1": 0.8299204149154968,
439
+ "eval_runtime": 0.6933,
440
+ "eval_samples_per_second": 457.219,
441
+ "eval_steps_per_second": 28.847,
442
+ "eval_weighted_f1": 0.8464478671575351,
443
+ "step": 1080
444
+ },
445
+ {
446
+ "epoch": 12.224089635854341,
447
+ "grad_norm": 0.07730665057897568,
448
+ "learning_rate": 8.654320987654322e-06,
449
+ "loss": 0.1066,
450
+ "step": 1100
451
+ },
452
+ {
453
+ "epoch": 12.504201680672269,
454
+ "grad_norm": 0.006811291445046663,
455
+ "learning_rate": 8.34567901234568e-06,
456
+ "loss": 0.1689,
457
+ "step": 1125
458
+ },
459
+ {
460
+ "epoch": 12.784313725490197,
461
+ "grad_norm": 26.207834243774414,
462
+ "learning_rate": 8.037037037037038e-06,
463
+ "loss": 0.102,
464
+ "step": 1150
465
+ },
466
+ {
467
+ "epoch": 13.0,
468
+ "eval_accuracy": 0.8454258675078864,
469
+ "eval_loss": 1.3052610158920288,
470
+ "eval_macro_f1": 0.8290917093029653,
471
+ "eval_runtime": 0.6859,
472
+ "eval_samples_per_second": 462.174,
473
+ "eval_steps_per_second": 29.159,
474
+ "eval_weighted_f1": 0.8445924920892679,
475
+ "step": 1170
476
+ },
477
+ {
478
+ "epoch": 13.056022408963585,
479
+ "grad_norm": 161.05384826660156,
480
+ "learning_rate": 7.728395061728395e-06,
481
+ "loss": 0.199,
482
+ "step": 1175
483
+ },
484
+ {
485
+ "epoch": 13.336134453781513,
486
+ "grad_norm": 158.7040557861328,
487
+ "learning_rate": 7.419753086419754e-06,
488
+ "loss": 0.1567,
489
+ "step": 1200
490
+ },
491
+ {
492
+ "epoch": 13.61624649859944,
493
+ "grad_norm": 0.009637095965445042,
494
+ "learning_rate": 7.111111111111112e-06,
495
+ "loss": 0.0224,
496
+ "step": 1225
497
+ },
498
+ {
499
+ "epoch": 13.896358543417367,
500
+ "grad_norm": 4.374759197235107,
501
+ "learning_rate": 6.80246913580247e-06,
502
+ "loss": 0.1211,
503
+ "step": 1250
504
+ },
505
+ {
506
+ "epoch": 14.0,
507
+ "eval_accuracy": 0.8391167192429022,
508
+ "eval_loss": 1.3063347339630127,
509
+ "eval_macro_f1": 0.8205361602930565,
510
+ "eval_runtime": 0.6592,
511
+ "eval_samples_per_second": 480.858,
512
+ "eval_steps_per_second": 30.338,
513
+ "eval_weighted_f1": 0.8374772581590924,
514
+ "step": 1260
515
+ },
516
+ {
517
+ "epoch": 14.168067226890756,
518
+ "grad_norm": 0.043752290308475494,
519
+ "learning_rate": 6.493827160493828e-06,
520
+ "loss": 0.1534,
521
+ "step": 1275
522
+ },
523
+ {
524
+ "epoch": 14.448179271708684,
525
+ "grad_norm": 0.2848133146762848,
526
+ "learning_rate": 6.1851851851851856e-06,
527
+ "loss": 0.1551,
528
+ "step": 1300
529
+ },
530
+ {
531
+ "epoch": 14.72829131652661,
532
+ "grad_norm": 0.20860472321510315,
533
+ "learning_rate": 5.876543209876544e-06,
534
+ "loss": 0.0696,
535
+ "step": 1325
536
+ },
537
+ {
538
+ "epoch": 15.0,
539
+ "grad_norm": 0.0009489697404205799,
540
+ "learning_rate": 5.567901234567901e-06,
541
+ "loss": 0.1178,
542
+ "step": 1350
543
+ },
544
+ {
545
+ "epoch": 15.0,
546
+ "eval_accuracy": 0.8485804416403786,
547
+ "eval_loss": 1.4265798330307007,
548
+ "eval_macro_f1": 0.8314726209463051,
549
+ "eval_runtime": 0.6825,
550
+ "eval_samples_per_second": 464.443,
551
+ "eval_steps_per_second": 29.302,
552
+ "eval_weighted_f1": 0.8472253667339172,
553
+ "step": 1350
554
+ },
555
+ {
556
+ "epoch": 15.280112044817928,
557
+ "grad_norm": 0.014971903525292873,
558
+ "learning_rate": 5.259259259259259e-06,
559
+ "loss": 0.105,
560
+ "step": 1375
561
+ },
562
+ {
563
+ "epoch": 15.560224089635854,
564
+ "grad_norm": 0.008831561543047428,
565
+ "learning_rate": 4.950617283950618e-06,
566
+ "loss": 0.0525,
567
+ "step": 1400
568
+ },
569
+ {
570
+ "epoch": 15.840336134453782,
571
+ "grad_norm": 0.004474582616239786,
572
+ "learning_rate": 4.641975308641975e-06,
573
+ "loss": 0.1044,
574
+ "step": 1425
575
+ },
576
+ {
577
+ "epoch": 16.0,
578
+ "eval_accuracy": 0.8485804416403786,
579
+ "eval_loss": 1.442765235900879,
580
+ "eval_macro_f1": 0.8314726209463051,
581
+ "eval_runtime": 0.781,
582
+ "eval_samples_per_second": 405.865,
583
+ "eval_steps_per_second": 25.607,
584
+ "eval_weighted_f1": 0.8472253667339172,
585
+ "step": 1440
586
+ },
587
+ {
588
+ "epoch": 16.11204481792717,
589
+ "grad_norm": 0.01411131490021944,
590
+ "learning_rate": 4.333333333333334e-06,
591
+ "loss": 0.0241,
592
+ "step": 1450
593
+ },
594
+ {
595
+ "epoch": 16.392156862745097,
596
+ "grad_norm": 0.029822485521435738,
597
+ "learning_rate": 4.024691358024692e-06,
598
+ "loss": 0.2201,
599
+ "step": 1475
600
+ },
601
+ {
602
+ "epoch": 16.672268907563026,
603
+ "grad_norm": 0.8752092123031616,
604
+ "learning_rate": 3.7160493827160497e-06,
605
+ "loss": 0.1631,
606
+ "step": 1500
607
+ },
608
+ {
609
+ "epoch": 16.952380952380953,
610
+ "grad_norm": 0.03659430518746376,
611
+ "learning_rate": 3.4074074074074077e-06,
612
+ "loss": 0.0123,
613
+ "step": 1525
614
+ },
615
+ {
616
+ "epoch": 17.0,
617
+ "eval_accuracy": 0.8454258675078864,
618
+ "eval_loss": 1.4357441663742065,
619
+ "eval_macro_f1": 0.8283437765916299,
620
+ "eval_runtime": 0.7362,
621
+ "eval_samples_per_second": 430.61,
622
+ "eval_steps_per_second": 27.168,
623
+ "eval_weighted_f1": 0.8442301211437485,
624
+ "step": 1530
625
+ },
626
+ {
627
+ "epoch": 17.22408963585434,
628
+ "grad_norm": 0.032733943313360214,
629
+ "learning_rate": 3.098765432098766e-06,
630
+ "loss": 0.0249,
631
+ "step": 1550
632
+ },
633
+ {
634
+ "epoch": 17.504201680672267,
635
+ "grad_norm": 0.008555321954190731,
636
+ "learning_rate": 2.7901234567901235e-06,
637
+ "loss": 0.0344,
638
+ "step": 1575
639
+ },
640
+ {
641
+ "epoch": 17.784313725490197,
642
+ "grad_norm": 0.003118259133771062,
643
+ "learning_rate": 2.481481481481482e-06,
644
+ "loss": 0.0059,
645
+ "step": 1600
646
+ },
647
+ {
648
+ "epoch": 18.0,
649
+ "eval_accuracy": 0.8454258675078864,
650
+ "eval_loss": 1.4717018604278564,
651
+ "eval_macro_f1": 0.827573957928623,
652
+ "eval_runtime": 0.6972,
653
+ "eval_samples_per_second": 454.657,
654
+ "eval_steps_per_second": 28.685,
655
+ "eval_weighted_f1": 0.8438506990155984,
656
+ "step": 1620
657
+ }
658
+ ],
659
+ "logging_steps": 25,
660
+ "max_steps": 1800,
661
+ "num_input_tokens_seen": 0,
662
+ "num_train_epochs": 20,
663
+ "save_steps": 500,
664
+ "stateful_callbacks": {
665
+ "EarlyStoppingCallback": {
666
+ "args": {
667
+ "early_stopping_patience": 3,
668
+ "early_stopping_threshold": 0.0
669
+ },
670
+ "attributes": {
671
+ "early_stopping_patience_counter": 3
672
+ }
673
+ },
674
+ "TrainerControl": {
675
+ "args": {
676
+ "should_epoch_stop": false,
677
+ "should_evaluate": false,
678
+ "should_log": false,
679
+ "should_save": true,
680
+ "should_training_stop": true
681
+ },
682
+ "attributes": {}
683
+ }
684
+ },
685
+ "total_flos": 0.0,
686
+ "train_batch_size": 8,
687
+ "trial_name": null,
688
+ "trial_params": null
689
+ }
ood-viclickbait2025-xlmr-source-source-06-s44-20260901T142033Z-6863e5c6/trainer/checkpoint-516/special_tokens_map.json ADDED
@@ -0,0 +1,15 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "bos_token": "<s>",
3
+ "cls_token": "<s>",
4
+ "eos_token": "</s>",
5
+ "mask_token": {
6
+ "content": "<mask>",
7
+ "lstrip": true,
8
+ "normalized": false,
9
+ "rstrip": false,
10
+ "single_word": false
11
+ },
12
+ "pad_token": "<pad>",
13
+ "sep_token": "</s>",
14
+ "unk_token": "<unk>"
15
+ }
ood-viclickbait2025-xlmr-source-source-06-s44-20260901T142033Z-6863e5c6/trainer/checkpoint-516/tokenizer_config.json ADDED
@@ -0,0 +1,55 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "added_tokens_decoder": {
3
+ "0": {
4
+ "content": "<s>",
5
+ "lstrip": false,
6
+ "normalized": false,
7
+ "rstrip": false,
8
+ "single_word": false,
9
+ "special": true
10
+ },
11
+ "1": {
12
+ "content": "<pad>",
13
+ "lstrip": false,
14
+ "normalized": false,
15
+ "rstrip": false,
16
+ "single_word": false,
17
+ "special": true
18
+ },
19
+ "2": {
20
+ "content": "</s>",
21
+ "lstrip": false,
22
+ "normalized": false,
23
+ "rstrip": false,
24
+ "single_word": false,
25
+ "special": true
26
+ },
27
+ "3": {
28
+ "content": "<unk>",
29
+ "lstrip": false,
30
+ "normalized": false,
31
+ "rstrip": false,
32
+ "single_word": false,
33
+ "special": true
34
+ },
35
+ "250001": {
36
+ "content": "<mask>",
37
+ "lstrip": true,
38
+ "normalized": false,
39
+ "rstrip": false,
40
+ "single_word": false,
41
+ "special": true
42
+ }
43
+ },
44
+ "bos_token": "<s>",
45
+ "clean_up_tokenization_spaces": false,
46
+ "cls_token": "<s>",
47
+ "eos_token": "</s>",
48
+ "extra_special_tokens": {},
49
+ "mask_token": "<mask>",
50
+ "model_max_length": 512,
51
+ "pad_token": "<pad>",
52
+ "sep_token": "</s>",
53
+ "tokenizer_class": "XLMRobertaTokenizer",
54
+ "unk_token": "<unk>"
55
+ }
ood-viclickbait2025-xlmr-source-source-06-s44-20260901T142033Z-6863e5c6/trainer/checkpoint-516/trainer_state.json ADDED
@@ -0,0 +1,249 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_global_step": 516,
3
+ "best_metric": 0.8428641025978643,
4
+ "best_model_checkpoint": "/root/VNEncoder/results/checkpoints/ood-viclickbait2025-xlmr-source-source-06-s44-20260901T142033Z-6863e5c6/trainer/checkpoint-516",
5
+ "epoch": 6.0,
6
+ "eval_steps": 500,
7
+ "global_step": 516,
8
+ "is_hyper_param_search": false,
9
+ "is_local_process_zero": true,
10
+ "is_world_process_zero": true,
11
+ "log_history": [
12
+ {
13
+ "epoch": 0.29069767441860467,
14
+ "grad_norm": 23.544675827026367,
15
+ "learning_rate": 2.790697674418605e-06,
16
+ "loss": 2.9263,
17
+ "step": 25
18
+ },
19
+ {
20
+ "epoch": 0.5813953488372093,
21
+ "grad_norm": 23.54766845703125,
22
+ "learning_rate": 5.697674418604652e-06,
23
+ "loss": 2.7261,
24
+ "step": 50
25
+ },
26
+ {
27
+ "epoch": 0.872093023255814,
28
+ "grad_norm": 61.286956787109375,
29
+ "learning_rate": 8.604651162790698e-06,
30
+ "loss": 2.4571,
31
+ "step": 75
32
+ },
33
+ {
34
+ "epoch": 1.0,
35
+ "eval_accuracy": 0.7254901960784313,
36
+ "eval_loss": 0.48061713576316833,
37
+ "eval_macro_f1": 0.42045454545454547,
38
+ "eval_runtime": 0.7673,
39
+ "eval_samples_per_second": 398.802,
40
+ "eval_steps_per_second": 26.065,
41
+ "eval_weighted_f1": 0.6100713012477719,
42
+ "step": 86
43
+ },
44
+ {
45
+ "epoch": 1.1627906976744187,
46
+ "grad_norm": 26.423561096191406,
47
+ "learning_rate": 1.1511627906976746e-05,
48
+ "loss": 2.3328,
49
+ "step": 100
50
+ },
51
+ {
52
+ "epoch": 1.4534883720930232,
53
+ "grad_norm": 60.45857620239258,
54
+ "learning_rate": 1.441860465116279e-05,
55
+ "loss": 1.9427,
56
+ "step": 125
57
+ },
58
+ {
59
+ "epoch": 1.744186046511628,
60
+ "grad_norm": 74.30782318115234,
61
+ "learning_rate": 1.7325581395348837e-05,
62
+ "loss": 1.9428,
63
+ "step": 150
64
+ },
65
+ {
66
+ "epoch": 2.0,
67
+ "eval_accuracy": 0.8562091503267973,
68
+ "eval_loss": 0.33015862107276917,
69
+ "eval_macro_f1": 0.8039375582479031,
70
+ "eval_runtime": 0.663,
71
+ "eval_samples_per_second": 461.567,
72
+ "eval_steps_per_second": 30.168,
73
+ "eval_weighted_f1": 0.8495924931016209,
74
+ "step": 172
75
+ },
76
+ {
77
+ "epoch": 2.0348837209302326,
78
+ "grad_norm": 66.0932846069336,
79
+ "learning_rate": 1.9974160206718347e-05,
80
+ "loss": 1.6777,
81
+ "step": 175
82
+ },
83
+ {
84
+ "epoch": 2.3255813953488373,
85
+ "grad_norm": 61.038414001464844,
86
+ "learning_rate": 1.9651162790697676e-05,
87
+ "loss": 1.667,
88
+ "step": 200
89
+ },
90
+ {
91
+ "epoch": 2.616279069767442,
92
+ "grad_norm": 84.13005828857422,
93
+ "learning_rate": 1.9328165374677004e-05,
94
+ "loss": 1.4979,
95
+ "step": 225
96
+ },
97
+ {
98
+ "epoch": 2.9069767441860463,
99
+ "grad_norm": 40.771522521972656,
100
+ "learning_rate": 1.9005167958656333e-05,
101
+ "loss": 1.4829,
102
+ "step": 250
103
+ },
104
+ {
105
+ "epoch": 3.0,
106
+ "eval_accuracy": 0.7745098039215687,
107
+ "eval_loss": 0.4348524510860443,
108
+ "eval_macro_f1": 0.7546794938826729,
109
+ "eval_runtime": 0.6897,
110
+ "eval_samples_per_second": 443.668,
111
+ "eval_steps_per_second": 28.998,
112
+ "eval_weighted_f1": 0.7861344684271283,
113
+ "step": 258
114
+ },
115
+ {
116
+ "epoch": 3.197674418604651,
117
+ "grad_norm": 63.65130615234375,
118
+ "learning_rate": 1.868217054263566e-05,
119
+ "loss": 1.2801,
120
+ "step": 275
121
+ },
122
+ {
123
+ "epoch": 3.488372093023256,
124
+ "grad_norm": 68.13922119140625,
125
+ "learning_rate": 1.8359173126614988e-05,
126
+ "loss": 1.2032,
127
+ "step": 300
128
+ },
129
+ {
130
+ "epoch": 3.7790697674418605,
131
+ "grad_norm": 73.62340545654297,
132
+ "learning_rate": 1.8036175710594317e-05,
133
+ "loss": 1.3543,
134
+ "step": 325
135
+ },
136
+ {
137
+ "epoch": 4.0,
138
+ "eval_accuracy": 0.8006535947712419,
139
+ "eval_loss": 0.4494991898536682,
140
+ "eval_macro_f1": 0.7740494607255692,
141
+ "eval_runtime": 0.6841,
142
+ "eval_samples_per_second": 447.328,
143
+ "eval_steps_per_second": 29.237,
144
+ "eval_weighted_f1": 0.8090148940427389,
145
+ "step": 344
146
+ },
147
+ {
148
+ "epoch": 4.069767441860465,
149
+ "grad_norm": 28.662683486938477,
150
+ "learning_rate": 1.7713178294573646e-05,
151
+ "loss": 1.135,
152
+ "step": 350
153
+ },
154
+ {
155
+ "epoch": 4.3604651162790695,
156
+ "grad_norm": 68.47048950195312,
157
+ "learning_rate": 1.7390180878552975e-05,
158
+ "loss": 0.8531,
159
+ "step": 375
160
+ },
161
+ {
162
+ "epoch": 4.651162790697675,
163
+ "grad_norm": 73.94048309326172,
164
+ "learning_rate": 1.70671834625323e-05,
165
+ "loss": 0.9591,
166
+ "step": 400
167
+ },
168
+ {
169
+ "epoch": 4.941860465116279,
170
+ "grad_norm": 81.35037994384766,
171
+ "learning_rate": 1.674418604651163e-05,
172
+ "loss": 0.9536,
173
+ "step": 425
174
+ },
175
+ {
176
+ "epoch": 5.0,
177
+ "eval_accuracy": 0.8300653594771242,
178
+ "eval_loss": 0.46315619349479675,
179
+ "eval_macro_f1": 0.8068932038834951,
180
+ "eval_runtime": 0.684,
181
+ "eval_samples_per_second": 447.359,
182
+ "eval_steps_per_second": 29.239,
183
+ "eval_weighted_f1": 0.8370607272035028,
184
+ "step": 430
185
+ },
186
+ {
187
+ "epoch": 5.232558139534884,
188
+ "grad_norm": 36.92371368408203,
189
+ "learning_rate": 1.6421188630490958e-05,
190
+ "loss": 0.719,
191
+ "step": 450
192
+ },
193
+ {
194
+ "epoch": 5.523255813953488,
195
+ "grad_norm": 89.27535247802734,
196
+ "learning_rate": 1.6098191214470284e-05,
197
+ "loss": 0.7397,
198
+ "step": 475
199
+ },
200
+ {
201
+ "epoch": 5.813953488372093,
202
+ "grad_norm": 102.30037689208984,
203
+ "learning_rate": 1.5775193798449613e-05,
204
+ "loss": 0.6656,
205
+ "step": 500
206
+ },
207
+ {
208
+ "epoch": 6.0,
209
+ "eval_accuracy": 0.8725490196078431,
210
+ "eval_loss": 0.4044148027896881,
211
+ "eval_macro_f1": 0.8428641025978643,
212
+ "eval_runtime": 0.6602,
213
+ "eval_samples_per_second": 463.488,
214
+ "eval_steps_per_second": 30.293,
215
+ "eval_weighted_f1": 0.8736649939315265,
216
+ "step": 516
217
+ }
218
+ ],
219
+ "logging_steps": 25,
220
+ "max_steps": 1720,
221
+ "num_input_tokens_seen": 0,
222
+ "num_train_epochs": 20,
223
+ "save_steps": 500,
224
+ "stateful_callbacks": {
225
+ "EarlyStoppingCallback": {
226
+ "args": {
227
+ "early_stopping_patience": 3,
228
+ "early_stopping_threshold": 0.0
229
+ },
230
+ "attributes": {
231
+ "early_stopping_patience_counter": 0
232
+ }
233
+ },
234
+ "TrainerControl": {
235
+ "args": {
236
+ "should_epoch_stop": false,
237
+ "should_evaluate": false,
238
+ "should_log": false,
239
+ "should_save": true,
240
+ "should_training_stop": false
241
+ },
242
+ "attributes": {}
243
+ }
244
+ },
245
+ "total_flos": 0.0,
246
+ "train_batch_size": 8,
247
+ "trial_name": null,
248
+ "trial_params": null
249
+ }
ood-viclickbait2025-xlmr-source-source-06-s44-20260901T142033Z-6863e5c6/trainer/checkpoint-774/special_tokens_map.json ADDED
@@ -0,0 +1,15 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "bos_token": "<s>",
3
+ "cls_token": "<s>",
4
+ "eos_token": "</s>",
5
+ "mask_token": {
6
+ "content": "<mask>",
7
+ "lstrip": true,
8
+ "normalized": false,
9
+ "rstrip": false,
10
+ "single_word": false
11
+ },
12
+ "pad_token": "<pad>",
13
+ "sep_token": "</s>",
14
+ "unk_token": "<unk>"
15
+ }
ood-viclickbait2025-xlmr-source-source-06-s44-20260901T142033Z-6863e5c6/trainer/checkpoint-774/tokenizer_config.json ADDED
@@ -0,0 +1,55 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "added_tokens_decoder": {
3
+ "0": {
4
+ "content": "<s>",
5
+ "lstrip": false,
6
+ "normalized": false,
7
+ "rstrip": false,
8
+ "single_word": false,
9
+ "special": true
10
+ },
11
+ "1": {
12
+ "content": "<pad>",
13
+ "lstrip": false,
14
+ "normalized": false,
15
+ "rstrip": false,
16
+ "single_word": false,
17
+ "special": true
18
+ },
19
+ "2": {
20
+ "content": "</s>",
21
+ "lstrip": false,
22
+ "normalized": false,
23
+ "rstrip": false,
24
+ "single_word": false,
25
+ "special": true
26
+ },
27
+ "3": {
28
+ "content": "<unk>",
29
+ "lstrip": false,
30
+ "normalized": false,
31
+ "rstrip": false,
32
+ "single_word": false,
33
+ "special": true
34
+ },
35
+ "250001": {
36
+ "content": "<mask>",
37
+ "lstrip": true,
38
+ "normalized": false,
39
+ "rstrip": false,
40
+ "single_word": false,
41
+ "special": true
42
+ }
43
+ },
44
+ "bos_token": "<s>",
45
+ "clean_up_tokenization_spaces": false,
46
+ "cls_token": "<s>",
47
+ "eos_token": "</s>",
48
+ "extra_special_tokens": {},
49
+ "mask_token": "<mask>",
50
+ "model_max_length": 512,
51
+ "pad_token": "<pad>",
52
+ "sep_token": "</s>",
53
+ "tokenizer_class": "XLMRobertaTokenizer",
54
+ "unk_token": "<unk>"
55
+ }
ood-viclickbait2025-xlmr-source-source-06-s44-20260901T142033Z-6863e5c6/trainer/checkpoint-774/trainer_state.json ADDED
@@ -0,0 +1,352 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_global_step": 516,
3
+ "best_metric": 0.8428641025978643,
4
+ "best_model_checkpoint": "/root/VNEncoder/results/checkpoints/ood-viclickbait2025-xlmr-source-source-06-s44-20260901T142033Z-6863e5c6/trainer/checkpoint-516",
5
+ "epoch": 9.0,
6
+ "eval_steps": 500,
7
+ "global_step": 774,
8
+ "is_hyper_param_search": false,
9
+ "is_local_process_zero": true,
10
+ "is_world_process_zero": true,
11
+ "log_history": [
12
+ {
13
+ "epoch": 0.29069767441860467,
14
+ "grad_norm": 23.544675827026367,
15
+ "learning_rate": 2.790697674418605e-06,
16
+ "loss": 2.9263,
17
+ "step": 25
18
+ },
19
+ {
20
+ "epoch": 0.5813953488372093,
21
+ "grad_norm": 23.54766845703125,
22
+ "learning_rate": 5.697674418604652e-06,
23
+ "loss": 2.7261,
24
+ "step": 50
25
+ },
26
+ {
27
+ "epoch": 0.872093023255814,
28
+ "grad_norm": 61.286956787109375,
29
+ "learning_rate": 8.604651162790698e-06,
30
+ "loss": 2.4571,
31
+ "step": 75
32
+ },
33
+ {
34
+ "epoch": 1.0,
35
+ "eval_accuracy": 0.7254901960784313,
36
+ "eval_loss": 0.48061713576316833,
37
+ "eval_macro_f1": 0.42045454545454547,
38
+ "eval_runtime": 0.7673,
39
+ "eval_samples_per_second": 398.802,
40
+ "eval_steps_per_second": 26.065,
41
+ "eval_weighted_f1": 0.6100713012477719,
42
+ "step": 86
43
+ },
44
+ {
45
+ "epoch": 1.1627906976744187,
46
+ "grad_norm": 26.423561096191406,
47
+ "learning_rate": 1.1511627906976746e-05,
48
+ "loss": 2.3328,
49
+ "step": 100
50
+ },
51
+ {
52
+ "epoch": 1.4534883720930232,
53
+ "grad_norm": 60.45857620239258,
54
+ "learning_rate": 1.441860465116279e-05,
55
+ "loss": 1.9427,
56
+ "step": 125
57
+ },
58
+ {
59
+ "epoch": 1.744186046511628,
60
+ "grad_norm": 74.30782318115234,
61
+ "learning_rate": 1.7325581395348837e-05,
62
+ "loss": 1.9428,
63
+ "step": 150
64
+ },
65
+ {
66
+ "epoch": 2.0,
67
+ "eval_accuracy": 0.8562091503267973,
68
+ "eval_loss": 0.33015862107276917,
69
+ "eval_macro_f1": 0.8039375582479031,
70
+ "eval_runtime": 0.663,
71
+ "eval_samples_per_second": 461.567,
72
+ "eval_steps_per_second": 30.168,
73
+ "eval_weighted_f1": 0.8495924931016209,
74
+ "step": 172
75
+ },
76
+ {
77
+ "epoch": 2.0348837209302326,
78
+ "grad_norm": 66.0932846069336,
79
+ "learning_rate": 1.9974160206718347e-05,
80
+ "loss": 1.6777,
81
+ "step": 175
82
+ },
83
+ {
84
+ "epoch": 2.3255813953488373,
85
+ "grad_norm": 61.038414001464844,
86
+ "learning_rate": 1.9651162790697676e-05,
87
+ "loss": 1.667,
88
+ "step": 200
89
+ },
90
+ {
91
+ "epoch": 2.616279069767442,
92
+ "grad_norm": 84.13005828857422,
93
+ "learning_rate": 1.9328165374677004e-05,
94
+ "loss": 1.4979,
95
+ "step": 225
96
+ },
97
+ {
98
+ "epoch": 2.9069767441860463,
99
+ "grad_norm": 40.771522521972656,
100
+ "learning_rate": 1.9005167958656333e-05,
101
+ "loss": 1.4829,
102
+ "step": 250
103
+ },
104
+ {
105
+ "epoch": 3.0,
106
+ "eval_accuracy": 0.7745098039215687,
107
+ "eval_loss": 0.4348524510860443,
108
+ "eval_macro_f1": 0.7546794938826729,
109
+ "eval_runtime": 0.6897,
110
+ "eval_samples_per_second": 443.668,
111
+ "eval_steps_per_second": 28.998,
112
+ "eval_weighted_f1": 0.7861344684271283,
113
+ "step": 258
114
+ },
115
+ {
116
+ "epoch": 3.197674418604651,
117
+ "grad_norm": 63.65130615234375,
118
+ "learning_rate": 1.868217054263566e-05,
119
+ "loss": 1.2801,
120
+ "step": 275
121
+ },
122
+ {
123
+ "epoch": 3.488372093023256,
124
+ "grad_norm": 68.13922119140625,
125
+ "learning_rate": 1.8359173126614988e-05,
126
+ "loss": 1.2032,
127
+ "step": 300
128
+ },
129
+ {
130
+ "epoch": 3.7790697674418605,
131
+ "grad_norm": 73.62340545654297,
132
+ "learning_rate": 1.8036175710594317e-05,
133
+ "loss": 1.3543,
134
+ "step": 325
135
+ },
136
+ {
137
+ "epoch": 4.0,
138
+ "eval_accuracy": 0.8006535947712419,
139
+ "eval_loss": 0.4494991898536682,
140
+ "eval_macro_f1": 0.7740494607255692,
141
+ "eval_runtime": 0.6841,
142
+ "eval_samples_per_second": 447.328,
143
+ "eval_steps_per_second": 29.237,
144
+ "eval_weighted_f1": 0.8090148940427389,
145
+ "step": 344
146
+ },
147
+ {
148
+ "epoch": 4.069767441860465,
149
+ "grad_norm": 28.662683486938477,
150
+ "learning_rate": 1.7713178294573646e-05,
151
+ "loss": 1.135,
152
+ "step": 350
153
+ },
154
+ {
155
+ "epoch": 4.3604651162790695,
156
+ "grad_norm": 68.47048950195312,
157
+ "learning_rate": 1.7390180878552975e-05,
158
+ "loss": 0.8531,
159
+ "step": 375
160
+ },
161
+ {
162
+ "epoch": 4.651162790697675,
163
+ "grad_norm": 73.94048309326172,
164
+ "learning_rate": 1.70671834625323e-05,
165
+ "loss": 0.9591,
166
+ "step": 400
167
+ },
168
+ {
169
+ "epoch": 4.941860465116279,
170
+ "grad_norm": 81.35037994384766,
171
+ "learning_rate": 1.674418604651163e-05,
172
+ "loss": 0.9536,
173
+ "step": 425
174
+ },
175
+ {
176
+ "epoch": 5.0,
177
+ "eval_accuracy": 0.8300653594771242,
178
+ "eval_loss": 0.46315619349479675,
179
+ "eval_macro_f1": 0.8068932038834951,
180
+ "eval_runtime": 0.684,
181
+ "eval_samples_per_second": 447.359,
182
+ "eval_steps_per_second": 29.239,
183
+ "eval_weighted_f1": 0.8370607272035028,
184
+ "step": 430
185
+ },
186
+ {
187
+ "epoch": 5.232558139534884,
188
+ "grad_norm": 36.92371368408203,
189
+ "learning_rate": 1.6421188630490958e-05,
190
+ "loss": 0.719,
191
+ "step": 450
192
+ },
193
+ {
194
+ "epoch": 5.523255813953488,
195
+ "grad_norm": 89.27535247802734,
196
+ "learning_rate": 1.6098191214470284e-05,
197
+ "loss": 0.7397,
198
+ "step": 475
199
+ },
200
+ {
201
+ "epoch": 5.813953488372093,
202
+ "grad_norm": 102.30037689208984,
203
+ "learning_rate": 1.5775193798449613e-05,
204
+ "loss": 0.6656,
205
+ "step": 500
206
+ },
207
+ {
208
+ "epoch": 6.0,
209
+ "eval_accuracy": 0.8725490196078431,
210
+ "eval_loss": 0.4044148027896881,
211
+ "eval_macro_f1": 0.8428641025978643,
212
+ "eval_runtime": 0.6602,
213
+ "eval_samples_per_second": 463.488,
214
+ "eval_steps_per_second": 30.293,
215
+ "eval_weighted_f1": 0.8736649939315265,
216
+ "step": 516
217
+ },
218
+ {
219
+ "epoch": 6.104651162790698,
220
+ "grad_norm": 104.36873626708984,
221
+ "learning_rate": 1.545219638242894e-05,
222
+ "loss": 0.6109,
223
+ "step": 525
224
+ },
225
+ {
226
+ "epoch": 6.395348837209302,
227
+ "grad_norm": 58.92158889770508,
228
+ "learning_rate": 1.5129198966408269e-05,
229
+ "loss": 0.3602,
230
+ "step": 550
231
+ },
232
+ {
233
+ "epoch": 6.686046511627907,
234
+ "grad_norm": 34.21396255493164,
235
+ "learning_rate": 1.4806201550387598e-05,
236
+ "loss": 0.4644,
237
+ "step": 575
238
+ },
239
+ {
240
+ "epoch": 6.976744186046512,
241
+ "grad_norm": 64.25516510009766,
242
+ "learning_rate": 1.4483204134366927e-05,
243
+ "loss": 0.6714,
244
+ "step": 600
245
+ },
246
+ {
247
+ "epoch": 7.0,
248
+ "eval_accuracy": 0.8594771241830066,
249
+ "eval_loss": 0.46995070576667786,
250
+ "eval_macro_f1": 0.8372883871047523,
251
+ "eval_runtime": 0.652,
252
+ "eval_samples_per_second": 469.33,
253
+ "eval_steps_per_second": 30.675,
254
+ "eval_weighted_f1": 0.8643861368109389,
255
+ "step": 602
256
+ },
257
+ {
258
+ "epoch": 7.267441860465116,
259
+ "grad_norm": 77.10468292236328,
260
+ "learning_rate": 1.4160206718346254e-05,
261
+ "loss": 0.2013,
262
+ "step": 625
263
+ },
264
+ {
265
+ "epoch": 7.558139534883721,
266
+ "grad_norm": 73.71111297607422,
267
+ "learning_rate": 1.3837209302325583e-05,
268
+ "loss": 0.4848,
269
+ "step": 650
270
+ },
271
+ {
272
+ "epoch": 7.848837209302325,
273
+ "grad_norm": 16.351274490356445,
274
+ "learning_rate": 1.3514211886304912e-05,
275
+ "loss": 0.367,
276
+ "step": 675
277
+ },
278
+ {
279
+ "epoch": 8.0,
280
+ "eval_accuracy": 0.8627450980392157,
281
+ "eval_loss": 0.6721513867378235,
282
+ "eval_macro_f1": 0.837800999545661,
283
+ "eval_runtime": 0.7237,
284
+ "eval_samples_per_second": 422.832,
285
+ "eval_steps_per_second": 27.636,
286
+ "eval_weighted_f1": 0.8664867128132488,
287
+ "step": 688
288
+ },
289
+ {
290
+ "epoch": 8.13953488372093,
291
+ "grad_norm": 162.64071655273438,
292
+ "learning_rate": 1.3191214470284239e-05,
293
+ "loss": 0.2724,
294
+ "step": 700
295
+ },
296
+ {
297
+ "epoch": 8.430232558139535,
298
+ "grad_norm": 18.183868408203125,
299
+ "learning_rate": 1.2868217054263568e-05,
300
+ "loss": 0.1065,
301
+ "step": 725
302
+ },
303
+ {
304
+ "epoch": 8.720930232558139,
305
+ "grad_norm": 96.52957916259766,
306
+ "learning_rate": 1.2545219638242895e-05,
307
+ "loss": 0.6261,
308
+ "step": 750
309
+ },
310
+ {
311
+ "epoch": 9.0,
312
+ "eval_accuracy": 0.8398692810457516,
313
+ "eval_loss": 0.9877539873123169,
314
+ "eval_macro_f1": 0.8184987471402114,
315
+ "eval_runtime": 0.7388,
316
+ "eval_samples_per_second": 414.212,
317
+ "eval_steps_per_second": 27.073,
318
+ "eval_weighted_f1": 0.8465857345589215,
319
+ "step": 774
320
+ }
321
+ ],
322
+ "logging_steps": 25,
323
+ "max_steps": 1720,
324
+ "num_input_tokens_seen": 0,
325
+ "num_train_epochs": 20,
326
+ "save_steps": 500,
327
+ "stateful_callbacks": {
328
+ "EarlyStoppingCallback": {
329
+ "args": {
330
+ "early_stopping_patience": 3,
331
+ "early_stopping_threshold": 0.0
332
+ },
333
+ "attributes": {
334
+ "early_stopping_patience_counter": 3
335
+ }
336
+ },
337
+ "TrainerControl": {
338
+ "args": {
339
+ "should_epoch_stop": false,
340
+ "should_evaluate": false,
341
+ "should_log": false,
342
+ "should_save": true,
343
+ "should_training_stop": true
344
+ },
345
+ "attributes": {}
346
+ }
347
+ },
348
+ "total_flos": 0.0,
349
+ "train_batch_size": 8,
350
+ "trial_name": null,
351
+ "trial_params": null
352
+ }
ood-viclickbait2025-xlmr-source-source-07-s44-20260901T142547Z-6863e5c6/trainer/checkpoint-711/special_tokens_map.json ADDED
@@ -0,0 +1,15 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "bos_token": "<s>",
3
+ "cls_token": "<s>",
4
+ "eos_token": "</s>",
5
+ "mask_token": {
6
+ "content": "<mask>",
7
+ "lstrip": true,
8
+ "normalized": false,
9
+ "rstrip": false,
10
+ "single_word": false
11
+ },
12
+ "pad_token": "<pad>",
13
+ "sep_token": "</s>",
14
+ "unk_token": "<unk>"
15
+ }
ood-viclickbait2025-xlmr-source-source-07-s44-20260901T142547Z-6863e5c6/trainer/checkpoint-711/tokenizer_config.json ADDED
@@ -0,0 +1,55 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "added_tokens_decoder": {
3
+ "0": {
4
+ "content": "<s>",
5
+ "lstrip": false,
6
+ "normalized": false,
7
+ "rstrip": false,
8
+ "single_word": false,
9
+ "special": true
10
+ },
11
+ "1": {
12
+ "content": "<pad>",
13
+ "lstrip": false,
14
+ "normalized": false,
15
+ "rstrip": false,
16
+ "single_word": false,
17
+ "special": true
18
+ },
19
+ "2": {
20
+ "content": "</s>",
21
+ "lstrip": false,
22
+ "normalized": false,
23
+ "rstrip": false,
24
+ "single_word": false,
25
+ "special": true
26
+ },
27
+ "3": {
28
+ "content": "<unk>",
29
+ "lstrip": false,
30
+ "normalized": false,
31
+ "rstrip": false,
32
+ "single_word": false,
33
+ "special": true
34
+ },
35
+ "250001": {
36
+ "content": "<mask>",
37
+ "lstrip": true,
38
+ "normalized": false,
39
+ "rstrip": false,
40
+ "single_word": false,
41
+ "special": true
42
+ }
43
+ },
44
+ "bos_token": "<s>",
45
+ "clean_up_tokenization_spaces": false,
46
+ "cls_token": "<s>",
47
+ "eos_token": "</s>",
48
+ "extra_special_tokens": {},
49
+ "mask_token": "<mask>",
50
+ "model_max_length": 512,
51
+ "pad_token": "<pad>",
52
+ "sep_token": "</s>",
53
+ "tokenizer_class": "XLMRobertaTokenizer",
54
+ "unk_token": "<unk>"
55
+ }
ood-viclickbait2025-xlmr-source-source-07-s44-20260901T142547Z-6863e5c6/trainer/checkpoint-711/trainer_state.json ADDED
@@ -0,0 +1,338 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_global_step": 711,
3
+ "best_metric": 0.8489247311827957,
4
+ "best_model_checkpoint": "/root/VNEncoder/results/checkpoints/ood-viclickbait2025-xlmr-source-source-07-s44-20260901T142547Z-6863e5c6/trainer/checkpoint-711",
5
+ "epoch": 9.0,
6
+ "eval_steps": 500,
7
+ "global_step": 711,
8
+ "is_hyper_param_search": false,
9
+ "is_local_process_zero": true,
10
+ "is_world_process_zero": true,
11
+ "log_history": [
12
+ {
13
+ "epoch": 0.31645569620253167,
14
+ "grad_norm": 10.859970092773438,
15
+ "learning_rate": 3.037974683544304e-06,
16
+ "loss": 2.894,
17
+ "step": 25
18
+ },
19
+ {
20
+ "epoch": 0.6329113924050633,
21
+ "grad_norm": 30.539888381958008,
22
+ "learning_rate": 6.20253164556962e-06,
23
+ "loss": 2.7578,
24
+ "step": 50
25
+ },
26
+ {
27
+ "epoch": 0.9493670886075949,
28
+ "grad_norm": 17.019699096679688,
29
+ "learning_rate": 9.367088607594937e-06,
30
+ "loss": 2.5701,
31
+ "step": 75
32
+ },
33
+ {
34
+ "epoch": 1.0,
35
+ "eval_accuracy": 0.6690391459074733,
36
+ "eval_loss": 0.5780642032623291,
37
+ "eval_macro_f1": 0.40085287846481876,
38
+ "eval_runtime": 0.6537,
39
+ "eval_samples_per_second": 429.883,
40
+ "eval_steps_per_second": 27.537,
41
+ "eval_weighted_f1": 0.5363725348853091,
42
+ "step": 79
43
+ },
44
+ {
45
+ "epoch": 1.2658227848101267,
46
+ "grad_norm": 89.87989807128906,
47
+ "learning_rate": 1.2531645569620255e-05,
48
+ "loss": 2.2146,
49
+ "step": 100
50
+ },
51
+ {
52
+ "epoch": 1.5822784810126582,
53
+ "grad_norm": 57.168434143066406,
54
+ "learning_rate": 1.5696202531645572e-05,
55
+ "loss": 2.1326,
56
+ "step": 125
57
+ },
58
+ {
59
+ "epoch": 1.8987341772151898,
60
+ "grad_norm": 90.69680786132812,
61
+ "learning_rate": 1.886075949367089e-05,
62
+ "loss": 1.9353,
63
+ "step": 150
64
+ },
65
+ {
66
+ "epoch": 2.0,
67
+ "eval_accuracy": 0.800711743772242,
68
+ "eval_loss": 0.46159303188323975,
69
+ "eval_macro_f1": 0.7854259845096542,
70
+ "eval_runtime": 0.6264,
71
+ "eval_samples_per_second": 448.565,
72
+ "eval_steps_per_second": 28.734,
73
+ "eval_weighted_f1": 0.8047879462422654,
74
+ "step": 158
75
+ },
76
+ {
77
+ "epoch": 2.2151898734177213,
78
+ "grad_norm": 27.01743507385254,
79
+ "learning_rate": 1.977496483825598e-05,
80
+ "loss": 1.7965,
81
+ "step": 175
82
+ },
83
+ {
84
+ "epoch": 2.5316455696202533,
85
+ "grad_norm": 53.67705535888672,
86
+ "learning_rate": 1.9423347398030942e-05,
87
+ "loss": 1.4839,
88
+ "step": 200
89
+ },
90
+ {
91
+ "epoch": 2.848101265822785,
92
+ "grad_norm": 53.62406921386719,
93
+ "learning_rate": 1.907172995780591e-05,
94
+ "loss": 1.6328,
95
+ "step": 225
96
+ },
97
+ {
98
+ "epoch": 3.0,
99
+ "eval_accuracy": 0.8434163701067615,
100
+ "eval_loss": 0.3271748423576355,
101
+ "eval_macro_f1": 0.8212261422787739,
102
+ "eval_runtime": 0.6805,
103
+ "eval_samples_per_second": 412.955,
104
+ "eval_steps_per_second": 26.453,
105
+ "eval_weighted_f1": 0.8425197952450246,
106
+ "step": 237
107
+ },
108
+ {
109
+ "epoch": 3.1645569620253164,
110
+ "grad_norm": 25.777297973632812,
111
+ "learning_rate": 1.872011251758087e-05,
112
+ "loss": 1.3556,
113
+ "step": 250
114
+ },
115
+ {
116
+ "epoch": 3.481012658227848,
117
+ "grad_norm": 45.83961868286133,
118
+ "learning_rate": 1.8368495077355838e-05,
119
+ "loss": 1.2223,
120
+ "step": 275
121
+ },
122
+ {
123
+ "epoch": 3.7974683544303796,
124
+ "grad_norm": 45.81952667236328,
125
+ "learning_rate": 1.8016877637130804e-05,
126
+ "loss": 1.3262,
127
+ "step": 300
128
+ },
129
+ {
130
+ "epoch": 4.0,
131
+ "eval_accuracy": 0.8220640569395018,
132
+ "eval_loss": 0.40762826800346375,
133
+ "eval_macro_f1": 0.8113384896336878,
134
+ "eval_runtime": 0.6108,
135
+ "eval_samples_per_second": 460.026,
136
+ "eval_steps_per_second": 29.468,
137
+ "eval_weighted_f1": 0.8265463835747672,
138
+ "step": 316
139
+ },
140
+ {
141
+ "epoch": 4.113924050632911,
142
+ "grad_norm": 27.679380416870117,
143
+ "learning_rate": 1.7665260196905767e-05,
144
+ "loss": 1.1138,
145
+ "step": 325
146
+ },
147
+ {
148
+ "epoch": 4.430379746835443,
149
+ "grad_norm": 28.408794403076172,
150
+ "learning_rate": 1.7313642756680733e-05,
151
+ "loss": 1.0092,
152
+ "step": 350
153
+ },
154
+ {
155
+ "epoch": 4.746835443037975,
156
+ "grad_norm": 65.93917846679688,
157
+ "learning_rate": 1.6962025316455696e-05,
158
+ "loss": 1.0344,
159
+ "step": 375
160
+ },
161
+ {
162
+ "epoch": 5.0,
163
+ "eval_accuracy": 0.8576512455516014,
164
+ "eval_loss": 0.407770574092865,
165
+ "eval_macro_f1": 0.8433143749303,
166
+ "eval_runtime": 0.6449,
167
+ "eval_samples_per_second": 435.704,
168
+ "eval_steps_per_second": 27.91,
169
+ "eval_weighted_f1": 0.859337936212931,
170
+ "step": 395
171
+ },
172
+ {
173
+ "epoch": 5.063291139240507,
174
+ "grad_norm": 81.56014251708984,
175
+ "learning_rate": 1.6610407876230663e-05,
176
+ "loss": 0.6673,
177
+ "step": 400
178
+ },
179
+ {
180
+ "epoch": 5.379746835443038,
181
+ "grad_norm": 225.88876342773438,
182
+ "learning_rate": 1.6258790436005626e-05,
183
+ "loss": 0.5635,
184
+ "step": 425
185
+ },
186
+ {
187
+ "epoch": 5.69620253164557,
188
+ "grad_norm": 28.108448028564453,
189
+ "learning_rate": 1.5907172995780592e-05,
190
+ "loss": 0.7501,
191
+ "step": 450
192
+ },
193
+ {
194
+ "epoch": 6.0,
195
+ "eval_accuracy": 0.8612099644128114,
196
+ "eval_loss": 0.47222700715065,
197
+ "eval_macro_f1": 0.8445465764500617,
198
+ "eval_runtime": 0.6649,
199
+ "eval_samples_per_second": 422.631,
200
+ "eval_steps_per_second": 27.072,
201
+ "eval_weighted_f1": 0.8617533357594228,
202
+ "step": 474
203
+ },
204
+ {
205
+ "epoch": 6.012658227848101,
206
+ "grad_norm": 30.494224548339844,
207
+ "learning_rate": 1.555555555555556e-05,
208
+ "loss": 0.7262,
209
+ "step": 475
210
+ },
211
+ {
212
+ "epoch": 6.329113924050633,
213
+ "grad_norm": 68.9117202758789,
214
+ "learning_rate": 1.5203938115330521e-05,
215
+ "loss": 0.4915,
216
+ "step": 500
217
+ },
218
+ {
219
+ "epoch": 6.6455696202531644,
220
+ "grad_norm": 126.9439468383789,
221
+ "learning_rate": 1.4852320675105488e-05,
222
+ "loss": 0.4423,
223
+ "step": 525
224
+ },
225
+ {
226
+ "epoch": 6.962025316455696,
227
+ "grad_norm": 60.305728912353516,
228
+ "learning_rate": 1.450070323488045e-05,
229
+ "loss": 0.4613,
230
+ "step": 550
231
+ },
232
+ {
233
+ "epoch": 7.0,
234
+ "eval_accuracy": 0.8612099644128114,
235
+ "eval_loss": 0.5797529220581055,
236
+ "eval_macro_f1": 0.8437219251336898,
237
+ "eval_runtime": 0.6399,
238
+ "eval_samples_per_second": 439.105,
239
+ "eval_steps_per_second": 28.128,
240
+ "eval_weighted_f1": 0.8613960073838659,
241
+ "step": 553
242
+ },
243
+ {
244
+ "epoch": 7.2784810126582276,
245
+ "grad_norm": 220.0026397705078,
246
+ "learning_rate": 1.4149085794655415e-05,
247
+ "loss": 0.2764,
248
+ "step": 575
249
+ },
250
+ {
251
+ "epoch": 7.594936708860759,
252
+ "grad_norm": 68.6791763305664,
253
+ "learning_rate": 1.3797468354430381e-05,
254
+ "loss": 0.5405,
255
+ "step": 600
256
+ },
257
+ {
258
+ "epoch": 7.911392405063291,
259
+ "grad_norm": 381.5016174316406,
260
+ "learning_rate": 1.3445850914205344e-05,
261
+ "loss": 0.3258,
262
+ "step": 625
263
+ },
264
+ {
265
+ "epoch": 8.0,
266
+ "eval_accuracy": 0.8469750889679716,
267
+ "eval_loss": 0.8536282777786255,
268
+ "eval_macro_f1": 0.8303354536136035,
269
+ "eval_runtime": 0.6626,
270
+ "eval_samples_per_second": 424.085,
271
+ "eval_steps_per_second": 27.166,
272
+ "eval_weighted_f1": 0.8482986963257053,
273
+ "step": 632
274
+ },
275
+ {
276
+ "epoch": 8.227848101265822,
277
+ "grad_norm": 131.03399658203125,
278
+ "learning_rate": 1.309423347398031e-05,
279
+ "loss": 0.2809,
280
+ "step": 650
281
+ },
282
+ {
283
+ "epoch": 8.544303797468354,
284
+ "grad_norm": 273.69573974609375,
285
+ "learning_rate": 1.2742616033755275e-05,
286
+ "loss": 0.171,
287
+ "step": 675
288
+ },
289
+ {
290
+ "epoch": 8.860759493670885,
291
+ "grad_norm": 287.41156005859375,
292
+ "learning_rate": 1.2390998593530238e-05,
293
+ "loss": 0.3972,
294
+ "step": 700
295
+ },
296
+ {
297
+ "epoch": 9.0,
298
+ "eval_accuracy": 0.8647686832740213,
299
+ "eval_loss": 0.957329273223877,
300
+ "eval_macro_f1": 0.8489247311827957,
301
+ "eval_runtime": 0.6499,
302
+ "eval_samples_per_second": 432.37,
303
+ "eval_steps_per_second": 27.696,
304
+ "eval_weighted_f1": 0.8654651207285807,
305
+ "step": 711
306
+ }
307
+ ],
308
+ "logging_steps": 25,
309
+ "max_steps": 1580,
310
+ "num_input_tokens_seen": 0,
311
+ "num_train_epochs": 20,
312
+ "save_steps": 500,
313
+ "stateful_callbacks": {
314
+ "EarlyStoppingCallback": {
315
+ "args": {
316
+ "early_stopping_patience": 3,
317
+ "early_stopping_threshold": 0.0
318
+ },
319
+ "attributes": {
320
+ "early_stopping_patience_counter": 0
321
+ }
322
+ },
323
+ "TrainerControl": {
324
+ "args": {
325
+ "should_epoch_stop": false,
326
+ "should_evaluate": false,
327
+ "should_log": false,
328
+ "should_save": true,
329
+ "should_training_stop": false
330
+ },
331
+ "attributes": {}
332
+ }
333
+ },
334
+ "total_flos": 0.0,
335
+ "train_batch_size": 8,
336
+ "trial_name": null,
337
+ "trial_params": null
338
+ }
ood-viclickbait2025-xlmr-source-source-07-s44-20260901T142547Z-6863e5c6/trainer/checkpoint-948/special_tokens_map.json ADDED
@@ -0,0 +1,15 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "bos_token": "<s>",
3
+ "cls_token": "<s>",
4
+ "eos_token": "</s>",
5
+ "mask_token": {
6
+ "content": "<mask>",
7
+ "lstrip": true,
8
+ "normalized": false,
9
+ "rstrip": false,
10
+ "single_word": false
11
+ },
12
+ "pad_token": "<pad>",
13
+ "sep_token": "</s>",
14
+ "unk_token": "<unk>"
15
+ }
ood-viclickbait2025-xlmr-source-source-07-s44-20260901T142547Z-6863e5c6/trainer/checkpoint-948/tokenizer_config.json ADDED
@@ -0,0 +1,55 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "added_tokens_decoder": {
3
+ "0": {
4
+ "content": "<s>",
5
+ "lstrip": false,
6
+ "normalized": false,
7
+ "rstrip": false,
8
+ "single_word": false,
9
+ "special": true
10
+ },
11
+ "1": {
12
+ "content": "<pad>",
13
+ "lstrip": false,
14
+ "normalized": false,
15
+ "rstrip": false,
16
+ "single_word": false,
17
+ "special": true
18
+ },
19
+ "2": {
20
+ "content": "</s>",
21
+ "lstrip": false,
22
+ "normalized": false,
23
+ "rstrip": false,
24
+ "single_word": false,
25
+ "special": true
26
+ },
27
+ "3": {
28
+ "content": "<unk>",
29
+ "lstrip": false,
30
+ "normalized": false,
31
+ "rstrip": false,
32
+ "single_word": false,
33
+ "special": true
34
+ },
35
+ "250001": {
36
+ "content": "<mask>",
37
+ "lstrip": true,
38
+ "normalized": false,
39
+ "rstrip": false,
40
+ "single_word": false,
41
+ "special": true
42
+ }
43
+ },
44
+ "bos_token": "<s>",
45
+ "clean_up_tokenization_spaces": false,
46
+ "cls_token": "<s>",
47
+ "eos_token": "</s>",
48
+ "extra_special_tokens": {},
49
+ "mask_token": "<mask>",
50
+ "model_max_length": 512,
51
+ "pad_token": "<pad>",
52
+ "sep_token": "</s>",
53
+ "tokenizer_class": "XLMRobertaTokenizer",
54
+ "unk_token": "<unk>"
55
+ }
ood-viclickbait2025-xlmr-source-source-07-s44-20260901T142547Z-6863e5c6/trainer/checkpoint-948/trainer_state.json ADDED
@@ -0,0 +1,434 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_global_step": 711,
3
+ "best_metric": 0.8489247311827957,
4
+ "best_model_checkpoint": "/root/VNEncoder/results/checkpoints/ood-viclickbait2025-xlmr-source-source-07-s44-20260901T142547Z-6863e5c6/trainer/checkpoint-711",
5
+ "epoch": 12.0,
6
+ "eval_steps": 500,
7
+ "global_step": 948,
8
+ "is_hyper_param_search": false,
9
+ "is_local_process_zero": true,
10
+ "is_world_process_zero": true,
11
+ "log_history": [
12
+ {
13
+ "epoch": 0.31645569620253167,
14
+ "grad_norm": 10.859970092773438,
15
+ "learning_rate": 3.037974683544304e-06,
16
+ "loss": 2.894,
17
+ "step": 25
18
+ },
19
+ {
20
+ "epoch": 0.6329113924050633,
21
+ "grad_norm": 30.539888381958008,
22
+ "learning_rate": 6.20253164556962e-06,
23
+ "loss": 2.7578,
24
+ "step": 50
25
+ },
26
+ {
27
+ "epoch": 0.9493670886075949,
28
+ "grad_norm": 17.019699096679688,
29
+ "learning_rate": 9.367088607594937e-06,
30
+ "loss": 2.5701,
31
+ "step": 75
32
+ },
33
+ {
34
+ "epoch": 1.0,
35
+ "eval_accuracy": 0.6690391459074733,
36
+ "eval_loss": 0.5780642032623291,
37
+ "eval_macro_f1": 0.40085287846481876,
38
+ "eval_runtime": 0.6537,
39
+ "eval_samples_per_second": 429.883,
40
+ "eval_steps_per_second": 27.537,
41
+ "eval_weighted_f1": 0.5363725348853091,
42
+ "step": 79
43
+ },
44
+ {
45
+ "epoch": 1.2658227848101267,
46
+ "grad_norm": 89.87989807128906,
47
+ "learning_rate": 1.2531645569620255e-05,
48
+ "loss": 2.2146,
49
+ "step": 100
50
+ },
51
+ {
52
+ "epoch": 1.5822784810126582,
53
+ "grad_norm": 57.168434143066406,
54
+ "learning_rate": 1.5696202531645572e-05,
55
+ "loss": 2.1326,
56
+ "step": 125
57
+ },
58
+ {
59
+ "epoch": 1.8987341772151898,
60
+ "grad_norm": 90.69680786132812,
61
+ "learning_rate": 1.886075949367089e-05,
62
+ "loss": 1.9353,
63
+ "step": 150
64
+ },
65
+ {
66
+ "epoch": 2.0,
67
+ "eval_accuracy": 0.800711743772242,
68
+ "eval_loss": 0.46159303188323975,
69
+ "eval_macro_f1": 0.7854259845096542,
70
+ "eval_runtime": 0.6264,
71
+ "eval_samples_per_second": 448.565,
72
+ "eval_steps_per_second": 28.734,
73
+ "eval_weighted_f1": 0.8047879462422654,
74
+ "step": 158
75
+ },
76
+ {
77
+ "epoch": 2.2151898734177213,
78
+ "grad_norm": 27.01743507385254,
79
+ "learning_rate": 1.977496483825598e-05,
80
+ "loss": 1.7965,
81
+ "step": 175
82
+ },
83
+ {
84
+ "epoch": 2.5316455696202533,
85
+ "grad_norm": 53.67705535888672,
86
+ "learning_rate": 1.9423347398030942e-05,
87
+ "loss": 1.4839,
88
+ "step": 200
89
+ },
90
+ {
91
+ "epoch": 2.848101265822785,
92
+ "grad_norm": 53.62406921386719,
93
+ "learning_rate": 1.907172995780591e-05,
94
+ "loss": 1.6328,
95
+ "step": 225
96
+ },
97
+ {
98
+ "epoch": 3.0,
99
+ "eval_accuracy": 0.8434163701067615,
100
+ "eval_loss": 0.3271748423576355,
101
+ "eval_macro_f1": 0.8212261422787739,
102
+ "eval_runtime": 0.6805,
103
+ "eval_samples_per_second": 412.955,
104
+ "eval_steps_per_second": 26.453,
105
+ "eval_weighted_f1": 0.8425197952450246,
106
+ "step": 237
107
+ },
108
+ {
109
+ "epoch": 3.1645569620253164,
110
+ "grad_norm": 25.777297973632812,
111
+ "learning_rate": 1.872011251758087e-05,
112
+ "loss": 1.3556,
113
+ "step": 250
114
+ },
115
+ {
116
+ "epoch": 3.481012658227848,
117
+ "grad_norm": 45.83961868286133,
118
+ "learning_rate": 1.8368495077355838e-05,
119
+ "loss": 1.2223,
120
+ "step": 275
121
+ },
122
+ {
123
+ "epoch": 3.7974683544303796,
124
+ "grad_norm": 45.81952667236328,
125
+ "learning_rate": 1.8016877637130804e-05,
126
+ "loss": 1.3262,
127
+ "step": 300
128
+ },
129
+ {
130
+ "epoch": 4.0,
131
+ "eval_accuracy": 0.8220640569395018,
132
+ "eval_loss": 0.40762826800346375,
133
+ "eval_macro_f1": 0.8113384896336878,
134
+ "eval_runtime": 0.6108,
135
+ "eval_samples_per_second": 460.026,
136
+ "eval_steps_per_second": 29.468,
137
+ "eval_weighted_f1": 0.8265463835747672,
138
+ "step": 316
139
+ },
140
+ {
141
+ "epoch": 4.113924050632911,
142
+ "grad_norm": 27.679380416870117,
143
+ "learning_rate": 1.7665260196905767e-05,
144
+ "loss": 1.1138,
145
+ "step": 325
146
+ },
147
+ {
148
+ "epoch": 4.430379746835443,
149
+ "grad_norm": 28.408794403076172,
150
+ "learning_rate": 1.7313642756680733e-05,
151
+ "loss": 1.0092,
152
+ "step": 350
153
+ },
154
+ {
155
+ "epoch": 4.746835443037975,
156
+ "grad_norm": 65.93917846679688,
157
+ "learning_rate": 1.6962025316455696e-05,
158
+ "loss": 1.0344,
159
+ "step": 375
160
+ },
161
+ {
162
+ "epoch": 5.0,
163
+ "eval_accuracy": 0.8576512455516014,
164
+ "eval_loss": 0.407770574092865,
165
+ "eval_macro_f1": 0.8433143749303,
166
+ "eval_runtime": 0.6449,
167
+ "eval_samples_per_second": 435.704,
168
+ "eval_steps_per_second": 27.91,
169
+ "eval_weighted_f1": 0.859337936212931,
170
+ "step": 395
171
+ },
172
+ {
173
+ "epoch": 5.063291139240507,
174
+ "grad_norm": 81.56014251708984,
175
+ "learning_rate": 1.6610407876230663e-05,
176
+ "loss": 0.6673,
177
+ "step": 400
178
+ },
179
+ {
180
+ "epoch": 5.379746835443038,
181
+ "grad_norm": 225.88876342773438,
182
+ "learning_rate": 1.6258790436005626e-05,
183
+ "loss": 0.5635,
184
+ "step": 425
185
+ },
186
+ {
187
+ "epoch": 5.69620253164557,
188
+ "grad_norm": 28.108448028564453,
189
+ "learning_rate": 1.5907172995780592e-05,
190
+ "loss": 0.7501,
191
+ "step": 450
192
+ },
193
+ {
194
+ "epoch": 6.0,
195
+ "eval_accuracy": 0.8612099644128114,
196
+ "eval_loss": 0.47222700715065,
197
+ "eval_macro_f1": 0.8445465764500617,
198
+ "eval_runtime": 0.6649,
199
+ "eval_samples_per_second": 422.631,
200
+ "eval_steps_per_second": 27.072,
201
+ "eval_weighted_f1": 0.8617533357594228,
202
+ "step": 474
203
+ },
204
+ {
205
+ "epoch": 6.012658227848101,
206
+ "grad_norm": 30.494224548339844,
207
+ "learning_rate": 1.555555555555556e-05,
208
+ "loss": 0.7262,
209
+ "step": 475
210
+ },
211
+ {
212
+ "epoch": 6.329113924050633,
213
+ "grad_norm": 68.9117202758789,
214
+ "learning_rate": 1.5203938115330521e-05,
215
+ "loss": 0.4915,
216
+ "step": 500
217
+ },
218
+ {
219
+ "epoch": 6.6455696202531644,
220
+ "grad_norm": 126.9439468383789,
221
+ "learning_rate": 1.4852320675105488e-05,
222
+ "loss": 0.4423,
223
+ "step": 525
224
+ },
225
+ {
226
+ "epoch": 6.962025316455696,
227
+ "grad_norm": 60.305728912353516,
228
+ "learning_rate": 1.450070323488045e-05,
229
+ "loss": 0.4613,
230
+ "step": 550
231
+ },
232
+ {
233
+ "epoch": 7.0,
234
+ "eval_accuracy": 0.8612099644128114,
235
+ "eval_loss": 0.5797529220581055,
236
+ "eval_macro_f1": 0.8437219251336898,
237
+ "eval_runtime": 0.6399,
238
+ "eval_samples_per_second": 439.105,
239
+ "eval_steps_per_second": 28.128,
240
+ "eval_weighted_f1": 0.8613960073838659,
241
+ "step": 553
242
+ },
243
+ {
244
+ "epoch": 7.2784810126582276,
245
+ "grad_norm": 220.0026397705078,
246
+ "learning_rate": 1.4149085794655415e-05,
247
+ "loss": 0.2764,
248
+ "step": 575
249
+ },
250
+ {
251
+ "epoch": 7.594936708860759,
252
+ "grad_norm": 68.6791763305664,
253
+ "learning_rate": 1.3797468354430381e-05,
254
+ "loss": 0.5405,
255
+ "step": 600
256
+ },
257
+ {
258
+ "epoch": 7.911392405063291,
259
+ "grad_norm": 381.5016174316406,
260
+ "learning_rate": 1.3445850914205344e-05,
261
+ "loss": 0.3258,
262
+ "step": 625
263
+ },
264
+ {
265
+ "epoch": 8.0,
266
+ "eval_accuracy": 0.8469750889679716,
267
+ "eval_loss": 0.8536282777786255,
268
+ "eval_macro_f1": 0.8303354536136035,
269
+ "eval_runtime": 0.6626,
270
+ "eval_samples_per_second": 424.085,
271
+ "eval_steps_per_second": 27.166,
272
+ "eval_weighted_f1": 0.8482986963257053,
273
+ "step": 632
274
+ },
275
+ {
276
+ "epoch": 8.227848101265822,
277
+ "grad_norm": 131.03399658203125,
278
+ "learning_rate": 1.309423347398031e-05,
279
+ "loss": 0.2809,
280
+ "step": 650
281
+ },
282
+ {
283
+ "epoch": 8.544303797468354,
284
+ "grad_norm": 273.69573974609375,
285
+ "learning_rate": 1.2742616033755275e-05,
286
+ "loss": 0.171,
287
+ "step": 675
288
+ },
289
+ {
290
+ "epoch": 8.860759493670885,
291
+ "grad_norm": 287.41156005859375,
292
+ "learning_rate": 1.2390998593530238e-05,
293
+ "loss": 0.3972,
294
+ "step": 700
295
+ },
296
+ {
297
+ "epoch": 9.0,
298
+ "eval_accuracy": 0.8647686832740213,
299
+ "eval_loss": 0.957329273223877,
300
+ "eval_macro_f1": 0.8489247311827957,
301
+ "eval_runtime": 0.6499,
302
+ "eval_samples_per_second": 432.37,
303
+ "eval_steps_per_second": 27.696,
304
+ "eval_weighted_f1": 0.8654651207285807,
305
+ "step": 711
306
+ },
307
+ {
308
+ "epoch": 9.177215189873417,
309
+ "grad_norm": 204.0855255126953,
310
+ "learning_rate": 1.2039381153305205e-05,
311
+ "loss": 0.4871,
312
+ "step": 725
313
+ },
314
+ {
315
+ "epoch": 9.49367088607595,
316
+ "grad_norm": 0.08458869904279709,
317
+ "learning_rate": 1.168776371308017e-05,
318
+ "loss": 0.1852,
319
+ "step": 750
320
+ },
321
+ {
322
+ "epoch": 9.810126582278482,
323
+ "grad_norm": 0.16510236263275146,
324
+ "learning_rate": 1.1336146272855136e-05,
325
+ "loss": 0.2526,
326
+ "step": 775
327
+ },
328
+ {
329
+ "epoch": 10.0,
330
+ "eval_accuracy": 0.8612099644128114,
331
+ "eval_loss": 1.0146349668502808,
332
+ "eval_macro_f1": 0.8360927895185534,
333
+ "eval_runtime": 0.6529,
334
+ "eval_samples_per_second": 430.404,
335
+ "eval_steps_per_second": 27.57,
336
+ "eval_weighted_f1": 0.8577848951090489,
337
+ "step": 790
338
+ },
339
+ {
340
+ "epoch": 10.126582278481013,
341
+ "grad_norm": 285.2706298828125,
342
+ "learning_rate": 1.0984528832630099e-05,
343
+ "loss": 0.3189,
344
+ "step": 800
345
+ },
346
+ {
347
+ "epoch": 10.443037974683545,
348
+ "grad_norm": 74.95194244384766,
349
+ "learning_rate": 1.0632911392405063e-05,
350
+ "loss": 0.0864,
351
+ "step": 825
352
+ },
353
+ {
354
+ "epoch": 10.759493670886076,
355
+ "grad_norm": 0.11407579481601715,
356
+ "learning_rate": 1.028129395218003e-05,
357
+ "loss": 0.2649,
358
+ "step": 850
359
+ },
360
+ {
361
+ "epoch": 11.0,
362
+ "eval_accuracy": 0.8469750889679716,
363
+ "eval_loss": 1.2006800174713135,
364
+ "eval_macro_f1": 0.8294830724940376,
365
+ "eval_runtime": 0.6345,
366
+ "eval_samples_per_second": 442.838,
367
+ "eval_steps_per_second": 28.367,
368
+ "eval_weighted_f1": 0.8479468676609678,
369
+ "step": 869
370
+ },
371
+ {
372
+ "epoch": 11.075949367088608,
373
+ "grad_norm": 268.8535461425781,
374
+ "learning_rate": 9.929676511954994e-06,
375
+ "loss": 0.2889,
376
+ "step": 875
377
+ },
378
+ {
379
+ "epoch": 11.39240506329114,
380
+ "grad_norm": 0.028925402089953423,
381
+ "learning_rate": 9.578059071729959e-06,
382
+ "loss": 0.2529,
383
+ "step": 900
384
+ },
385
+ {
386
+ "epoch": 11.708860759493671,
387
+ "grad_norm": 0.23740242421627045,
388
+ "learning_rate": 9.226441631504923e-06,
389
+ "loss": 0.2063,
390
+ "step": 925
391
+ },
392
+ {
393
+ "epoch": 12.0,
394
+ "eval_accuracy": 0.8505338078291815,
395
+ "eval_loss": 1.2759236097335815,
396
+ "eval_macro_f1": 0.8362470862470862,
397
+ "eval_runtime": 0.6721,
398
+ "eval_samples_per_second": 418.076,
399
+ "eval_steps_per_second": 26.781,
400
+ "eval_weighted_f1": 0.8525993579374362,
401
+ "step": 948
402
+ }
403
+ ],
404
+ "logging_steps": 25,
405
+ "max_steps": 1580,
406
+ "num_input_tokens_seen": 0,
407
+ "num_train_epochs": 20,
408
+ "save_steps": 500,
409
+ "stateful_callbacks": {
410
+ "EarlyStoppingCallback": {
411
+ "args": {
412
+ "early_stopping_patience": 3,
413
+ "early_stopping_threshold": 0.0
414
+ },
415
+ "attributes": {
416
+ "early_stopping_patience_counter": 3
417
+ }
418
+ },
419
+ "TrainerControl": {
420
+ "args": {
421
+ "should_epoch_stop": false,
422
+ "should_evaluate": false,
423
+ "should_log": false,
424
+ "should_save": true,
425
+ "should_training_stop": true
426
+ },
427
+ "attributes": {}
428
+ }
429
+ },
430
+ "total_flos": 0.0,
431
+ "train_batch_size": 8,
432
+ "trial_name": null,
433
+ "trial_params": null
434
+ }
ood-vifactcheck-xlmr-source-source-00-s42-20260901T150602Z-27bda108/trainer/checkpoint-2548/special_tokens_map.json ADDED
@@ -0,0 +1,15 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "bos_token": "<s>",
3
+ "cls_token": "<s>",
4
+ "eos_token": "</s>",
5
+ "mask_token": {
6
+ "content": "<mask>",
7
+ "lstrip": true,
8
+ "normalized": false,
9
+ "rstrip": false,
10
+ "single_word": false
11
+ },
12
+ "pad_token": "<pad>",
13
+ "sep_token": "</s>",
14
+ "unk_token": "<unk>"
15
+ }
ood-vifactcheck-xlmr-source-source-00-s42-20260901T150602Z-27bda108/trainer/checkpoint-2548/tokenizer_config.json ADDED
@@ -0,0 +1,55 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "added_tokens_decoder": {
3
+ "0": {
4
+ "content": "<s>",
5
+ "lstrip": false,
6
+ "normalized": false,
7
+ "rstrip": false,
8
+ "single_word": false,
9
+ "special": true
10
+ },
11
+ "1": {
12
+ "content": "<pad>",
13
+ "lstrip": false,
14
+ "normalized": false,
15
+ "rstrip": false,
16
+ "single_word": false,
17
+ "special": true
18
+ },
19
+ "2": {
20
+ "content": "</s>",
21
+ "lstrip": false,
22
+ "normalized": false,
23
+ "rstrip": false,
24
+ "single_word": false,
25
+ "special": true
26
+ },
27
+ "3": {
28
+ "content": "<unk>",
29
+ "lstrip": false,
30
+ "normalized": false,
31
+ "rstrip": false,
32
+ "single_word": false,
33
+ "special": true
34
+ },
35
+ "250001": {
36
+ "content": "<mask>",
37
+ "lstrip": true,
38
+ "normalized": false,
39
+ "rstrip": false,
40
+ "single_word": false,
41
+ "special": true
42
+ }
43
+ },
44
+ "bos_token": "<s>",
45
+ "clean_up_tokenization_spaces": false,
46
+ "cls_token": "<s>",
47
+ "eos_token": "</s>",
48
+ "extra_special_tokens": {},
49
+ "mask_token": "<mask>",
50
+ "model_max_length": 512,
51
+ "pad_token": "<pad>",
52
+ "sep_token": "</s>",
53
+ "tokenizer_class": "XLMRobertaTokenizer",
54
+ "unk_token": "<unk>"
55
+ }
ood-vifactcheck-xlmr-source-source-00-s42-20260901T150602Z-27bda108/trainer/checkpoint-2548/trainer_state.json ADDED
@@ -0,0 +1,904 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_global_step": 2548,
3
+ "best_metric": 0.8084286581699676,
4
+ "best_model_checkpoint": "/root/VNEncoder/results/checkpoints/ood-vifactcheck-xlmr-source-source-00-s42-20260901T150602Z-27bda108/trainer/checkpoint-2548",
5
+ "epoch": 14.0,
6
+ "eval_steps": 500,
7
+ "global_step": 2548,
8
+ "is_hyper_param_search": false,
9
+ "is_local_process_zero": true,
10
+ "is_world_process_zero": true,
11
+ "log_history": [
12
+ {
13
+ "epoch": 0.13793103448275862,
14
+ "grad_norm": 39.19527816772461,
15
+ "learning_rate": 1.3186813186813187e-06,
16
+ "loss": 4.9093,
17
+ "step": 25
18
+ },
19
+ {
20
+ "epoch": 0.27586206896551724,
21
+ "grad_norm": 52.46269226074219,
22
+ "learning_rate": 2.6923076923076923e-06,
23
+ "loss": 4.7032,
24
+ "step": 50
25
+ },
26
+ {
27
+ "epoch": 0.41379310344827586,
28
+ "grad_norm": 62.08519744873047,
29
+ "learning_rate": 4.065934065934066e-06,
30
+ "loss": 4.6141,
31
+ "step": 75
32
+ },
33
+ {
34
+ "epoch": 0.5517241379310345,
35
+ "grad_norm": 52.28927993774414,
36
+ "learning_rate": 5.43956043956044e-06,
37
+ "loss": 4.4695,
38
+ "step": 100
39
+ },
40
+ {
41
+ "epoch": 0.6896551724137931,
42
+ "grad_norm": 72.97164916992188,
43
+ "learning_rate": 6.813186813186814e-06,
44
+ "loss": 4.445,
45
+ "step": 125
46
+ },
47
+ {
48
+ "epoch": 0.8275862068965517,
49
+ "grad_norm": 135.36875915527344,
50
+ "learning_rate": 8.186813186813188e-06,
51
+ "loss": 4.3505,
52
+ "step": 150
53
+ },
54
+ {
55
+ "epoch": 0.9655172413793104,
56
+ "grad_norm": 87.52588653564453,
57
+ "learning_rate": 9.560439560439562e-06,
58
+ "loss": 4.3182,
59
+ "step": 175
60
+ },
61
+ {
62
+ "epoch": 1.0,
63
+ "eval_accuracy": 0.5582137161084529,
64
+ "eval_loss": 0.9493994116783142,
65
+ "eval_macro_f1": 0.5214585756924014,
66
+ "eval_runtime": 1.145,
67
+ "eval_samples_per_second": 547.614,
68
+ "eval_steps_per_second": 34.935,
69
+ "eval_weighted_f1": 0.5239640857538724,
70
+ "step": 182
71
+ },
72
+ {
73
+ "epoch": 1.0993103448275863,
74
+ "grad_norm": 18.75910186767578,
75
+ "learning_rate": 1.0934065934065933e-05,
76
+ "loss": 3.8659,
77
+ "step": 200
78
+ },
79
+ {
80
+ "epoch": 1.237241379310345,
81
+ "grad_norm": 103.00005340576172,
82
+ "learning_rate": 1.230769230769231e-05,
83
+ "loss": 3.9323,
84
+ "step": 225
85
+ },
86
+ {
87
+ "epoch": 1.3751724137931034,
88
+ "grad_norm": 461.6681823730469,
89
+ "learning_rate": 1.3681318681318683e-05,
90
+ "loss": 3.9205,
91
+ "step": 250
92
+ },
93
+ {
94
+ "epoch": 1.513103448275862,
95
+ "grad_norm": 86.98558807373047,
96
+ "learning_rate": 1.5054945054945056e-05,
97
+ "loss": 3.9262,
98
+ "step": 275
99
+ },
100
+ {
101
+ "epoch": 1.6510344827586207,
102
+ "grad_norm": 118.53007507324219,
103
+ "learning_rate": 1.642857142857143e-05,
104
+ "loss": 3.6482,
105
+ "step": 300
106
+ },
107
+ {
108
+ "epoch": 1.7889655172413792,
109
+ "grad_norm": 61.446815490722656,
110
+ "learning_rate": 1.78021978021978e-05,
111
+ "loss": 3.3495,
112
+ "step": 325
113
+ },
114
+ {
115
+ "epoch": 1.926896551724138,
116
+ "grad_norm": 36.16039276123047,
117
+ "learning_rate": 1.9175824175824178e-05,
118
+ "loss": 3.1645,
119
+ "step": 350
120
+ },
121
+ {
122
+ "epoch": 2.0,
123
+ "eval_accuracy": 0.6778309409888357,
124
+ "eval_loss": 0.7728503346443176,
125
+ "eval_macro_f1": 0.663754797826036,
126
+ "eval_runtime": 1.2012,
127
+ "eval_samples_per_second": 521.966,
128
+ "eval_steps_per_second": 33.299,
129
+ "eval_weighted_f1": 0.664461982274911,
130
+ "step": 364
131
+ },
132
+ {
133
+ "epoch": 2.0606896551724136,
134
+ "grad_norm": 90.64247131347656,
135
+ "learning_rate": 1.993894993894994e-05,
136
+ "loss": 3.1183,
137
+ "step": 375
138
+ },
139
+ {
140
+ "epoch": 2.1986206896551725,
141
+ "grad_norm": 132.85733032226562,
142
+ "learning_rate": 1.9786324786324788e-05,
143
+ "loss": 3.028,
144
+ "step": 400
145
+ },
146
+ {
147
+ "epoch": 2.336551724137931,
148
+ "grad_norm": 68.87568664550781,
149
+ "learning_rate": 1.9633699633699634e-05,
150
+ "loss": 2.7834,
151
+ "step": 425
152
+ },
153
+ {
154
+ "epoch": 2.47448275862069,
155
+ "grad_norm": 213.333251953125,
156
+ "learning_rate": 1.948107448107448e-05,
157
+ "loss": 2.8182,
158
+ "step": 450
159
+ },
160
+ {
161
+ "epoch": 2.6124137931034483,
162
+ "grad_norm": 68.24656677246094,
163
+ "learning_rate": 1.932844932844933e-05,
164
+ "loss": 2.7826,
165
+ "step": 475
166
+ },
167
+ {
168
+ "epoch": 2.750344827586207,
169
+ "grad_norm": 57.20626449584961,
170
+ "learning_rate": 1.9175824175824178e-05,
171
+ "loss": 2.4841,
172
+ "step": 500
173
+ },
174
+ {
175
+ "epoch": 2.8882758620689657,
176
+ "grad_norm": 86.33191680908203,
177
+ "learning_rate": 1.9023199023199025e-05,
178
+ "loss": 2.3844,
179
+ "step": 525
180
+ },
181
+ {
182
+ "epoch": 3.0,
183
+ "eval_accuracy": 0.759170653907496,
184
+ "eval_loss": 0.634452223777771,
185
+ "eval_macro_f1": 0.7576138458491398,
186
+ "eval_runtime": 1.2596,
187
+ "eval_samples_per_second": 497.763,
188
+ "eval_steps_per_second": 31.755,
189
+ "eval_weighted_f1": 0.7578268121471049,
190
+ "step": 546
191
+ },
192
+ {
193
+ "epoch": 3.0220689655172412,
194
+ "grad_norm": 46.42818832397461,
195
+ "learning_rate": 1.887057387057387e-05,
196
+ "loss": 2.1964,
197
+ "step": 550
198
+ },
199
+ {
200
+ "epoch": 3.16,
201
+ "grad_norm": 85.85521697998047,
202
+ "learning_rate": 1.8717948717948718e-05,
203
+ "loss": 1.9962,
204
+ "step": 575
205
+ },
206
+ {
207
+ "epoch": 3.2979310344827586,
208
+ "grad_norm": 74.1902847290039,
209
+ "learning_rate": 1.8565323565323568e-05,
210
+ "loss": 1.9985,
211
+ "step": 600
212
+ },
213
+ {
214
+ "epoch": 3.435862068965517,
215
+ "grad_norm": 55.97420120239258,
216
+ "learning_rate": 1.8412698412698415e-05,
217
+ "loss": 2.1766,
218
+ "step": 625
219
+ },
220
+ {
221
+ "epoch": 3.573793103448276,
222
+ "grad_norm": 61.05451202392578,
223
+ "learning_rate": 1.826007326007326e-05,
224
+ "loss": 1.9655,
225
+ "step": 650
226
+ },
227
+ {
228
+ "epoch": 3.7117241379310344,
229
+ "grad_norm": 74.15213012695312,
230
+ "learning_rate": 1.8107448107448108e-05,
231
+ "loss": 1.8523,
232
+ "step": 675
233
+ },
234
+ {
235
+ "epoch": 3.849655172413793,
236
+ "grad_norm": 123.62406921386719,
237
+ "learning_rate": 1.7954822954822958e-05,
238
+ "loss": 2.1683,
239
+ "step": 700
240
+ },
241
+ {
242
+ "epoch": 3.9875862068965517,
243
+ "grad_norm": 112.76321411132812,
244
+ "learning_rate": 1.78021978021978e-05,
245
+ "loss": 2.263,
246
+ "step": 725
247
+ },
248
+ {
249
+ "epoch": 4.0,
250
+ "eval_accuracy": 0.7623604465709729,
251
+ "eval_loss": 0.6158862113952637,
252
+ "eval_macro_f1": 0.761909154501918,
253
+ "eval_runtime": 1.1601,
254
+ "eval_samples_per_second": 540.485,
255
+ "eval_steps_per_second": 34.481,
256
+ "eval_weighted_f1": 0.7622521658984954,
257
+ "step": 728
258
+ },
259
+ {
260
+ "epoch": 4.121379310344827,
261
+ "grad_norm": 87.3993148803711,
262
+ "learning_rate": 1.764957264957265e-05,
263
+ "loss": 1.6419,
264
+ "step": 750
265
+ },
266
+ {
267
+ "epoch": 4.259310344827586,
268
+ "grad_norm": 37.41790771484375,
269
+ "learning_rate": 1.74969474969475e-05,
270
+ "loss": 1.6413,
271
+ "step": 775
272
+ },
273
+ {
274
+ "epoch": 4.397241379310345,
275
+ "grad_norm": 64.60355377197266,
276
+ "learning_rate": 1.7344322344322345e-05,
277
+ "loss": 1.4674,
278
+ "step": 800
279
+ },
280
+ {
281
+ "epoch": 4.535172413793103,
282
+ "grad_norm": 30.588787078857422,
283
+ "learning_rate": 1.7191697191697195e-05,
284
+ "loss": 1.7474,
285
+ "step": 825
286
+ },
287
+ {
288
+ "epoch": 4.673103448275862,
289
+ "grad_norm": 92.7600326538086,
290
+ "learning_rate": 1.703907203907204e-05,
291
+ "loss": 1.4887,
292
+ "step": 850
293
+ },
294
+ {
295
+ "epoch": 4.811034482758621,
296
+ "grad_norm": 119.30988311767578,
297
+ "learning_rate": 1.688644688644689e-05,
298
+ "loss": 1.7374,
299
+ "step": 875
300
+ },
301
+ {
302
+ "epoch": 4.94896551724138,
303
+ "grad_norm": 58.519447326660156,
304
+ "learning_rate": 1.6733821733821735e-05,
305
+ "loss": 1.8617,
306
+ "step": 900
307
+ },
308
+ {
309
+ "epoch": 5.0,
310
+ "eval_accuracy": 0.7783094098883573,
311
+ "eval_loss": 0.6327180862426758,
312
+ "eval_macro_f1": 0.779588954162009,
313
+ "eval_runtime": 1.1419,
314
+ "eval_samples_per_second": 549.099,
315
+ "eval_steps_per_second": 35.03,
316
+ "eval_weighted_f1": 0.779605143168874,
317
+ "step": 910
318
+ },
319
+ {
320
+ "epoch": 5.082758620689655,
321
+ "grad_norm": 42.61728286743164,
322
+ "learning_rate": 1.6581196581196585e-05,
323
+ "loss": 1.5041,
324
+ "step": 925
325
+ },
326
+ {
327
+ "epoch": 5.220689655172414,
328
+ "grad_norm": 154.97837829589844,
329
+ "learning_rate": 1.642857142857143e-05,
330
+ "loss": 1.2823,
331
+ "step": 950
332
+ },
333
+ {
334
+ "epoch": 5.358620689655172,
335
+ "grad_norm": 129.80027770996094,
336
+ "learning_rate": 1.627594627594628e-05,
337
+ "loss": 1.2453,
338
+ "step": 975
339
+ },
340
+ {
341
+ "epoch": 5.496551724137931,
342
+ "grad_norm": 54.04684829711914,
343
+ "learning_rate": 1.6123321123321125e-05,
344
+ "loss": 1.4213,
345
+ "step": 1000
346
+ },
347
+ {
348
+ "epoch": 5.63448275862069,
349
+ "grad_norm": 80.3543472290039,
350
+ "learning_rate": 1.5970695970695972e-05,
351
+ "loss": 1.1575,
352
+ "step": 1025
353
+ },
354
+ {
355
+ "epoch": 5.772413793103448,
356
+ "grad_norm": 88.98538208007812,
357
+ "learning_rate": 1.581807081807082e-05,
358
+ "loss": 1.1128,
359
+ "step": 1050
360
+ },
361
+ {
362
+ "epoch": 5.910344827586207,
363
+ "grad_norm": 85.2730941772461,
364
+ "learning_rate": 1.5665445665445665e-05,
365
+ "loss": 1.2682,
366
+ "step": 1075
367
+ },
368
+ {
369
+ "epoch": 6.0,
370
+ "eval_accuracy": 0.7910685805422647,
371
+ "eval_loss": 0.6248478293418884,
372
+ "eval_macro_f1": 0.7915072260253936,
373
+ "eval_runtime": 1.1875,
374
+ "eval_samples_per_second": 527.995,
375
+ "eval_steps_per_second": 33.684,
376
+ "eval_weighted_f1": 0.791812298590797,
377
+ "step": 1092
378
+ },
379
+ {
380
+ "epoch": 6.0441379310344825,
381
+ "grad_norm": 53.63780212402344,
382
+ "learning_rate": 1.5512820512820516e-05,
383
+ "loss": 1.1473,
384
+ "step": 1100
385
+ },
386
+ {
387
+ "epoch": 6.182068965517241,
388
+ "grad_norm": 63.029563903808594,
389
+ "learning_rate": 1.5360195360195362e-05,
390
+ "loss": 0.8161,
391
+ "step": 1125
392
+ },
393
+ {
394
+ "epoch": 6.32,
395
+ "grad_norm": 28.43917465209961,
396
+ "learning_rate": 1.5207570207570207e-05,
397
+ "loss": 0.8535,
398
+ "step": 1150
399
+ },
400
+ {
401
+ "epoch": 6.457931034482758,
402
+ "grad_norm": 158.05908203125,
403
+ "learning_rate": 1.5054945054945056e-05,
404
+ "loss": 1.1149,
405
+ "step": 1175
406
+ },
407
+ {
408
+ "epoch": 6.595862068965517,
409
+ "grad_norm": 70.95556640625,
410
+ "learning_rate": 1.4902319902319904e-05,
411
+ "loss": 0.8912,
412
+ "step": 1200
413
+ },
414
+ {
415
+ "epoch": 6.733793103448276,
416
+ "grad_norm": 74.47285461425781,
417
+ "learning_rate": 1.4749694749694752e-05,
418
+ "loss": 0.8943,
419
+ "step": 1225
420
+ },
421
+ {
422
+ "epoch": 6.871724137931034,
423
+ "grad_norm": 163.60885620117188,
424
+ "learning_rate": 1.4597069597069597e-05,
425
+ "loss": 1.049,
426
+ "step": 1250
427
+ },
428
+ {
429
+ "epoch": 7.0,
430
+ "eval_accuracy": 0.7990430622009569,
431
+ "eval_loss": 0.7322165966033936,
432
+ "eval_macro_f1": 0.7991770461017825,
433
+ "eval_runtime": 1.304,
434
+ "eval_samples_per_second": 480.823,
435
+ "eval_steps_per_second": 30.675,
436
+ "eval_weighted_f1": 0.7994727057439845,
437
+ "step": 1274
438
+ },
439
+ {
440
+ "epoch": 7.0055172413793105,
441
+ "grad_norm": 54.150089263916016,
442
+ "learning_rate": 1.4444444444444446e-05,
443
+ "loss": 1.021,
444
+ "step": 1275
445
+ },
446
+ {
447
+ "epoch": 7.1434482758620685,
448
+ "grad_norm": 34.459228515625,
449
+ "learning_rate": 1.4291819291819293e-05,
450
+ "loss": 0.6191,
451
+ "step": 1300
452
+ },
453
+ {
454
+ "epoch": 7.281379310344827,
455
+ "grad_norm": 67.76248168945312,
456
+ "learning_rate": 1.4139194139194141e-05,
457
+ "loss": 0.7633,
458
+ "step": 1325
459
+ },
460
+ {
461
+ "epoch": 7.419310344827586,
462
+ "grad_norm": 21.25878143310547,
463
+ "learning_rate": 1.3986568986568988e-05,
464
+ "loss": 0.6288,
465
+ "step": 1350
466
+ },
467
+ {
468
+ "epoch": 7.557241379310344,
469
+ "grad_norm": 62.710269927978516,
470
+ "learning_rate": 1.3833943833943834e-05,
471
+ "loss": 0.6954,
472
+ "step": 1375
473
+ },
474
+ {
475
+ "epoch": 7.695172413793103,
476
+ "grad_norm": 194.7527618408203,
477
+ "learning_rate": 1.3681318681318683e-05,
478
+ "loss": 0.4977,
479
+ "step": 1400
480
+ },
481
+ {
482
+ "epoch": 7.833103448275862,
483
+ "grad_norm": 118.31316375732422,
484
+ "learning_rate": 1.3528693528693531e-05,
485
+ "loss": 0.9496,
486
+ "step": 1425
487
+ },
488
+ {
489
+ "epoch": 7.971034482758621,
490
+ "grad_norm": 90.14521026611328,
491
+ "learning_rate": 1.3376068376068376e-05,
492
+ "loss": 0.7248,
493
+ "step": 1450
494
+ },
495
+ {
496
+ "epoch": 8.0,
497
+ "eval_accuracy": 0.8022328548644339,
498
+ "eval_loss": 0.7689692378044128,
499
+ "eval_macro_f1": 0.8004514623124823,
500
+ "eval_runtime": 1.1568,
501
+ "eval_samples_per_second": 542.001,
502
+ "eval_steps_per_second": 34.577,
503
+ "eval_weighted_f1": 0.801057614157989,
504
+ "step": 1456
505
+ },
506
+ {
507
+ "epoch": 8.104827586206897,
508
+ "grad_norm": 301.8183898925781,
509
+ "learning_rate": 1.3223443223443224e-05,
510
+ "loss": 0.5541,
511
+ "step": 1475
512
+ },
513
+ {
514
+ "epoch": 8.242758620689655,
515
+ "grad_norm": 203.32354736328125,
516
+ "learning_rate": 1.3070818070818073e-05,
517
+ "loss": 0.5226,
518
+ "step": 1500
519
+ },
520
+ {
521
+ "epoch": 8.380689655172414,
522
+ "grad_norm": 15.354279518127441,
523
+ "learning_rate": 1.291819291819292e-05,
524
+ "loss": 0.4858,
525
+ "step": 1525
526
+ },
527
+ {
528
+ "epoch": 8.518620689655172,
529
+ "grad_norm": 10.226881980895996,
530
+ "learning_rate": 1.2765567765567766e-05,
531
+ "loss": 0.5951,
532
+ "step": 1550
533
+ },
534
+ {
535
+ "epoch": 8.65655172413793,
536
+ "grad_norm": 83.22173309326172,
537
+ "learning_rate": 1.2612942612942613e-05,
538
+ "loss": 0.5951,
539
+ "step": 1575
540
+ },
541
+ {
542
+ "epoch": 8.79448275862069,
543
+ "grad_norm": 12.615023612976074,
544
+ "learning_rate": 1.2460317460317461e-05,
545
+ "loss": 0.489,
546
+ "step": 1600
547
+ },
548
+ {
549
+ "epoch": 8.932413793103448,
550
+ "grad_norm": 54.15406799316406,
551
+ "learning_rate": 1.230769230769231e-05,
552
+ "loss": 0.4364,
553
+ "step": 1625
554
+ },
555
+ {
556
+ "epoch": 9.0,
557
+ "eval_accuracy": 0.8022328548644339,
558
+ "eval_loss": 0.9064925909042358,
559
+ "eval_macro_f1": 0.8025287782234445,
560
+ "eval_runtime": 1.1773,
561
+ "eval_samples_per_second": 532.59,
562
+ "eval_steps_per_second": 33.977,
563
+ "eval_weighted_f1": 0.8023587030058814,
564
+ "step": 1638
565
+ },
566
+ {
567
+ "epoch": 9.066206896551725,
568
+ "grad_norm": 87.1186294555664,
569
+ "learning_rate": 1.2155067155067155e-05,
570
+ "loss": 0.4131,
571
+ "step": 1650
572
+ },
573
+ {
574
+ "epoch": 9.204137931034483,
575
+ "grad_norm": 118.65606689453125,
576
+ "learning_rate": 1.2002442002442003e-05,
577
+ "loss": 0.3966,
578
+ "step": 1675
579
+ },
580
+ {
581
+ "epoch": 9.34206896551724,
582
+ "grad_norm": 43.27682113647461,
583
+ "learning_rate": 1.1849816849816852e-05,
584
+ "loss": 0.3994,
585
+ "step": 1700
586
+ },
587
+ {
588
+ "epoch": 9.48,
589
+ "grad_norm": 46.55112075805664,
590
+ "learning_rate": 1.1697191697191698e-05,
591
+ "loss": 0.3002,
592
+ "step": 1725
593
+ },
594
+ {
595
+ "epoch": 9.617931034482758,
596
+ "grad_norm": 299.67840576171875,
597
+ "learning_rate": 1.1544566544566545e-05,
598
+ "loss": 0.4456,
599
+ "step": 1750
600
+ },
601
+ {
602
+ "epoch": 9.755862068965516,
603
+ "grad_norm": 94.89724731445312,
604
+ "learning_rate": 1.1391941391941392e-05,
605
+ "loss": 0.4069,
606
+ "step": 1775
607
+ },
608
+ {
609
+ "epoch": 9.893793103448276,
610
+ "grad_norm": 83.06227111816406,
611
+ "learning_rate": 1.123931623931624e-05,
612
+ "loss": 0.3748,
613
+ "step": 1800
614
+ },
615
+ {
616
+ "epoch": 10.0,
617
+ "eval_accuracy": 0.773524720893142,
618
+ "eval_loss": 1.3610498905181885,
619
+ "eval_macro_f1": 0.7716054942384281,
620
+ "eval_runtime": 1.1853,
621
+ "eval_samples_per_second": 528.97,
622
+ "eval_steps_per_second": 33.746,
623
+ "eval_weighted_f1": 0.7719216877725383,
624
+ "step": 1820
625
+ },
626
+ {
627
+ "epoch": 10.027586206896551,
628
+ "grad_norm": 171.37094116210938,
629
+ "learning_rate": 1.1086691086691088e-05,
630
+ "loss": 0.398,
631
+ "step": 1825
632
+ },
633
+ {
634
+ "epoch": 10.16551724137931,
635
+ "grad_norm": 31.03558349609375,
636
+ "learning_rate": 1.0934065934065933e-05,
637
+ "loss": 0.2433,
638
+ "step": 1850
639
+ },
640
+ {
641
+ "epoch": 10.303448275862069,
642
+ "grad_norm": 5.404947757720947,
643
+ "learning_rate": 1.0781440781440782e-05,
644
+ "loss": 0.3894,
645
+ "step": 1875
646
+ },
647
+ {
648
+ "epoch": 10.441379310344828,
649
+ "grad_norm": 1.9526923894882202,
650
+ "learning_rate": 1.062881562881563e-05,
651
+ "loss": 0.2768,
652
+ "step": 1900
653
+ },
654
+ {
655
+ "epoch": 10.579310344827586,
656
+ "grad_norm": 3.5537567138671875,
657
+ "learning_rate": 1.0476190476190477e-05,
658
+ "loss": 0.3726,
659
+ "step": 1925
660
+ },
661
+ {
662
+ "epoch": 10.717241379310344,
663
+ "grad_norm": 218.355712890625,
664
+ "learning_rate": 1.0323565323565325e-05,
665
+ "loss": 0.2008,
666
+ "step": 1950
667
+ },
668
+ {
669
+ "epoch": 10.855172413793104,
670
+ "grad_norm": 0.5096381306648254,
671
+ "learning_rate": 1.0170940170940172e-05,
672
+ "loss": 0.5065,
673
+ "step": 1975
674
+ },
675
+ {
676
+ "epoch": 10.993103448275862,
677
+ "grad_norm": 0.12985798716545105,
678
+ "learning_rate": 1.0018315018315019e-05,
679
+ "loss": 0.2639,
680
+ "step": 2000
681
+ },
682
+ {
683
+ "epoch": 11.0,
684
+ "eval_accuracy": 0.79585326953748,
685
+ "eval_loss": 1.401996374130249,
686
+ "eval_macro_f1": 0.794472041204085,
687
+ "eval_runtime": 1.1683,
688
+ "eval_samples_per_second": 536.672,
689
+ "eval_steps_per_second": 34.237,
690
+ "eval_weighted_f1": 0.7946472239407022,
691
+ "step": 2002
692
+ },
693
+ {
694
+ "epoch": 11.126896551724139,
695
+ "grad_norm": 244.31393432617188,
696
+ "learning_rate": 9.865689865689867e-06,
697
+ "loss": 0.4166,
698
+ "step": 2025
699
+ },
700
+ {
701
+ "epoch": 11.264827586206897,
702
+ "grad_norm": 0.7453848123550415,
703
+ "learning_rate": 9.713064713064714e-06,
704
+ "loss": 0.1639,
705
+ "step": 2050
706
+ },
707
+ {
708
+ "epoch": 11.402758620689655,
709
+ "grad_norm": 1.1450600624084473,
710
+ "learning_rate": 9.560439560439562e-06,
711
+ "loss": 0.2126,
712
+ "step": 2075
713
+ },
714
+ {
715
+ "epoch": 11.540689655172415,
716
+ "grad_norm": 94.49410247802734,
717
+ "learning_rate": 9.407814407814409e-06,
718
+ "loss": 0.4276,
719
+ "step": 2100
720
+ },
721
+ {
722
+ "epoch": 11.678620689655173,
723
+ "grad_norm": 53.80741500854492,
724
+ "learning_rate": 9.255189255189256e-06,
725
+ "loss": 0.4125,
726
+ "step": 2125
727
+ },
728
+ {
729
+ "epoch": 11.81655172413793,
730
+ "grad_norm": 2.2708137035369873,
731
+ "learning_rate": 9.102564102564104e-06,
732
+ "loss": 0.3553,
733
+ "step": 2150
734
+ },
735
+ {
736
+ "epoch": 11.95448275862069,
737
+ "grad_norm": 16.314722061157227,
738
+ "learning_rate": 8.94993894993895e-06,
739
+ "loss": 0.213,
740
+ "step": 2175
741
+ },
742
+ {
743
+ "epoch": 12.0,
744
+ "eval_accuracy": 0.8086124401913876,
745
+ "eval_loss": 1.4752185344696045,
746
+ "eval_macro_f1": 0.8073371962771456,
747
+ "eval_runtime": 1.1292,
748
+ "eval_samples_per_second": 555.264,
749
+ "eval_steps_per_second": 35.424,
750
+ "eval_weighted_f1": 0.8074475414562201,
751
+ "step": 2184
752
+ },
753
+ {
754
+ "epoch": 12.088275862068965,
755
+ "grad_norm": 4.032958507537842,
756
+ "learning_rate": 8.797313797313797e-06,
757
+ "loss": 0.1548,
758
+ "step": 2200
759
+ },
760
+ {
761
+ "epoch": 12.226206896551725,
762
+ "grad_norm": 11.657094955444336,
763
+ "learning_rate": 8.644688644688646e-06,
764
+ "loss": 0.2706,
765
+ "step": 2225
766
+ },
767
+ {
768
+ "epoch": 12.364137931034483,
769
+ "grad_norm": 12.007180213928223,
770
+ "learning_rate": 8.492063492063492e-06,
771
+ "loss": 0.1725,
772
+ "step": 2250
773
+ },
774
+ {
775
+ "epoch": 12.50206896551724,
776
+ "grad_norm": 3.103738307952881,
777
+ "learning_rate": 8.33943833943834e-06,
778
+ "loss": 0.1077,
779
+ "step": 2275
780
+ },
781
+ {
782
+ "epoch": 12.64,
783
+ "grad_norm": 0.05534272640943527,
784
+ "learning_rate": 8.186813186813188e-06,
785
+ "loss": 0.2332,
786
+ "step": 2300
787
+ },
788
+ {
789
+ "epoch": 12.777931034482759,
790
+ "grad_norm": 0.27866512537002563,
791
+ "learning_rate": 8.034188034188036e-06,
792
+ "loss": 0.0977,
793
+ "step": 2325
794
+ },
795
+ {
796
+ "epoch": 12.915862068965517,
797
+ "grad_norm": 16.027671813964844,
798
+ "learning_rate": 7.881562881562883e-06,
799
+ "loss": 0.3268,
800
+ "step": 2350
801
+ },
802
+ {
803
+ "epoch": 13.0,
804
+ "eval_accuracy": 0.7974481658692185,
805
+ "eval_loss": 1.5549489259719849,
806
+ "eval_macro_f1": 0.7969154183788802,
807
+ "eval_runtime": 1.1851,
808
+ "eval_samples_per_second": 529.085,
809
+ "eval_steps_per_second": 33.753,
810
+ "eval_weighted_f1": 0.7970637749059234,
811
+ "step": 2366
812
+ },
813
+ {
814
+ "epoch": 13.049655172413793,
815
+ "grad_norm": 167.1862030029297,
816
+ "learning_rate": 7.72893772893773e-06,
817
+ "loss": 0.1916,
818
+ "step": 2375
819
+ },
820
+ {
821
+ "epoch": 13.187586206896551,
822
+ "grad_norm": 578.3486328125,
823
+ "learning_rate": 7.576312576312577e-06,
824
+ "loss": 0.2089,
825
+ "step": 2400
826
+ },
827
+ {
828
+ "epoch": 13.32551724137931,
829
+ "grad_norm": 2.9909238815307617,
830
+ "learning_rate": 7.423687423687424e-06,
831
+ "loss": 0.1702,
832
+ "step": 2425
833
+ },
834
+ {
835
+ "epoch": 13.463448275862069,
836
+ "grad_norm": 12.770936012268066,
837
+ "learning_rate": 7.271062271062271e-06,
838
+ "loss": 0.1574,
839
+ "step": 2450
840
+ },
841
+ {
842
+ "epoch": 13.601379310344827,
843
+ "grad_norm": 0.8478724360466003,
844
+ "learning_rate": 7.1184371184371195e-06,
845
+ "loss": 0.1238,
846
+ "step": 2475
847
+ },
848
+ {
849
+ "epoch": 13.739310344827587,
850
+ "grad_norm": 0.18776558339595795,
851
+ "learning_rate": 6.965811965811966e-06,
852
+ "loss": 0.0629,
853
+ "step": 2500
854
+ },
855
+ {
856
+ "epoch": 13.877241379310345,
857
+ "grad_norm": 10.055283546447754,
858
+ "learning_rate": 6.813186813186814e-06,
859
+ "loss": 0.3233,
860
+ "step": 2525
861
+ },
862
+ {
863
+ "epoch": 14.0,
864
+ "eval_accuracy": 0.8086124401913876,
865
+ "eval_loss": 1.6799039840698242,
866
+ "eval_macro_f1": 0.8084286581699676,
867
+ "eval_runtime": 1.1687,
868
+ "eval_samples_per_second": 536.477,
869
+ "eval_steps_per_second": 34.225,
870
+ "eval_weighted_f1": 0.8085553790065865,
871
+ "step": 2548
872
+ }
873
+ ],
874
+ "logging_steps": 25,
875
+ "max_steps": 3640,
876
+ "num_input_tokens_seen": 0,
877
+ "num_train_epochs": 20,
878
+ "save_steps": 500,
879
+ "stateful_callbacks": {
880
+ "EarlyStoppingCallback": {
881
+ "args": {
882
+ "early_stopping_patience": 3,
883
+ "early_stopping_threshold": 0.0
884
+ },
885
+ "attributes": {
886
+ "early_stopping_patience_counter": 0
887
+ }
888
+ },
889
+ "TrainerControl": {
890
+ "args": {
891
+ "should_epoch_stop": false,
892
+ "should_evaluate": false,
893
+ "should_log": false,
894
+ "should_save": true,
895
+ "should_training_stop": false
896
+ },
897
+ "attributes": {}
898
+ }
899
+ },
900
+ "total_flos": 0.0,
901
+ "train_batch_size": 8,
902
+ "trial_name": null,
903
+ "trial_params": null
904
+ }
ood-vifactcheck-xlmr-source-source-00-s42-20260901T150602Z-27bda108/trainer/checkpoint-3094/special_tokens_map.json ADDED
@@ -0,0 +1,15 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "bos_token": "<s>",
3
+ "cls_token": "<s>",
4
+ "eos_token": "</s>",
5
+ "mask_token": {
6
+ "content": "<mask>",
7
+ "lstrip": true,
8
+ "normalized": false,
9
+ "rstrip": false,
10
+ "single_word": false
11
+ },
12
+ "pad_token": "<pad>",
13
+ "sep_token": "</s>",
14
+ "unk_token": "<unk>"
15
+ }
ood-vifactcheck-xlmr-source-source-00-s42-20260901T150602Z-27bda108/trainer/checkpoint-3094/tokenizer_config.json ADDED
@@ -0,0 +1,55 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "added_tokens_decoder": {
3
+ "0": {
4
+ "content": "<s>",
5
+ "lstrip": false,
6
+ "normalized": false,
7
+ "rstrip": false,
8
+ "single_word": false,
9
+ "special": true
10
+ },
11
+ "1": {
12
+ "content": "<pad>",
13
+ "lstrip": false,
14
+ "normalized": false,
15
+ "rstrip": false,
16
+ "single_word": false,
17
+ "special": true
18
+ },
19
+ "2": {
20
+ "content": "</s>",
21
+ "lstrip": false,
22
+ "normalized": false,
23
+ "rstrip": false,
24
+ "single_word": false,
25
+ "special": true
26
+ },
27
+ "3": {
28
+ "content": "<unk>",
29
+ "lstrip": false,
30
+ "normalized": false,
31
+ "rstrip": false,
32
+ "single_word": false,
33
+ "special": true
34
+ },
35
+ "250001": {
36
+ "content": "<mask>",
37
+ "lstrip": true,
38
+ "normalized": false,
39
+ "rstrip": false,
40
+ "single_word": false,
41
+ "special": true
42
+ }
43
+ },
44
+ "bos_token": "<s>",
45
+ "clean_up_tokenization_spaces": false,
46
+ "cls_token": "<s>",
47
+ "eos_token": "</s>",
48
+ "extra_special_tokens": {},
49
+ "mask_token": "<mask>",
50
+ "model_max_length": 512,
51
+ "pad_token": "<pad>",
52
+ "sep_token": "</s>",
53
+ "tokenizer_class": "XLMRobertaTokenizer",
54
+ "unk_token": "<unk>"
55
+ }
ood-vifactcheck-xlmr-source-source-00-s42-20260901T150602Z-27bda108/trainer/checkpoint-3094/trainer_state.json ADDED
@@ -0,0 +1,1091 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_global_step": 2548,
3
+ "best_metric": 0.8084286581699676,
4
+ "best_model_checkpoint": "/root/VNEncoder/results/checkpoints/ood-vifactcheck-xlmr-source-source-00-s42-20260901T150602Z-27bda108/trainer/checkpoint-2548",
5
+ "epoch": 17.0,
6
+ "eval_steps": 500,
7
+ "global_step": 3094,
8
+ "is_hyper_param_search": false,
9
+ "is_local_process_zero": true,
10
+ "is_world_process_zero": true,
11
+ "log_history": [
12
+ {
13
+ "epoch": 0.13793103448275862,
14
+ "grad_norm": 39.19527816772461,
15
+ "learning_rate": 1.3186813186813187e-06,
16
+ "loss": 4.9093,
17
+ "step": 25
18
+ },
19
+ {
20
+ "epoch": 0.27586206896551724,
21
+ "grad_norm": 52.46269226074219,
22
+ "learning_rate": 2.6923076923076923e-06,
23
+ "loss": 4.7032,
24
+ "step": 50
25
+ },
26
+ {
27
+ "epoch": 0.41379310344827586,
28
+ "grad_norm": 62.08519744873047,
29
+ "learning_rate": 4.065934065934066e-06,
30
+ "loss": 4.6141,
31
+ "step": 75
32
+ },
33
+ {
34
+ "epoch": 0.5517241379310345,
35
+ "grad_norm": 52.28927993774414,
36
+ "learning_rate": 5.43956043956044e-06,
37
+ "loss": 4.4695,
38
+ "step": 100
39
+ },
40
+ {
41
+ "epoch": 0.6896551724137931,
42
+ "grad_norm": 72.97164916992188,
43
+ "learning_rate": 6.813186813186814e-06,
44
+ "loss": 4.445,
45
+ "step": 125
46
+ },
47
+ {
48
+ "epoch": 0.8275862068965517,
49
+ "grad_norm": 135.36875915527344,
50
+ "learning_rate": 8.186813186813188e-06,
51
+ "loss": 4.3505,
52
+ "step": 150
53
+ },
54
+ {
55
+ "epoch": 0.9655172413793104,
56
+ "grad_norm": 87.52588653564453,
57
+ "learning_rate": 9.560439560439562e-06,
58
+ "loss": 4.3182,
59
+ "step": 175
60
+ },
61
+ {
62
+ "epoch": 1.0,
63
+ "eval_accuracy": 0.5582137161084529,
64
+ "eval_loss": 0.9493994116783142,
65
+ "eval_macro_f1": 0.5214585756924014,
66
+ "eval_runtime": 1.145,
67
+ "eval_samples_per_second": 547.614,
68
+ "eval_steps_per_second": 34.935,
69
+ "eval_weighted_f1": 0.5239640857538724,
70
+ "step": 182
71
+ },
72
+ {
73
+ "epoch": 1.0993103448275863,
74
+ "grad_norm": 18.75910186767578,
75
+ "learning_rate": 1.0934065934065933e-05,
76
+ "loss": 3.8659,
77
+ "step": 200
78
+ },
79
+ {
80
+ "epoch": 1.237241379310345,
81
+ "grad_norm": 103.00005340576172,
82
+ "learning_rate": 1.230769230769231e-05,
83
+ "loss": 3.9323,
84
+ "step": 225
85
+ },
86
+ {
87
+ "epoch": 1.3751724137931034,
88
+ "grad_norm": 461.6681823730469,
89
+ "learning_rate": 1.3681318681318683e-05,
90
+ "loss": 3.9205,
91
+ "step": 250
92
+ },
93
+ {
94
+ "epoch": 1.513103448275862,
95
+ "grad_norm": 86.98558807373047,
96
+ "learning_rate": 1.5054945054945056e-05,
97
+ "loss": 3.9262,
98
+ "step": 275
99
+ },
100
+ {
101
+ "epoch": 1.6510344827586207,
102
+ "grad_norm": 118.53007507324219,
103
+ "learning_rate": 1.642857142857143e-05,
104
+ "loss": 3.6482,
105
+ "step": 300
106
+ },
107
+ {
108
+ "epoch": 1.7889655172413792,
109
+ "grad_norm": 61.446815490722656,
110
+ "learning_rate": 1.78021978021978e-05,
111
+ "loss": 3.3495,
112
+ "step": 325
113
+ },
114
+ {
115
+ "epoch": 1.926896551724138,
116
+ "grad_norm": 36.16039276123047,
117
+ "learning_rate": 1.9175824175824178e-05,
118
+ "loss": 3.1645,
119
+ "step": 350
120
+ },
121
+ {
122
+ "epoch": 2.0,
123
+ "eval_accuracy": 0.6778309409888357,
124
+ "eval_loss": 0.7728503346443176,
125
+ "eval_macro_f1": 0.663754797826036,
126
+ "eval_runtime": 1.2012,
127
+ "eval_samples_per_second": 521.966,
128
+ "eval_steps_per_second": 33.299,
129
+ "eval_weighted_f1": 0.664461982274911,
130
+ "step": 364
131
+ },
132
+ {
133
+ "epoch": 2.0606896551724136,
134
+ "grad_norm": 90.64247131347656,
135
+ "learning_rate": 1.993894993894994e-05,
136
+ "loss": 3.1183,
137
+ "step": 375
138
+ },
139
+ {
140
+ "epoch": 2.1986206896551725,
141
+ "grad_norm": 132.85733032226562,
142
+ "learning_rate": 1.9786324786324788e-05,
143
+ "loss": 3.028,
144
+ "step": 400
145
+ },
146
+ {
147
+ "epoch": 2.336551724137931,
148
+ "grad_norm": 68.87568664550781,
149
+ "learning_rate": 1.9633699633699634e-05,
150
+ "loss": 2.7834,
151
+ "step": 425
152
+ },
153
+ {
154
+ "epoch": 2.47448275862069,
155
+ "grad_norm": 213.333251953125,
156
+ "learning_rate": 1.948107448107448e-05,
157
+ "loss": 2.8182,
158
+ "step": 450
159
+ },
160
+ {
161
+ "epoch": 2.6124137931034483,
162
+ "grad_norm": 68.24656677246094,
163
+ "learning_rate": 1.932844932844933e-05,
164
+ "loss": 2.7826,
165
+ "step": 475
166
+ },
167
+ {
168
+ "epoch": 2.750344827586207,
169
+ "grad_norm": 57.20626449584961,
170
+ "learning_rate": 1.9175824175824178e-05,
171
+ "loss": 2.4841,
172
+ "step": 500
173
+ },
174
+ {
175
+ "epoch": 2.8882758620689657,
176
+ "grad_norm": 86.33191680908203,
177
+ "learning_rate": 1.9023199023199025e-05,
178
+ "loss": 2.3844,
179
+ "step": 525
180
+ },
181
+ {
182
+ "epoch": 3.0,
183
+ "eval_accuracy": 0.759170653907496,
184
+ "eval_loss": 0.634452223777771,
185
+ "eval_macro_f1": 0.7576138458491398,
186
+ "eval_runtime": 1.2596,
187
+ "eval_samples_per_second": 497.763,
188
+ "eval_steps_per_second": 31.755,
189
+ "eval_weighted_f1": 0.7578268121471049,
190
+ "step": 546
191
+ },
192
+ {
193
+ "epoch": 3.0220689655172412,
194
+ "grad_norm": 46.42818832397461,
195
+ "learning_rate": 1.887057387057387e-05,
196
+ "loss": 2.1964,
197
+ "step": 550
198
+ },
199
+ {
200
+ "epoch": 3.16,
201
+ "grad_norm": 85.85521697998047,
202
+ "learning_rate": 1.8717948717948718e-05,
203
+ "loss": 1.9962,
204
+ "step": 575
205
+ },
206
+ {
207
+ "epoch": 3.2979310344827586,
208
+ "grad_norm": 74.1902847290039,
209
+ "learning_rate": 1.8565323565323568e-05,
210
+ "loss": 1.9985,
211
+ "step": 600
212
+ },
213
+ {
214
+ "epoch": 3.435862068965517,
215
+ "grad_norm": 55.97420120239258,
216
+ "learning_rate": 1.8412698412698415e-05,
217
+ "loss": 2.1766,
218
+ "step": 625
219
+ },
220
+ {
221
+ "epoch": 3.573793103448276,
222
+ "grad_norm": 61.05451202392578,
223
+ "learning_rate": 1.826007326007326e-05,
224
+ "loss": 1.9655,
225
+ "step": 650
226
+ },
227
+ {
228
+ "epoch": 3.7117241379310344,
229
+ "grad_norm": 74.15213012695312,
230
+ "learning_rate": 1.8107448107448108e-05,
231
+ "loss": 1.8523,
232
+ "step": 675
233
+ },
234
+ {
235
+ "epoch": 3.849655172413793,
236
+ "grad_norm": 123.62406921386719,
237
+ "learning_rate": 1.7954822954822958e-05,
238
+ "loss": 2.1683,
239
+ "step": 700
240
+ },
241
+ {
242
+ "epoch": 3.9875862068965517,
243
+ "grad_norm": 112.76321411132812,
244
+ "learning_rate": 1.78021978021978e-05,
245
+ "loss": 2.263,
246
+ "step": 725
247
+ },
248
+ {
249
+ "epoch": 4.0,
250
+ "eval_accuracy": 0.7623604465709729,
251
+ "eval_loss": 0.6158862113952637,
252
+ "eval_macro_f1": 0.761909154501918,
253
+ "eval_runtime": 1.1601,
254
+ "eval_samples_per_second": 540.485,
255
+ "eval_steps_per_second": 34.481,
256
+ "eval_weighted_f1": 0.7622521658984954,
257
+ "step": 728
258
+ },
259
+ {
260
+ "epoch": 4.121379310344827,
261
+ "grad_norm": 87.3993148803711,
262
+ "learning_rate": 1.764957264957265e-05,
263
+ "loss": 1.6419,
264
+ "step": 750
265
+ },
266
+ {
267
+ "epoch": 4.259310344827586,
268
+ "grad_norm": 37.41790771484375,
269
+ "learning_rate": 1.74969474969475e-05,
270
+ "loss": 1.6413,
271
+ "step": 775
272
+ },
273
+ {
274
+ "epoch": 4.397241379310345,
275
+ "grad_norm": 64.60355377197266,
276
+ "learning_rate": 1.7344322344322345e-05,
277
+ "loss": 1.4674,
278
+ "step": 800
279
+ },
280
+ {
281
+ "epoch": 4.535172413793103,
282
+ "grad_norm": 30.588787078857422,
283
+ "learning_rate": 1.7191697191697195e-05,
284
+ "loss": 1.7474,
285
+ "step": 825
286
+ },
287
+ {
288
+ "epoch": 4.673103448275862,
289
+ "grad_norm": 92.7600326538086,
290
+ "learning_rate": 1.703907203907204e-05,
291
+ "loss": 1.4887,
292
+ "step": 850
293
+ },
294
+ {
295
+ "epoch": 4.811034482758621,
296
+ "grad_norm": 119.30988311767578,
297
+ "learning_rate": 1.688644688644689e-05,
298
+ "loss": 1.7374,
299
+ "step": 875
300
+ },
301
+ {
302
+ "epoch": 4.94896551724138,
303
+ "grad_norm": 58.519447326660156,
304
+ "learning_rate": 1.6733821733821735e-05,
305
+ "loss": 1.8617,
306
+ "step": 900
307
+ },
308
+ {
309
+ "epoch": 5.0,
310
+ "eval_accuracy": 0.7783094098883573,
311
+ "eval_loss": 0.6327180862426758,
312
+ "eval_macro_f1": 0.779588954162009,
313
+ "eval_runtime": 1.1419,
314
+ "eval_samples_per_second": 549.099,
315
+ "eval_steps_per_second": 35.03,
316
+ "eval_weighted_f1": 0.779605143168874,
317
+ "step": 910
318
+ },
319
+ {
320
+ "epoch": 5.082758620689655,
321
+ "grad_norm": 42.61728286743164,
322
+ "learning_rate": 1.6581196581196585e-05,
323
+ "loss": 1.5041,
324
+ "step": 925
325
+ },
326
+ {
327
+ "epoch": 5.220689655172414,
328
+ "grad_norm": 154.97837829589844,
329
+ "learning_rate": 1.642857142857143e-05,
330
+ "loss": 1.2823,
331
+ "step": 950
332
+ },
333
+ {
334
+ "epoch": 5.358620689655172,
335
+ "grad_norm": 129.80027770996094,
336
+ "learning_rate": 1.627594627594628e-05,
337
+ "loss": 1.2453,
338
+ "step": 975
339
+ },
340
+ {
341
+ "epoch": 5.496551724137931,
342
+ "grad_norm": 54.04684829711914,
343
+ "learning_rate": 1.6123321123321125e-05,
344
+ "loss": 1.4213,
345
+ "step": 1000
346
+ },
347
+ {
348
+ "epoch": 5.63448275862069,
349
+ "grad_norm": 80.3543472290039,
350
+ "learning_rate": 1.5970695970695972e-05,
351
+ "loss": 1.1575,
352
+ "step": 1025
353
+ },
354
+ {
355
+ "epoch": 5.772413793103448,
356
+ "grad_norm": 88.98538208007812,
357
+ "learning_rate": 1.581807081807082e-05,
358
+ "loss": 1.1128,
359
+ "step": 1050
360
+ },
361
+ {
362
+ "epoch": 5.910344827586207,
363
+ "grad_norm": 85.2730941772461,
364
+ "learning_rate": 1.5665445665445665e-05,
365
+ "loss": 1.2682,
366
+ "step": 1075
367
+ },
368
+ {
369
+ "epoch": 6.0,
370
+ "eval_accuracy": 0.7910685805422647,
371
+ "eval_loss": 0.6248478293418884,
372
+ "eval_macro_f1": 0.7915072260253936,
373
+ "eval_runtime": 1.1875,
374
+ "eval_samples_per_second": 527.995,
375
+ "eval_steps_per_second": 33.684,
376
+ "eval_weighted_f1": 0.791812298590797,
377
+ "step": 1092
378
+ },
379
+ {
380
+ "epoch": 6.0441379310344825,
381
+ "grad_norm": 53.63780212402344,
382
+ "learning_rate": 1.5512820512820516e-05,
383
+ "loss": 1.1473,
384
+ "step": 1100
385
+ },
386
+ {
387
+ "epoch": 6.182068965517241,
388
+ "grad_norm": 63.029563903808594,
389
+ "learning_rate": 1.5360195360195362e-05,
390
+ "loss": 0.8161,
391
+ "step": 1125
392
+ },
393
+ {
394
+ "epoch": 6.32,
395
+ "grad_norm": 28.43917465209961,
396
+ "learning_rate": 1.5207570207570207e-05,
397
+ "loss": 0.8535,
398
+ "step": 1150
399
+ },
400
+ {
401
+ "epoch": 6.457931034482758,
402
+ "grad_norm": 158.05908203125,
403
+ "learning_rate": 1.5054945054945056e-05,
404
+ "loss": 1.1149,
405
+ "step": 1175
406
+ },
407
+ {
408
+ "epoch": 6.595862068965517,
409
+ "grad_norm": 70.95556640625,
410
+ "learning_rate": 1.4902319902319904e-05,
411
+ "loss": 0.8912,
412
+ "step": 1200
413
+ },
414
+ {
415
+ "epoch": 6.733793103448276,
416
+ "grad_norm": 74.47285461425781,
417
+ "learning_rate": 1.4749694749694752e-05,
418
+ "loss": 0.8943,
419
+ "step": 1225
420
+ },
421
+ {
422
+ "epoch": 6.871724137931034,
423
+ "grad_norm": 163.60885620117188,
424
+ "learning_rate": 1.4597069597069597e-05,
425
+ "loss": 1.049,
426
+ "step": 1250
427
+ },
428
+ {
429
+ "epoch": 7.0,
430
+ "eval_accuracy": 0.7990430622009569,
431
+ "eval_loss": 0.7322165966033936,
432
+ "eval_macro_f1": 0.7991770461017825,
433
+ "eval_runtime": 1.304,
434
+ "eval_samples_per_second": 480.823,
435
+ "eval_steps_per_second": 30.675,
436
+ "eval_weighted_f1": 0.7994727057439845,
437
+ "step": 1274
438
+ },
439
+ {
440
+ "epoch": 7.0055172413793105,
441
+ "grad_norm": 54.150089263916016,
442
+ "learning_rate": 1.4444444444444446e-05,
443
+ "loss": 1.021,
444
+ "step": 1275
445
+ },
446
+ {
447
+ "epoch": 7.1434482758620685,
448
+ "grad_norm": 34.459228515625,
449
+ "learning_rate": 1.4291819291819293e-05,
450
+ "loss": 0.6191,
451
+ "step": 1300
452
+ },
453
+ {
454
+ "epoch": 7.281379310344827,
455
+ "grad_norm": 67.76248168945312,
456
+ "learning_rate": 1.4139194139194141e-05,
457
+ "loss": 0.7633,
458
+ "step": 1325
459
+ },
460
+ {
461
+ "epoch": 7.419310344827586,
462
+ "grad_norm": 21.25878143310547,
463
+ "learning_rate": 1.3986568986568988e-05,
464
+ "loss": 0.6288,
465
+ "step": 1350
466
+ },
467
+ {
468
+ "epoch": 7.557241379310344,
469
+ "grad_norm": 62.710269927978516,
470
+ "learning_rate": 1.3833943833943834e-05,
471
+ "loss": 0.6954,
472
+ "step": 1375
473
+ },
474
+ {
475
+ "epoch": 7.695172413793103,
476
+ "grad_norm": 194.7527618408203,
477
+ "learning_rate": 1.3681318681318683e-05,
478
+ "loss": 0.4977,
479
+ "step": 1400
480
+ },
481
+ {
482
+ "epoch": 7.833103448275862,
483
+ "grad_norm": 118.31316375732422,
484
+ "learning_rate": 1.3528693528693531e-05,
485
+ "loss": 0.9496,
486
+ "step": 1425
487
+ },
488
+ {
489
+ "epoch": 7.971034482758621,
490
+ "grad_norm": 90.14521026611328,
491
+ "learning_rate": 1.3376068376068376e-05,
492
+ "loss": 0.7248,
493
+ "step": 1450
494
+ },
495
+ {
496
+ "epoch": 8.0,
497
+ "eval_accuracy": 0.8022328548644339,
498
+ "eval_loss": 0.7689692378044128,
499
+ "eval_macro_f1": 0.8004514623124823,
500
+ "eval_runtime": 1.1568,
501
+ "eval_samples_per_second": 542.001,
502
+ "eval_steps_per_second": 34.577,
503
+ "eval_weighted_f1": 0.801057614157989,
504
+ "step": 1456
505
+ },
506
+ {
507
+ "epoch": 8.104827586206897,
508
+ "grad_norm": 301.8183898925781,
509
+ "learning_rate": 1.3223443223443224e-05,
510
+ "loss": 0.5541,
511
+ "step": 1475
512
+ },
513
+ {
514
+ "epoch": 8.242758620689655,
515
+ "grad_norm": 203.32354736328125,
516
+ "learning_rate": 1.3070818070818073e-05,
517
+ "loss": 0.5226,
518
+ "step": 1500
519
+ },
520
+ {
521
+ "epoch": 8.380689655172414,
522
+ "grad_norm": 15.354279518127441,
523
+ "learning_rate": 1.291819291819292e-05,
524
+ "loss": 0.4858,
525
+ "step": 1525
526
+ },
527
+ {
528
+ "epoch": 8.518620689655172,
529
+ "grad_norm": 10.226881980895996,
530
+ "learning_rate": 1.2765567765567766e-05,
531
+ "loss": 0.5951,
532
+ "step": 1550
533
+ },
534
+ {
535
+ "epoch": 8.65655172413793,
536
+ "grad_norm": 83.22173309326172,
537
+ "learning_rate": 1.2612942612942613e-05,
538
+ "loss": 0.5951,
539
+ "step": 1575
540
+ },
541
+ {
542
+ "epoch": 8.79448275862069,
543
+ "grad_norm": 12.615023612976074,
544
+ "learning_rate": 1.2460317460317461e-05,
545
+ "loss": 0.489,
546
+ "step": 1600
547
+ },
548
+ {
549
+ "epoch": 8.932413793103448,
550
+ "grad_norm": 54.15406799316406,
551
+ "learning_rate": 1.230769230769231e-05,
552
+ "loss": 0.4364,
553
+ "step": 1625
554
+ },
555
+ {
556
+ "epoch": 9.0,
557
+ "eval_accuracy": 0.8022328548644339,
558
+ "eval_loss": 0.9064925909042358,
559
+ "eval_macro_f1": 0.8025287782234445,
560
+ "eval_runtime": 1.1773,
561
+ "eval_samples_per_second": 532.59,
562
+ "eval_steps_per_second": 33.977,
563
+ "eval_weighted_f1": 0.8023587030058814,
564
+ "step": 1638
565
+ },
566
+ {
567
+ "epoch": 9.066206896551725,
568
+ "grad_norm": 87.1186294555664,
569
+ "learning_rate": 1.2155067155067155e-05,
570
+ "loss": 0.4131,
571
+ "step": 1650
572
+ },
573
+ {
574
+ "epoch": 9.204137931034483,
575
+ "grad_norm": 118.65606689453125,
576
+ "learning_rate": 1.2002442002442003e-05,
577
+ "loss": 0.3966,
578
+ "step": 1675
579
+ },
580
+ {
581
+ "epoch": 9.34206896551724,
582
+ "grad_norm": 43.27682113647461,
583
+ "learning_rate": 1.1849816849816852e-05,
584
+ "loss": 0.3994,
585
+ "step": 1700
586
+ },
587
+ {
588
+ "epoch": 9.48,
589
+ "grad_norm": 46.55112075805664,
590
+ "learning_rate": 1.1697191697191698e-05,
591
+ "loss": 0.3002,
592
+ "step": 1725
593
+ },
594
+ {
595
+ "epoch": 9.617931034482758,
596
+ "grad_norm": 299.67840576171875,
597
+ "learning_rate": 1.1544566544566545e-05,
598
+ "loss": 0.4456,
599
+ "step": 1750
600
+ },
601
+ {
602
+ "epoch": 9.755862068965516,
603
+ "grad_norm": 94.89724731445312,
604
+ "learning_rate": 1.1391941391941392e-05,
605
+ "loss": 0.4069,
606
+ "step": 1775
607
+ },
608
+ {
609
+ "epoch": 9.893793103448276,
610
+ "grad_norm": 83.06227111816406,
611
+ "learning_rate": 1.123931623931624e-05,
612
+ "loss": 0.3748,
613
+ "step": 1800
614
+ },
615
+ {
616
+ "epoch": 10.0,
617
+ "eval_accuracy": 0.773524720893142,
618
+ "eval_loss": 1.3610498905181885,
619
+ "eval_macro_f1": 0.7716054942384281,
620
+ "eval_runtime": 1.1853,
621
+ "eval_samples_per_second": 528.97,
622
+ "eval_steps_per_second": 33.746,
623
+ "eval_weighted_f1": 0.7719216877725383,
624
+ "step": 1820
625
+ },
626
+ {
627
+ "epoch": 10.027586206896551,
628
+ "grad_norm": 171.37094116210938,
629
+ "learning_rate": 1.1086691086691088e-05,
630
+ "loss": 0.398,
631
+ "step": 1825
632
+ },
633
+ {
634
+ "epoch": 10.16551724137931,
635
+ "grad_norm": 31.03558349609375,
636
+ "learning_rate": 1.0934065934065933e-05,
637
+ "loss": 0.2433,
638
+ "step": 1850
639
+ },
640
+ {
641
+ "epoch": 10.303448275862069,
642
+ "grad_norm": 5.404947757720947,
643
+ "learning_rate": 1.0781440781440782e-05,
644
+ "loss": 0.3894,
645
+ "step": 1875
646
+ },
647
+ {
648
+ "epoch": 10.441379310344828,
649
+ "grad_norm": 1.9526923894882202,
650
+ "learning_rate": 1.062881562881563e-05,
651
+ "loss": 0.2768,
652
+ "step": 1900
653
+ },
654
+ {
655
+ "epoch": 10.579310344827586,
656
+ "grad_norm": 3.5537567138671875,
657
+ "learning_rate": 1.0476190476190477e-05,
658
+ "loss": 0.3726,
659
+ "step": 1925
660
+ },
661
+ {
662
+ "epoch": 10.717241379310344,
663
+ "grad_norm": 218.355712890625,
664
+ "learning_rate": 1.0323565323565325e-05,
665
+ "loss": 0.2008,
666
+ "step": 1950
667
+ },
668
+ {
669
+ "epoch": 10.855172413793104,
670
+ "grad_norm": 0.5096381306648254,
671
+ "learning_rate": 1.0170940170940172e-05,
672
+ "loss": 0.5065,
673
+ "step": 1975
674
+ },
675
+ {
676
+ "epoch": 10.993103448275862,
677
+ "grad_norm": 0.12985798716545105,
678
+ "learning_rate": 1.0018315018315019e-05,
679
+ "loss": 0.2639,
680
+ "step": 2000
681
+ },
682
+ {
683
+ "epoch": 11.0,
684
+ "eval_accuracy": 0.79585326953748,
685
+ "eval_loss": 1.401996374130249,
686
+ "eval_macro_f1": 0.794472041204085,
687
+ "eval_runtime": 1.1683,
688
+ "eval_samples_per_second": 536.672,
689
+ "eval_steps_per_second": 34.237,
690
+ "eval_weighted_f1": 0.7946472239407022,
691
+ "step": 2002
692
+ },
693
+ {
694
+ "epoch": 11.126896551724139,
695
+ "grad_norm": 244.31393432617188,
696
+ "learning_rate": 9.865689865689867e-06,
697
+ "loss": 0.4166,
698
+ "step": 2025
699
+ },
700
+ {
701
+ "epoch": 11.264827586206897,
702
+ "grad_norm": 0.7453848123550415,
703
+ "learning_rate": 9.713064713064714e-06,
704
+ "loss": 0.1639,
705
+ "step": 2050
706
+ },
707
+ {
708
+ "epoch": 11.402758620689655,
709
+ "grad_norm": 1.1450600624084473,
710
+ "learning_rate": 9.560439560439562e-06,
711
+ "loss": 0.2126,
712
+ "step": 2075
713
+ },
714
+ {
715
+ "epoch": 11.540689655172415,
716
+ "grad_norm": 94.49410247802734,
717
+ "learning_rate": 9.407814407814409e-06,
718
+ "loss": 0.4276,
719
+ "step": 2100
720
+ },
721
+ {
722
+ "epoch": 11.678620689655173,
723
+ "grad_norm": 53.80741500854492,
724
+ "learning_rate": 9.255189255189256e-06,
725
+ "loss": 0.4125,
726
+ "step": 2125
727
+ },
728
+ {
729
+ "epoch": 11.81655172413793,
730
+ "grad_norm": 2.2708137035369873,
731
+ "learning_rate": 9.102564102564104e-06,
732
+ "loss": 0.3553,
733
+ "step": 2150
734
+ },
735
+ {
736
+ "epoch": 11.95448275862069,
737
+ "grad_norm": 16.314722061157227,
738
+ "learning_rate": 8.94993894993895e-06,
739
+ "loss": 0.213,
740
+ "step": 2175
741
+ },
742
+ {
743
+ "epoch": 12.0,
744
+ "eval_accuracy": 0.8086124401913876,
745
+ "eval_loss": 1.4752185344696045,
746
+ "eval_macro_f1": 0.8073371962771456,
747
+ "eval_runtime": 1.1292,
748
+ "eval_samples_per_second": 555.264,
749
+ "eval_steps_per_second": 35.424,
750
+ "eval_weighted_f1": 0.8074475414562201,
751
+ "step": 2184
752
+ },
753
+ {
754
+ "epoch": 12.088275862068965,
755
+ "grad_norm": 4.032958507537842,
756
+ "learning_rate": 8.797313797313797e-06,
757
+ "loss": 0.1548,
758
+ "step": 2200
759
+ },
760
+ {
761
+ "epoch": 12.226206896551725,
762
+ "grad_norm": 11.657094955444336,
763
+ "learning_rate": 8.644688644688646e-06,
764
+ "loss": 0.2706,
765
+ "step": 2225
766
+ },
767
+ {
768
+ "epoch": 12.364137931034483,
769
+ "grad_norm": 12.007180213928223,
770
+ "learning_rate": 8.492063492063492e-06,
771
+ "loss": 0.1725,
772
+ "step": 2250
773
+ },
774
+ {
775
+ "epoch": 12.50206896551724,
776
+ "grad_norm": 3.103738307952881,
777
+ "learning_rate": 8.33943833943834e-06,
778
+ "loss": 0.1077,
779
+ "step": 2275
780
+ },
781
+ {
782
+ "epoch": 12.64,
783
+ "grad_norm": 0.05534272640943527,
784
+ "learning_rate": 8.186813186813188e-06,
785
+ "loss": 0.2332,
786
+ "step": 2300
787
+ },
788
+ {
789
+ "epoch": 12.777931034482759,
790
+ "grad_norm": 0.27866512537002563,
791
+ "learning_rate": 8.034188034188036e-06,
792
+ "loss": 0.0977,
793
+ "step": 2325
794
+ },
795
+ {
796
+ "epoch": 12.915862068965517,
797
+ "grad_norm": 16.027671813964844,
798
+ "learning_rate": 7.881562881562883e-06,
799
+ "loss": 0.3268,
800
+ "step": 2350
801
+ },
802
+ {
803
+ "epoch": 13.0,
804
+ "eval_accuracy": 0.7974481658692185,
805
+ "eval_loss": 1.5549489259719849,
806
+ "eval_macro_f1": 0.7969154183788802,
807
+ "eval_runtime": 1.1851,
808
+ "eval_samples_per_second": 529.085,
809
+ "eval_steps_per_second": 33.753,
810
+ "eval_weighted_f1": 0.7970637749059234,
811
+ "step": 2366
812
+ },
813
+ {
814
+ "epoch": 13.049655172413793,
815
+ "grad_norm": 167.1862030029297,
816
+ "learning_rate": 7.72893772893773e-06,
817
+ "loss": 0.1916,
818
+ "step": 2375
819
+ },
820
+ {
821
+ "epoch": 13.187586206896551,
822
+ "grad_norm": 578.3486328125,
823
+ "learning_rate": 7.576312576312577e-06,
824
+ "loss": 0.2089,
825
+ "step": 2400
826
+ },
827
+ {
828
+ "epoch": 13.32551724137931,
829
+ "grad_norm": 2.9909238815307617,
830
+ "learning_rate": 7.423687423687424e-06,
831
+ "loss": 0.1702,
832
+ "step": 2425
833
+ },
834
+ {
835
+ "epoch": 13.463448275862069,
836
+ "grad_norm": 12.770936012268066,
837
+ "learning_rate": 7.271062271062271e-06,
838
+ "loss": 0.1574,
839
+ "step": 2450
840
+ },
841
+ {
842
+ "epoch": 13.601379310344827,
843
+ "grad_norm": 0.8478724360466003,
844
+ "learning_rate": 7.1184371184371195e-06,
845
+ "loss": 0.1238,
846
+ "step": 2475
847
+ },
848
+ {
849
+ "epoch": 13.739310344827587,
850
+ "grad_norm": 0.18776558339595795,
851
+ "learning_rate": 6.965811965811966e-06,
852
+ "loss": 0.0629,
853
+ "step": 2500
854
+ },
855
+ {
856
+ "epoch": 13.877241379310345,
857
+ "grad_norm": 10.055283546447754,
858
+ "learning_rate": 6.813186813186814e-06,
859
+ "loss": 0.3233,
860
+ "step": 2525
861
+ },
862
+ {
863
+ "epoch": 14.0,
864
+ "eval_accuracy": 0.8086124401913876,
865
+ "eval_loss": 1.6799039840698242,
866
+ "eval_macro_f1": 0.8084286581699676,
867
+ "eval_runtime": 1.1687,
868
+ "eval_samples_per_second": 536.477,
869
+ "eval_steps_per_second": 34.225,
870
+ "eval_weighted_f1": 0.8085553790065865,
871
+ "step": 2548
872
+ },
873
+ {
874
+ "epoch": 14.011034482758621,
875
+ "grad_norm": 276.6418762207031,
876
+ "learning_rate": 6.66056166056166e-06,
877
+ "loss": 0.2604,
878
+ "step": 2550
879
+ },
880
+ {
881
+ "epoch": 14.148965517241379,
882
+ "grad_norm": 23.674741744995117,
883
+ "learning_rate": 6.507936507936509e-06,
884
+ "loss": 0.1366,
885
+ "step": 2575
886
+ },
887
+ {
888
+ "epoch": 14.286896551724137,
889
+ "grad_norm": 456.3885803222656,
890
+ "learning_rate": 6.3553113553113555e-06,
891
+ "loss": 0.2356,
892
+ "step": 2600
893
+ },
894
+ {
895
+ "epoch": 14.424827586206897,
896
+ "grad_norm": 0.32852277159690857,
897
+ "learning_rate": 6.202686202686203e-06,
898
+ "loss": 0.1828,
899
+ "step": 2625
900
+ },
901
+ {
902
+ "epoch": 14.562758620689655,
903
+ "grad_norm": 0.16003744304180145,
904
+ "learning_rate": 6.05006105006105e-06,
905
+ "loss": 0.121,
906
+ "step": 2650
907
+ },
908
+ {
909
+ "epoch": 14.700689655172415,
910
+ "grad_norm": 199.4854278564453,
911
+ "learning_rate": 5.897435897435898e-06,
912
+ "loss": 0.26,
913
+ "step": 2675
914
+ },
915
+ {
916
+ "epoch": 14.838620689655173,
917
+ "grad_norm": 0.23486696183681488,
918
+ "learning_rate": 5.744810744810745e-06,
919
+ "loss": 0.2724,
920
+ "step": 2700
921
+ },
922
+ {
923
+ "epoch": 14.97655172413793,
924
+ "grad_norm": 552.55029296875,
925
+ "learning_rate": 5.592185592185592e-06,
926
+ "loss": 0.106,
927
+ "step": 2725
928
+ },
929
+ {
930
+ "epoch": 15.0,
931
+ "eval_accuracy": 0.8006379585326954,
932
+ "eval_loss": 1.7526615858078003,
933
+ "eval_macro_f1": 0.8013367773515058,
934
+ "eval_runtime": 1.1993,
935
+ "eval_samples_per_second": 522.785,
936
+ "eval_steps_per_second": 33.352,
937
+ "eval_weighted_f1": 0.8011896085394442,
938
+ "step": 2730
939
+ },
940
+ {
941
+ "epoch": 15.110344827586207,
942
+ "grad_norm": 0.046297576278448105,
943
+ "learning_rate": 5.43956043956044e-06,
944
+ "loss": 0.1396,
945
+ "step": 2750
946
+ },
947
+ {
948
+ "epoch": 15.248275862068965,
949
+ "grad_norm": 0.0772584080696106,
950
+ "learning_rate": 5.2869352869352875e-06,
951
+ "loss": 0.0958,
952
+ "step": 2775
953
+ },
954
+ {
955
+ "epoch": 15.386206896551725,
956
+ "grad_norm": 265.6419372558594,
957
+ "learning_rate": 5.134310134310134e-06,
958
+ "loss": 0.2009,
959
+ "step": 2800
960
+ },
961
+ {
962
+ "epoch": 15.524137931034483,
963
+ "grad_norm": 19.894752502441406,
964
+ "learning_rate": 4.9816849816849826e-06,
965
+ "loss": 0.073,
966
+ "step": 2825
967
+ },
968
+ {
969
+ "epoch": 15.662068965517241,
970
+ "grad_norm": 341.0168762207031,
971
+ "learning_rate": 4.829059829059829e-06,
972
+ "loss": 0.2365,
973
+ "step": 2850
974
+ },
975
+ {
976
+ "epoch": 15.8,
977
+ "grad_norm": 152.31944274902344,
978
+ "learning_rate": 4.676434676434677e-06,
979
+ "loss": 0.0926,
980
+ "step": 2875
981
+ },
982
+ {
983
+ "epoch": 15.937931034482759,
984
+ "grad_norm": 0.21693311631679535,
985
+ "learning_rate": 4.523809523809524e-06,
986
+ "loss": 0.159,
987
+ "step": 2900
988
+ },
989
+ {
990
+ "epoch": 16.0,
991
+ "eval_accuracy": 0.7990430622009569,
992
+ "eval_loss": 1.797470211982727,
993
+ "eval_macro_f1": 0.799145012661444,
994
+ "eval_runtime": 1.1154,
995
+ "eval_samples_per_second": 562.148,
996
+ "eval_steps_per_second": 35.863,
997
+ "eval_weighted_f1": 0.7992051011258572,
998
+ "step": 2912
999
+ },
1000
+ {
1001
+ "epoch": 16.071724137931035,
1002
+ "grad_norm": 395.2620849609375,
1003
+ "learning_rate": 4.371184371184372e-06,
1004
+ "loss": 0.1296,
1005
+ "step": 2925
1006
+ },
1007
+ {
1008
+ "epoch": 16.209655172413793,
1009
+ "grad_norm": 0.18170881271362305,
1010
+ "learning_rate": 4.218559218559219e-06,
1011
+ "loss": 0.1053,
1012
+ "step": 2950
1013
+ },
1014
+ {
1015
+ "epoch": 16.34758620689655,
1016
+ "grad_norm": 420.1138610839844,
1017
+ "learning_rate": 4.065934065934066e-06,
1018
+ "loss": 0.0738,
1019
+ "step": 2975
1020
+ },
1021
+ {
1022
+ "epoch": 16.48551724137931,
1023
+ "grad_norm": 0.03433357924222946,
1024
+ "learning_rate": 3.913308913308914e-06,
1025
+ "loss": 0.0067,
1026
+ "step": 3000
1027
+ },
1028
+ {
1029
+ "epoch": 16.623448275862067,
1030
+ "grad_norm": 0.047096770256757736,
1031
+ "learning_rate": 3.760683760683761e-06,
1032
+ "loss": 0.1146,
1033
+ "step": 3025
1034
+ },
1035
+ {
1036
+ "epoch": 16.76137931034483,
1037
+ "grad_norm": 0.02566961944103241,
1038
+ "learning_rate": 3.6080586080586083e-06,
1039
+ "loss": 0.0096,
1040
+ "step": 3050
1041
+ },
1042
+ {
1043
+ "epoch": 16.899310344827587,
1044
+ "grad_norm": 1.223204255104065,
1045
+ "learning_rate": 3.4554334554334555e-06,
1046
+ "loss": 0.0274,
1047
+ "step": 3075
1048
+ },
1049
+ {
1050
+ "epoch": 17.0,
1051
+ "eval_accuracy": 0.8022328548644339,
1052
+ "eval_loss": 1.8602946996688843,
1053
+ "eval_macro_f1": 0.8019165313579076,
1054
+ "eval_runtime": 1.18,
1055
+ "eval_samples_per_second": 531.372,
1056
+ "eval_steps_per_second": 33.899,
1057
+ "eval_weighted_f1": 0.802071879926345,
1058
+ "step": 3094
1059
+ }
1060
+ ],
1061
+ "logging_steps": 25,
1062
+ "max_steps": 3640,
1063
+ "num_input_tokens_seen": 0,
1064
+ "num_train_epochs": 20,
1065
+ "save_steps": 500,
1066
+ "stateful_callbacks": {
1067
+ "EarlyStoppingCallback": {
1068
+ "args": {
1069
+ "early_stopping_patience": 3,
1070
+ "early_stopping_threshold": 0.0
1071
+ },
1072
+ "attributes": {
1073
+ "early_stopping_patience_counter": 3
1074
+ }
1075
+ },
1076
+ "TrainerControl": {
1077
+ "args": {
1078
+ "should_epoch_stop": false,
1079
+ "should_evaluate": false,
1080
+ "should_log": false,
1081
+ "should_save": true,
1082
+ "should_training_stop": true
1083
+ },
1084
+ "attributes": {}
1085
+ }
1086
+ },
1087
+ "total_flos": 0.0,
1088
+ "train_batch_size": 8,
1089
+ "trial_name": null,
1090
+ "trial_params": null
1091
+ }
ood-vifactcheck-xlmr-source-source-00-s43-20260901T164831Z-fbd63439/trainer/checkpoint-2548/special_tokens_map.json ADDED
@@ -0,0 +1,15 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "bos_token": "<s>",
3
+ "cls_token": "<s>",
4
+ "eos_token": "</s>",
5
+ "mask_token": {
6
+ "content": "<mask>",
7
+ "lstrip": true,
8
+ "normalized": false,
9
+ "rstrip": false,
10
+ "single_word": false
11
+ },
12
+ "pad_token": "<pad>",
13
+ "sep_token": "</s>",
14
+ "unk_token": "<unk>"
15
+ }
ood-vifactcheck-xlmr-source-source-00-s43-20260901T164831Z-fbd63439/trainer/checkpoint-2548/tokenizer_config.json ADDED
@@ -0,0 +1,55 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "added_tokens_decoder": {
3
+ "0": {
4
+ "content": "<s>",
5
+ "lstrip": false,
6
+ "normalized": false,
7
+ "rstrip": false,
8
+ "single_word": false,
9
+ "special": true
10
+ },
11
+ "1": {
12
+ "content": "<pad>",
13
+ "lstrip": false,
14
+ "normalized": false,
15
+ "rstrip": false,
16
+ "single_word": false,
17
+ "special": true
18
+ },
19
+ "2": {
20
+ "content": "</s>",
21
+ "lstrip": false,
22
+ "normalized": false,
23
+ "rstrip": false,
24
+ "single_word": false,
25
+ "special": true
26
+ },
27
+ "3": {
28
+ "content": "<unk>",
29
+ "lstrip": false,
30
+ "normalized": false,
31
+ "rstrip": false,
32
+ "single_word": false,
33
+ "special": true
34
+ },
35
+ "250001": {
36
+ "content": "<mask>",
37
+ "lstrip": true,
38
+ "normalized": false,
39
+ "rstrip": false,
40
+ "single_word": false,
41
+ "special": true
42
+ }
43
+ },
44
+ "bos_token": "<s>",
45
+ "clean_up_tokenization_spaces": false,
46
+ "cls_token": "<s>",
47
+ "eos_token": "</s>",
48
+ "extra_special_tokens": {},
49
+ "mask_token": "<mask>",
50
+ "model_max_length": 512,
51
+ "pad_token": "<pad>",
52
+ "sep_token": "</s>",
53
+ "tokenizer_class": "XLMRobertaTokenizer",
54
+ "unk_token": "<unk>"
55
+ }
ood-vifactcheck-xlmr-source-source-00-s43-20260901T164831Z-fbd63439/trainer/checkpoint-2548/trainer_state.json ADDED
@@ -0,0 +1,904 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_global_step": 2548,
3
+ "best_metric": 0.8244676436475006,
4
+ "best_model_checkpoint": "/root/VNEncoder/results/checkpoints/ood-vifactcheck-xlmr-source-source-00-s43-20260901T164831Z-fbd63439/trainer/checkpoint-2548",
5
+ "epoch": 14.0,
6
+ "eval_steps": 500,
7
+ "global_step": 2548,
8
+ "is_hyper_param_search": false,
9
+ "is_local_process_zero": true,
10
+ "is_world_process_zero": true,
11
+ "log_history": [
12
+ {
13
+ "epoch": 0.13793103448275862,
14
+ "grad_norm": 20.084362030029297,
15
+ "learning_rate": 1.3186813186813187e-06,
16
+ "loss": 4.7288,
17
+ "step": 25
18
+ },
19
+ {
20
+ "epoch": 0.27586206896551724,
21
+ "grad_norm": 5.729361534118652,
22
+ "learning_rate": 2.6923076923076923e-06,
23
+ "loss": 4.684,
24
+ "step": 50
25
+ },
26
+ {
27
+ "epoch": 0.41379310344827586,
28
+ "grad_norm": 7.694893836975098,
29
+ "learning_rate": 4.065934065934066e-06,
30
+ "loss": 4.5741,
31
+ "step": 75
32
+ },
33
+ {
34
+ "epoch": 0.5517241379310345,
35
+ "grad_norm": 8.551424980163574,
36
+ "learning_rate": 5.43956043956044e-06,
37
+ "loss": 4.4816,
38
+ "step": 100
39
+ },
40
+ {
41
+ "epoch": 0.6896551724137931,
42
+ "grad_norm": 35.368587493896484,
43
+ "learning_rate": 6.813186813186814e-06,
44
+ "loss": 4.5308,
45
+ "step": 125
46
+ },
47
+ {
48
+ "epoch": 0.8275862068965517,
49
+ "grad_norm": 44.43610763549805,
50
+ "learning_rate": 8.186813186813188e-06,
51
+ "loss": 4.444,
52
+ "step": 150
53
+ },
54
+ {
55
+ "epoch": 0.9655172413793104,
56
+ "grad_norm": 24.932857513427734,
57
+ "learning_rate": 9.560439560439562e-06,
58
+ "loss": 4.3658,
59
+ "step": 175
60
+ },
61
+ {
62
+ "epoch": 1.0,
63
+ "eval_accuracy": 0.4258373205741627,
64
+ "eval_loss": 1.0057348012924194,
65
+ "eval_macro_f1": 0.3397112691230338,
66
+ "eval_runtime": 1.1171,
67
+ "eval_samples_per_second": 561.294,
68
+ "eval_steps_per_second": 35.808,
69
+ "eval_weighted_f1": 0.3524679654710614,
70
+ "step": 182
71
+ },
72
+ {
73
+ "epoch": 1.0993103448275863,
74
+ "grad_norm": 61.700660705566406,
75
+ "learning_rate": 1.0934065934065933e-05,
76
+ "loss": 4.1429,
77
+ "step": 200
78
+ },
79
+ {
80
+ "epoch": 1.237241379310345,
81
+ "grad_norm": 46.13859558105469,
82
+ "learning_rate": 1.230769230769231e-05,
83
+ "loss": 3.9467,
84
+ "step": 225
85
+ },
86
+ {
87
+ "epoch": 1.3751724137931034,
88
+ "grad_norm": 46.452125549316406,
89
+ "learning_rate": 1.3681318681318683e-05,
90
+ "loss": 3.6127,
91
+ "step": 250
92
+ },
93
+ {
94
+ "epoch": 1.513103448275862,
95
+ "grad_norm": 59.73109436035156,
96
+ "learning_rate": 1.5054945054945056e-05,
97
+ "loss": 3.5063,
98
+ "step": 275
99
+ },
100
+ {
101
+ "epoch": 1.6510344827586207,
102
+ "grad_norm": 145.90650939941406,
103
+ "learning_rate": 1.642857142857143e-05,
104
+ "loss": 3.3779,
105
+ "step": 300
106
+ },
107
+ {
108
+ "epoch": 1.7889655172413792,
109
+ "grad_norm": 54.86250305175781,
110
+ "learning_rate": 1.78021978021978e-05,
111
+ "loss": 3.0618,
112
+ "step": 325
113
+ },
114
+ {
115
+ "epoch": 1.926896551724138,
116
+ "grad_norm": 52.89915084838867,
117
+ "learning_rate": 1.9175824175824178e-05,
118
+ "loss": 3.1781,
119
+ "step": 350
120
+ },
121
+ {
122
+ "epoch": 2.0,
123
+ "eval_accuracy": 0.7113237639553429,
124
+ "eval_loss": 0.7212908864021301,
125
+ "eval_macro_f1": 0.7123417832526394,
126
+ "eval_runtime": 1.0768,
127
+ "eval_samples_per_second": 582.292,
128
+ "eval_steps_per_second": 37.148,
129
+ "eval_weighted_f1": 0.712662576565873,
130
+ "step": 364
131
+ },
132
+ {
133
+ "epoch": 2.0606896551724136,
134
+ "grad_norm": 84.83079528808594,
135
+ "learning_rate": 1.993894993894994e-05,
136
+ "loss": 2.6491,
137
+ "step": 375
138
+ },
139
+ {
140
+ "epoch": 2.1986206896551725,
141
+ "grad_norm": 108.39063262939453,
142
+ "learning_rate": 1.9786324786324788e-05,
143
+ "loss": 2.7958,
144
+ "step": 400
145
+ },
146
+ {
147
+ "epoch": 2.336551724137931,
148
+ "grad_norm": 96.23841094970703,
149
+ "learning_rate": 1.9633699633699634e-05,
150
+ "loss": 2.4406,
151
+ "step": 425
152
+ },
153
+ {
154
+ "epoch": 2.47448275862069,
155
+ "grad_norm": 85.79228973388672,
156
+ "learning_rate": 1.948107448107448e-05,
157
+ "loss": 2.6577,
158
+ "step": 450
159
+ },
160
+ {
161
+ "epoch": 2.6124137931034483,
162
+ "grad_norm": 42.99919509887695,
163
+ "learning_rate": 1.932844932844933e-05,
164
+ "loss": 2.5698,
165
+ "step": 475
166
+ },
167
+ {
168
+ "epoch": 2.750344827586207,
169
+ "grad_norm": 82.11553192138672,
170
+ "learning_rate": 1.9175824175824178e-05,
171
+ "loss": 2.465,
172
+ "step": 500
173
+ },
174
+ {
175
+ "epoch": 2.8882758620689657,
176
+ "grad_norm": 64.50354766845703,
177
+ "learning_rate": 1.9023199023199025e-05,
178
+ "loss": 2.6201,
179
+ "step": 525
180
+ },
181
+ {
182
+ "epoch": 3.0,
183
+ "eval_accuracy": 0.7543859649122807,
184
+ "eval_loss": 0.5967164039611816,
185
+ "eval_macro_f1": 0.7556428199642172,
186
+ "eval_runtime": 1.0678,
187
+ "eval_samples_per_second": 587.187,
188
+ "eval_steps_per_second": 37.46,
189
+ "eval_weighted_f1": 0.7552112588758818,
190
+ "step": 546
191
+ },
192
+ {
193
+ "epoch": 3.0220689655172412,
194
+ "grad_norm": 57.78512954711914,
195
+ "learning_rate": 1.887057387057387e-05,
196
+ "loss": 2.2413,
197
+ "step": 550
198
+ },
199
+ {
200
+ "epoch": 3.16,
201
+ "grad_norm": 49.69960403442383,
202
+ "learning_rate": 1.8717948717948718e-05,
203
+ "loss": 2.0975,
204
+ "step": 575
205
+ },
206
+ {
207
+ "epoch": 3.2979310344827586,
208
+ "grad_norm": 70.23579406738281,
209
+ "learning_rate": 1.8565323565323568e-05,
210
+ "loss": 1.8645,
211
+ "step": 600
212
+ },
213
+ {
214
+ "epoch": 3.435862068965517,
215
+ "grad_norm": 139.99871826171875,
216
+ "learning_rate": 1.8412698412698415e-05,
217
+ "loss": 1.9666,
218
+ "step": 625
219
+ },
220
+ {
221
+ "epoch": 3.573793103448276,
222
+ "grad_norm": 41.8527717590332,
223
+ "learning_rate": 1.826007326007326e-05,
224
+ "loss": 1.8205,
225
+ "step": 650
226
+ },
227
+ {
228
+ "epoch": 3.7117241379310344,
229
+ "grad_norm": 109.82010650634766,
230
+ "learning_rate": 1.8107448107448108e-05,
231
+ "loss": 1.6387,
232
+ "step": 675
233
+ },
234
+ {
235
+ "epoch": 3.849655172413793,
236
+ "grad_norm": 107.30518341064453,
237
+ "learning_rate": 1.7954822954822958e-05,
238
+ "loss": 2.1496,
239
+ "step": 700
240
+ },
241
+ {
242
+ "epoch": 3.9875862068965517,
243
+ "grad_norm": 83.41844177246094,
244
+ "learning_rate": 1.78021978021978e-05,
245
+ "loss": 1.9329,
246
+ "step": 725
247
+ },
248
+ {
249
+ "epoch": 4.0,
250
+ "eval_accuracy": 0.7767145135566188,
251
+ "eval_loss": 0.5591686964035034,
252
+ "eval_macro_f1": 0.7763996474814833,
253
+ "eval_runtime": 1.1489,
254
+ "eval_samples_per_second": 545.719,
255
+ "eval_steps_per_second": 34.815,
256
+ "eval_weighted_f1": 0.7766707740817129,
257
+ "step": 728
258
+ },
259
+ {
260
+ "epoch": 4.121379310344827,
261
+ "grad_norm": 38.20743942260742,
262
+ "learning_rate": 1.764957264957265e-05,
263
+ "loss": 1.501,
264
+ "step": 750
265
+ },
266
+ {
267
+ "epoch": 4.259310344827586,
268
+ "grad_norm": 101.95227813720703,
269
+ "learning_rate": 1.74969474969475e-05,
270
+ "loss": 1.5836,
271
+ "step": 775
272
+ },
273
+ {
274
+ "epoch": 4.397241379310345,
275
+ "grad_norm": 48.2017936706543,
276
+ "learning_rate": 1.7344322344322345e-05,
277
+ "loss": 1.3153,
278
+ "step": 800
279
+ },
280
+ {
281
+ "epoch": 4.535172413793103,
282
+ "grad_norm": 56.81407928466797,
283
+ "learning_rate": 1.7191697191697195e-05,
284
+ "loss": 1.5429,
285
+ "step": 825
286
+ },
287
+ {
288
+ "epoch": 4.673103448275862,
289
+ "grad_norm": 101.91703796386719,
290
+ "learning_rate": 1.703907203907204e-05,
291
+ "loss": 1.3026,
292
+ "step": 850
293
+ },
294
+ {
295
+ "epoch": 4.811034482758621,
296
+ "grad_norm": 75.37382507324219,
297
+ "learning_rate": 1.688644688644689e-05,
298
+ "loss": 1.2796,
299
+ "step": 875
300
+ },
301
+ {
302
+ "epoch": 4.94896551724138,
303
+ "grad_norm": 123.11703491210938,
304
+ "learning_rate": 1.6733821733821735e-05,
305
+ "loss": 1.4378,
306
+ "step": 900
307
+ },
308
+ {
309
+ "epoch": 5.0,
310
+ "eval_accuracy": 0.7862838915470495,
311
+ "eval_loss": 0.7190022468566895,
312
+ "eval_macro_f1": 0.7867678171448537,
313
+ "eval_runtime": 1.1303,
314
+ "eval_samples_per_second": 554.703,
315
+ "eval_steps_per_second": 35.388,
316
+ "eval_weighted_f1": 0.7868702555186253,
317
+ "step": 910
318
+ },
319
+ {
320
+ "epoch": 5.082758620689655,
321
+ "grad_norm": 111.80149841308594,
322
+ "learning_rate": 1.6581196581196585e-05,
323
+ "loss": 1.0587,
324
+ "step": 925
325
+ },
326
+ {
327
+ "epoch": 5.220689655172414,
328
+ "grad_norm": 101.65257263183594,
329
+ "learning_rate": 1.642857142857143e-05,
330
+ "loss": 1.0306,
331
+ "step": 950
332
+ },
333
+ {
334
+ "epoch": 5.358620689655172,
335
+ "grad_norm": 58.97084045410156,
336
+ "learning_rate": 1.627594627594628e-05,
337
+ "loss": 1.1596,
338
+ "step": 975
339
+ },
340
+ {
341
+ "epoch": 5.496551724137931,
342
+ "grad_norm": 63.96665573120117,
343
+ "learning_rate": 1.6123321123321125e-05,
344
+ "loss": 1.2873,
345
+ "step": 1000
346
+ },
347
+ {
348
+ "epoch": 5.63448275862069,
349
+ "grad_norm": 104.3017349243164,
350
+ "learning_rate": 1.5970695970695972e-05,
351
+ "loss": 1.0208,
352
+ "step": 1025
353
+ },
354
+ {
355
+ "epoch": 5.772413793103448,
356
+ "grad_norm": 108.5999526977539,
357
+ "learning_rate": 1.581807081807082e-05,
358
+ "loss": 1.2335,
359
+ "step": 1050
360
+ },
361
+ {
362
+ "epoch": 5.910344827586207,
363
+ "grad_norm": 94.29696655273438,
364
+ "learning_rate": 1.5665445665445665e-05,
365
+ "loss": 1.2207,
366
+ "step": 1075
367
+ },
368
+ {
369
+ "epoch": 6.0,
370
+ "eval_accuracy": 0.7878787878787878,
371
+ "eval_loss": 0.6836919784545898,
372
+ "eval_macro_f1": 0.7879398797194143,
373
+ "eval_runtime": 1.1922,
374
+ "eval_samples_per_second": 525.905,
375
+ "eval_steps_per_second": 33.551,
376
+ "eval_weighted_f1": 0.7879111724300937,
377
+ "step": 1092
378
+ },
379
+ {
380
+ "epoch": 6.0441379310344825,
381
+ "grad_norm": 103.21680450439453,
382
+ "learning_rate": 1.5512820512820516e-05,
383
+ "loss": 0.9723,
384
+ "step": 1100
385
+ },
386
+ {
387
+ "epoch": 6.182068965517241,
388
+ "grad_norm": 104.1303482055664,
389
+ "learning_rate": 1.5360195360195362e-05,
390
+ "loss": 0.7962,
391
+ "step": 1125
392
+ },
393
+ {
394
+ "epoch": 6.32,
395
+ "grad_norm": 56.65522003173828,
396
+ "learning_rate": 1.5207570207570207e-05,
397
+ "loss": 0.7656,
398
+ "step": 1150
399
+ },
400
+ {
401
+ "epoch": 6.457931034482758,
402
+ "grad_norm": 136.6097869873047,
403
+ "learning_rate": 1.5054945054945056e-05,
404
+ "loss": 0.6219,
405
+ "step": 1175
406
+ },
407
+ {
408
+ "epoch": 6.595862068965517,
409
+ "grad_norm": 116.55268096923828,
410
+ "learning_rate": 1.4902319902319904e-05,
411
+ "loss": 0.7777,
412
+ "step": 1200
413
+ },
414
+ {
415
+ "epoch": 6.733793103448276,
416
+ "grad_norm": 61.935054779052734,
417
+ "learning_rate": 1.4749694749694752e-05,
418
+ "loss": 0.8592,
419
+ "step": 1225
420
+ },
421
+ {
422
+ "epoch": 6.871724137931034,
423
+ "grad_norm": 34.48295593261719,
424
+ "learning_rate": 1.4597069597069597e-05,
425
+ "loss": 0.9455,
426
+ "step": 1250
427
+ },
428
+ {
429
+ "epoch": 7.0,
430
+ "eval_accuracy": 0.7830940988835726,
431
+ "eval_loss": 0.7971011996269226,
432
+ "eval_macro_f1": 0.7841797417521805,
433
+ "eval_runtime": 1.1578,
434
+ "eval_samples_per_second": 541.551,
435
+ "eval_steps_per_second": 34.549,
436
+ "eval_weighted_f1": 0.7845438631046493,
437
+ "step": 1274
438
+ },
439
+ {
440
+ "epoch": 7.0055172413793105,
441
+ "grad_norm": 178.21438598632812,
442
+ "learning_rate": 1.4444444444444446e-05,
443
+ "loss": 0.7349,
444
+ "step": 1275
445
+ },
446
+ {
447
+ "epoch": 7.1434482758620685,
448
+ "grad_norm": 101.17195892333984,
449
+ "learning_rate": 1.4291819291819293e-05,
450
+ "loss": 0.6712,
451
+ "step": 1300
452
+ },
453
+ {
454
+ "epoch": 7.281379310344827,
455
+ "grad_norm": 60.77378845214844,
456
+ "learning_rate": 1.4139194139194141e-05,
457
+ "loss": 0.6681,
458
+ "step": 1325
459
+ },
460
+ {
461
+ "epoch": 7.419310344827586,
462
+ "grad_norm": 227.9224090576172,
463
+ "learning_rate": 1.3986568986568988e-05,
464
+ "loss": 0.5815,
465
+ "step": 1350
466
+ },
467
+ {
468
+ "epoch": 7.557241379310344,
469
+ "grad_norm": 93.1969223022461,
470
+ "learning_rate": 1.3833943833943834e-05,
471
+ "loss": 0.7168,
472
+ "step": 1375
473
+ },
474
+ {
475
+ "epoch": 7.695172413793103,
476
+ "grad_norm": 53.236114501953125,
477
+ "learning_rate": 1.3681318681318683e-05,
478
+ "loss": 0.8488,
479
+ "step": 1400
480
+ },
481
+ {
482
+ "epoch": 7.833103448275862,
483
+ "grad_norm": 84.6240005493164,
484
+ "learning_rate": 1.3528693528693531e-05,
485
+ "loss": 0.5962,
486
+ "step": 1425
487
+ },
488
+ {
489
+ "epoch": 7.971034482758621,
490
+ "grad_norm": 70.12305450439453,
491
+ "learning_rate": 1.3376068376068376e-05,
492
+ "loss": 0.6278,
493
+ "step": 1450
494
+ },
495
+ {
496
+ "epoch": 8.0,
497
+ "eval_accuracy": 0.7910685805422647,
498
+ "eval_loss": 0.9096579551696777,
499
+ "eval_macro_f1": 0.7890004378104845,
500
+ "eval_runtime": 1.0983,
501
+ "eval_samples_per_second": 570.88,
502
+ "eval_steps_per_second": 36.42,
503
+ "eval_weighted_f1": 0.7894828129259822,
504
+ "step": 1456
505
+ },
506
+ {
507
+ "epoch": 8.104827586206897,
508
+ "grad_norm": 181.92086791992188,
509
+ "learning_rate": 1.3223443223443224e-05,
510
+ "loss": 0.5441,
511
+ "step": 1475
512
+ },
513
+ {
514
+ "epoch": 8.242758620689655,
515
+ "grad_norm": 91.28218841552734,
516
+ "learning_rate": 1.3070818070818073e-05,
517
+ "loss": 0.6527,
518
+ "step": 1500
519
+ },
520
+ {
521
+ "epoch": 8.380689655172414,
522
+ "grad_norm": 211.7768096923828,
523
+ "learning_rate": 1.291819291819292e-05,
524
+ "loss": 0.4916,
525
+ "step": 1525
526
+ },
527
+ {
528
+ "epoch": 8.518620689655172,
529
+ "grad_norm": 32.13935089111328,
530
+ "learning_rate": 1.2765567765567766e-05,
531
+ "loss": 0.4391,
532
+ "step": 1550
533
+ },
534
+ {
535
+ "epoch": 8.65655172413793,
536
+ "grad_norm": 295.61663818359375,
537
+ "learning_rate": 1.2612942612942613e-05,
538
+ "loss": 0.5198,
539
+ "step": 1575
540
+ },
541
+ {
542
+ "epoch": 8.79448275862069,
543
+ "grad_norm": 32.63080596923828,
544
+ "learning_rate": 1.2460317460317461e-05,
545
+ "loss": 0.4442,
546
+ "step": 1600
547
+ },
548
+ {
549
+ "epoch": 8.932413793103448,
550
+ "grad_norm": 83.26895904541016,
551
+ "learning_rate": 1.230769230769231e-05,
552
+ "loss": 0.6965,
553
+ "step": 1625
554
+ },
555
+ {
556
+ "epoch": 9.0,
557
+ "eval_accuracy": 0.7974481658692185,
558
+ "eval_loss": 1.1046565771102905,
559
+ "eval_macro_f1": 0.7953705175062238,
560
+ "eval_runtime": 1.1067,
561
+ "eval_samples_per_second": 566.552,
562
+ "eval_steps_per_second": 36.144,
563
+ "eval_weighted_f1": 0.7957420601762267,
564
+ "step": 1638
565
+ },
566
+ {
567
+ "epoch": 9.066206896551725,
568
+ "grad_norm": 0.9721019268035889,
569
+ "learning_rate": 1.2155067155067155e-05,
570
+ "loss": 0.3916,
571
+ "step": 1650
572
+ },
573
+ {
574
+ "epoch": 9.204137931034483,
575
+ "grad_norm": 246.64122009277344,
576
+ "learning_rate": 1.2002442002442003e-05,
577
+ "loss": 0.5128,
578
+ "step": 1675
579
+ },
580
+ {
581
+ "epoch": 9.34206896551724,
582
+ "grad_norm": 66.1498794555664,
583
+ "learning_rate": 1.1849816849816852e-05,
584
+ "loss": 0.2748,
585
+ "step": 1700
586
+ },
587
+ {
588
+ "epoch": 9.48,
589
+ "grad_norm": 18.061092376708984,
590
+ "learning_rate": 1.1697191697191698e-05,
591
+ "loss": 0.421,
592
+ "step": 1725
593
+ },
594
+ {
595
+ "epoch": 9.617931034482758,
596
+ "grad_norm": 7.748656272888184,
597
+ "learning_rate": 1.1544566544566545e-05,
598
+ "loss": 0.3691,
599
+ "step": 1750
600
+ },
601
+ {
602
+ "epoch": 9.755862068965516,
603
+ "grad_norm": 139.55853271484375,
604
+ "learning_rate": 1.1391941391941392e-05,
605
+ "loss": 0.3909,
606
+ "step": 1775
607
+ },
608
+ {
609
+ "epoch": 9.893793103448276,
610
+ "grad_norm": 165.58914184570312,
611
+ "learning_rate": 1.123931623931624e-05,
612
+ "loss": 0.3689,
613
+ "step": 1800
614
+ },
615
+ {
616
+ "epoch": 10.0,
617
+ "eval_accuracy": 0.7926634768740032,
618
+ "eval_loss": 1.3370226621627808,
619
+ "eval_macro_f1": 0.7928866612894071,
620
+ "eval_runtime": 1.0674,
621
+ "eval_samples_per_second": 587.409,
622
+ "eval_steps_per_second": 37.474,
623
+ "eval_weighted_f1": 0.7930633611359907,
624
+ "step": 1820
625
+ },
626
+ {
627
+ "epoch": 10.027586206896551,
628
+ "grad_norm": 2.90893292427063,
629
+ "learning_rate": 1.1086691086691088e-05,
630
+ "loss": 0.2536,
631
+ "step": 1825
632
+ },
633
+ {
634
+ "epoch": 10.16551724137931,
635
+ "grad_norm": 69.7308120727539,
636
+ "learning_rate": 1.0934065934065933e-05,
637
+ "loss": 0.4142,
638
+ "step": 1850
639
+ },
640
+ {
641
+ "epoch": 10.303448275862069,
642
+ "grad_norm": 0.6135218739509583,
643
+ "learning_rate": 1.0781440781440782e-05,
644
+ "loss": 0.2497,
645
+ "step": 1875
646
+ },
647
+ {
648
+ "epoch": 10.441379310344828,
649
+ "grad_norm": 1.3770861625671387,
650
+ "learning_rate": 1.062881562881563e-05,
651
+ "loss": 0.2115,
652
+ "step": 1900
653
+ },
654
+ {
655
+ "epoch": 10.579310344827586,
656
+ "grad_norm": 284.72198486328125,
657
+ "learning_rate": 1.0476190476190477e-05,
658
+ "loss": 0.596,
659
+ "step": 1925
660
+ },
661
+ {
662
+ "epoch": 10.717241379310344,
663
+ "grad_norm": 2.9492316246032715,
664
+ "learning_rate": 1.0323565323565325e-05,
665
+ "loss": 0.4342,
666
+ "step": 1950
667
+ },
668
+ {
669
+ "epoch": 10.855172413793104,
670
+ "grad_norm": 68.05549621582031,
671
+ "learning_rate": 1.0170940170940172e-05,
672
+ "loss": 0.2668,
673
+ "step": 1975
674
+ },
675
+ {
676
+ "epoch": 10.993103448275862,
677
+ "grad_norm": 106.35601043701172,
678
+ "learning_rate": 1.0018315018315019e-05,
679
+ "loss": 0.4008,
680
+ "step": 2000
681
+ },
682
+ {
683
+ "epoch": 11.0,
684
+ "eval_accuracy": 0.773524720893142,
685
+ "eval_loss": 1.292651653289795,
686
+ "eval_macro_f1": 0.7737703583664469,
687
+ "eval_runtime": 1.2845,
688
+ "eval_samples_per_second": 488.141,
689
+ "eval_steps_per_second": 31.141,
690
+ "eval_weighted_f1": 0.7740535430140322,
691
+ "step": 2002
692
+ },
693
+ {
694
+ "epoch": 11.126896551724139,
695
+ "grad_norm": 12.272939682006836,
696
+ "learning_rate": 9.865689865689867e-06,
697
+ "loss": 0.1852,
698
+ "step": 2025
699
+ },
700
+ {
701
+ "epoch": 11.264827586206897,
702
+ "grad_norm": 0.17907734215259552,
703
+ "learning_rate": 9.713064713064714e-06,
704
+ "loss": 0.2774,
705
+ "step": 2050
706
+ },
707
+ {
708
+ "epoch": 11.402758620689655,
709
+ "grad_norm": 143.93467712402344,
710
+ "learning_rate": 9.560439560439562e-06,
711
+ "loss": 0.2405,
712
+ "step": 2075
713
+ },
714
+ {
715
+ "epoch": 11.540689655172415,
716
+ "grad_norm": 139.8834991455078,
717
+ "learning_rate": 9.407814407814409e-06,
718
+ "loss": 0.2925,
719
+ "step": 2100
720
+ },
721
+ {
722
+ "epoch": 11.678620689655173,
723
+ "grad_norm": 12.48167610168457,
724
+ "learning_rate": 9.255189255189256e-06,
725
+ "loss": 0.3291,
726
+ "step": 2125
727
+ },
728
+ {
729
+ "epoch": 11.81655172413793,
730
+ "grad_norm": 213.54396057128906,
731
+ "learning_rate": 9.102564102564104e-06,
732
+ "loss": 0.3811,
733
+ "step": 2150
734
+ },
735
+ {
736
+ "epoch": 11.95448275862069,
737
+ "grad_norm": 3.9226012229919434,
738
+ "learning_rate": 8.94993894993895e-06,
739
+ "loss": 0.272,
740
+ "step": 2175
741
+ },
742
+ {
743
+ "epoch": 12.0,
744
+ "eval_accuracy": 0.8070175438596491,
745
+ "eval_loss": 1.3421677350997925,
746
+ "eval_macro_f1": 0.8049096987990539,
747
+ "eval_runtime": 1.1904,
748
+ "eval_samples_per_second": 526.708,
749
+ "eval_steps_per_second": 33.602,
750
+ "eval_weighted_f1": 0.8049808148241474,
751
+ "step": 2184
752
+ },
753
+ {
754
+ "epoch": 12.088275862068965,
755
+ "grad_norm": 164.14903259277344,
756
+ "learning_rate": 8.797313797313797e-06,
757
+ "loss": 0.2303,
758
+ "step": 2200
759
+ },
760
+ {
761
+ "epoch": 12.226206896551725,
762
+ "grad_norm": 1.2080793380737305,
763
+ "learning_rate": 8.644688644688646e-06,
764
+ "loss": 0.2818,
765
+ "step": 2225
766
+ },
767
+ {
768
+ "epoch": 12.364137931034483,
769
+ "grad_norm": 0.2107601761817932,
770
+ "learning_rate": 8.492063492063492e-06,
771
+ "loss": 0.166,
772
+ "step": 2250
773
+ },
774
+ {
775
+ "epoch": 12.50206896551724,
776
+ "grad_norm": 597.9608154296875,
777
+ "learning_rate": 8.33943833943834e-06,
778
+ "loss": 0.2335,
779
+ "step": 2275
780
+ },
781
+ {
782
+ "epoch": 12.64,
783
+ "grad_norm": 0.06262794137001038,
784
+ "learning_rate": 8.186813186813188e-06,
785
+ "loss": 0.3141,
786
+ "step": 2300
787
+ },
788
+ {
789
+ "epoch": 12.777931034482759,
790
+ "grad_norm": 183.90464782714844,
791
+ "learning_rate": 8.034188034188036e-06,
792
+ "loss": 0.1393,
793
+ "step": 2325
794
+ },
795
+ {
796
+ "epoch": 12.915862068965517,
797
+ "grad_norm": 37.281959533691406,
798
+ "learning_rate": 7.881562881562883e-06,
799
+ "loss": 0.1174,
800
+ "step": 2350
801
+ },
802
+ {
803
+ "epoch": 13.0,
804
+ "eval_accuracy": 0.79585326953748,
805
+ "eval_loss": 1.623354434967041,
806
+ "eval_macro_f1": 0.7948055254552955,
807
+ "eval_runtime": 1.2964,
808
+ "eval_samples_per_second": 483.644,
809
+ "eval_steps_per_second": 30.854,
810
+ "eval_weighted_f1": 0.7949443345209599,
811
+ "step": 2366
812
+ },
813
+ {
814
+ "epoch": 13.049655172413793,
815
+ "grad_norm": 3.0382513999938965,
816
+ "learning_rate": 7.72893772893773e-06,
817
+ "loss": 0.1771,
818
+ "step": 2375
819
+ },
820
+ {
821
+ "epoch": 13.187586206896551,
822
+ "grad_norm": 0.5465341210365295,
823
+ "learning_rate": 7.576312576312577e-06,
824
+ "loss": 0.2488,
825
+ "step": 2400
826
+ },
827
+ {
828
+ "epoch": 13.32551724137931,
829
+ "grad_norm": 44.515506744384766,
830
+ "learning_rate": 7.423687423687424e-06,
831
+ "loss": 0.2077,
832
+ "step": 2425
833
+ },
834
+ {
835
+ "epoch": 13.463448275862069,
836
+ "grad_norm": 27.11043357849121,
837
+ "learning_rate": 7.271062271062271e-06,
838
+ "loss": 0.1455,
839
+ "step": 2450
840
+ },
841
+ {
842
+ "epoch": 13.601379310344827,
843
+ "grad_norm": 313.0473937988281,
844
+ "learning_rate": 7.1184371184371195e-06,
845
+ "loss": 0.1181,
846
+ "step": 2475
847
+ },
848
+ {
849
+ "epoch": 13.739310344827587,
850
+ "grad_norm": 0.022376621142029762,
851
+ "learning_rate": 6.965811965811966e-06,
852
+ "loss": 0.1992,
853
+ "step": 2500
854
+ },
855
+ {
856
+ "epoch": 13.877241379310345,
857
+ "grad_norm": 0.028782488778233528,
858
+ "learning_rate": 6.813186813186814e-06,
859
+ "loss": 0.0495,
860
+ "step": 2525
861
+ },
862
+ {
863
+ "epoch": 14.0,
864
+ "eval_accuracy": 0.8261562998405104,
865
+ "eval_loss": 1.483603596687317,
866
+ "eval_macro_f1": 0.8244676436475006,
867
+ "eval_runtime": 1.1393,
868
+ "eval_samples_per_second": 550.344,
869
+ "eval_steps_per_second": 35.11,
870
+ "eval_weighted_f1": 0.82470379163135,
871
+ "step": 2548
872
+ }
873
+ ],
874
+ "logging_steps": 25,
875
+ "max_steps": 3640,
876
+ "num_input_tokens_seen": 0,
877
+ "num_train_epochs": 20,
878
+ "save_steps": 500,
879
+ "stateful_callbacks": {
880
+ "EarlyStoppingCallback": {
881
+ "args": {
882
+ "early_stopping_patience": 3,
883
+ "early_stopping_threshold": 0.0
884
+ },
885
+ "attributes": {
886
+ "early_stopping_patience_counter": 0
887
+ }
888
+ },
889
+ "TrainerControl": {
890
+ "args": {
891
+ "should_epoch_stop": false,
892
+ "should_evaluate": false,
893
+ "should_log": false,
894
+ "should_save": true,
895
+ "should_training_stop": false
896
+ },
897
+ "attributes": {}
898
+ }
899
+ },
900
+ "total_flos": 0.0,
901
+ "train_batch_size": 8,
902
+ "trial_name": null,
903
+ "trial_params": null
904
+ }
ood-vifactcheck-xlmr-source-source-00-s43-20260901T164831Z-fbd63439/trainer/checkpoint-3094/special_tokens_map.json ADDED
@@ -0,0 +1,15 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "bos_token": "<s>",
3
+ "cls_token": "<s>",
4
+ "eos_token": "</s>",
5
+ "mask_token": {
6
+ "content": "<mask>",
7
+ "lstrip": true,
8
+ "normalized": false,
9
+ "rstrip": false,
10
+ "single_word": false
11
+ },
12
+ "pad_token": "<pad>",
13
+ "sep_token": "</s>",
14
+ "unk_token": "<unk>"
15
+ }
ood-vifactcheck-xlmr-source-source-00-s43-20260901T164831Z-fbd63439/trainer/checkpoint-3094/tokenizer_config.json ADDED
@@ -0,0 +1,55 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "added_tokens_decoder": {
3
+ "0": {
4
+ "content": "<s>",
5
+ "lstrip": false,
6
+ "normalized": false,
7
+ "rstrip": false,
8
+ "single_word": false,
9
+ "special": true
10
+ },
11
+ "1": {
12
+ "content": "<pad>",
13
+ "lstrip": false,
14
+ "normalized": false,
15
+ "rstrip": false,
16
+ "single_word": false,
17
+ "special": true
18
+ },
19
+ "2": {
20
+ "content": "</s>",
21
+ "lstrip": false,
22
+ "normalized": false,
23
+ "rstrip": false,
24
+ "single_word": false,
25
+ "special": true
26
+ },
27
+ "3": {
28
+ "content": "<unk>",
29
+ "lstrip": false,
30
+ "normalized": false,
31
+ "rstrip": false,
32
+ "single_word": false,
33
+ "special": true
34
+ },
35
+ "250001": {
36
+ "content": "<mask>",
37
+ "lstrip": true,
38
+ "normalized": false,
39
+ "rstrip": false,
40
+ "single_word": false,
41
+ "special": true
42
+ }
43
+ },
44
+ "bos_token": "<s>",
45
+ "clean_up_tokenization_spaces": false,
46
+ "cls_token": "<s>",
47
+ "eos_token": "</s>",
48
+ "extra_special_tokens": {},
49
+ "mask_token": "<mask>",
50
+ "model_max_length": 512,
51
+ "pad_token": "<pad>",
52
+ "sep_token": "</s>",
53
+ "tokenizer_class": "XLMRobertaTokenizer",
54
+ "unk_token": "<unk>"
55
+ }
ood-vifactcheck-xlmr-source-source-00-s43-20260901T164831Z-fbd63439/trainer/checkpoint-3094/trainer_state.json ADDED
@@ -0,0 +1,1091 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_global_step": 2548,
3
+ "best_metric": 0.8244676436475006,
4
+ "best_model_checkpoint": "/root/VNEncoder/results/checkpoints/ood-vifactcheck-xlmr-source-source-00-s43-20260901T164831Z-fbd63439/trainer/checkpoint-2548",
5
+ "epoch": 17.0,
6
+ "eval_steps": 500,
7
+ "global_step": 3094,
8
+ "is_hyper_param_search": false,
9
+ "is_local_process_zero": true,
10
+ "is_world_process_zero": true,
11
+ "log_history": [
12
+ {
13
+ "epoch": 0.13793103448275862,
14
+ "grad_norm": 20.084362030029297,
15
+ "learning_rate": 1.3186813186813187e-06,
16
+ "loss": 4.7288,
17
+ "step": 25
18
+ },
19
+ {
20
+ "epoch": 0.27586206896551724,
21
+ "grad_norm": 5.729361534118652,
22
+ "learning_rate": 2.6923076923076923e-06,
23
+ "loss": 4.684,
24
+ "step": 50
25
+ },
26
+ {
27
+ "epoch": 0.41379310344827586,
28
+ "grad_norm": 7.694893836975098,
29
+ "learning_rate": 4.065934065934066e-06,
30
+ "loss": 4.5741,
31
+ "step": 75
32
+ },
33
+ {
34
+ "epoch": 0.5517241379310345,
35
+ "grad_norm": 8.551424980163574,
36
+ "learning_rate": 5.43956043956044e-06,
37
+ "loss": 4.4816,
38
+ "step": 100
39
+ },
40
+ {
41
+ "epoch": 0.6896551724137931,
42
+ "grad_norm": 35.368587493896484,
43
+ "learning_rate": 6.813186813186814e-06,
44
+ "loss": 4.5308,
45
+ "step": 125
46
+ },
47
+ {
48
+ "epoch": 0.8275862068965517,
49
+ "grad_norm": 44.43610763549805,
50
+ "learning_rate": 8.186813186813188e-06,
51
+ "loss": 4.444,
52
+ "step": 150
53
+ },
54
+ {
55
+ "epoch": 0.9655172413793104,
56
+ "grad_norm": 24.932857513427734,
57
+ "learning_rate": 9.560439560439562e-06,
58
+ "loss": 4.3658,
59
+ "step": 175
60
+ },
61
+ {
62
+ "epoch": 1.0,
63
+ "eval_accuracy": 0.4258373205741627,
64
+ "eval_loss": 1.0057348012924194,
65
+ "eval_macro_f1": 0.3397112691230338,
66
+ "eval_runtime": 1.1171,
67
+ "eval_samples_per_second": 561.294,
68
+ "eval_steps_per_second": 35.808,
69
+ "eval_weighted_f1": 0.3524679654710614,
70
+ "step": 182
71
+ },
72
+ {
73
+ "epoch": 1.0993103448275863,
74
+ "grad_norm": 61.700660705566406,
75
+ "learning_rate": 1.0934065934065933e-05,
76
+ "loss": 4.1429,
77
+ "step": 200
78
+ },
79
+ {
80
+ "epoch": 1.237241379310345,
81
+ "grad_norm": 46.13859558105469,
82
+ "learning_rate": 1.230769230769231e-05,
83
+ "loss": 3.9467,
84
+ "step": 225
85
+ },
86
+ {
87
+ "epoch": 1.3751724137931034,
88
+ "grad_norm": 46.452125549316406,
89
+ "learning_rate": 1.3681318681318683e-05,
90
+ "loss": 3.6127,
91
+ "step": 250
92
+ },
93
+ {
94
+ "epoch": 1.513103448275862,
95
+ "grad_norm": 59.73109436035156,
96
+ "learning_rate": 1.5054945054945056e-05,
97
+ "loss": 3.5063,
98
+ "step": 275
99
+ },
100
+ {
101
+ "epoch": 1.6510344827586207,
102
+ "grad_norm": 145.90650939941406,
103
+ "learning_rate": 1.642857142857143e-05,
104
+ "loss": 3.3779,
105
+ "step": 300
106
+ },
107
+ {
108
+ "epoch": 1.7889655172413792,
109
+ "grad_norm": 54.86250305175781,
110
+ "learning_rate": 1.78021978021978e-05,
111
+ "loss": 3.0618,
112
+ "step": 325
113
+ },
114
+ {
115
+ "epoch": 1.926896551724138,
116
+ "grad_norm": 52.89915084838867,
117
+ "learning_rate": 1.9175824175824178e-05,
118
+ "loss": 3.1781,
119
+ "step": 350
120
+ },
121
+ {
122
+ "epoch": 2.0,
123
+ "eval_accuracy": 0.7113237639553429,
124
+ "eval_loss": 0.7212908864021301,
125
+ "eval_macro_f1": 0.7123417832526394,
126
+ "eval_runtime": 1.0768,
127
+ "eval_samples_per_second": 582.292,
128
+ "eval_steps_per_second": 37.148,
129
+ "eval_weighted_f1": 0.712662576565873,
130
+ "step": 364
131
+ },
132
+ {
133
+ "epoch": 2.0606896551724136,
134
+ "grad_norm": 84.83079528808594,
135
+ "learning_rate": 1.993894993894994e-05,
136
+ "loss": 2.6491,
137
+ "step": 375
138
+ },
139
+ {
140
+ "epoch": 2.1986206896551725,
141
+ "grad_norm": 108.39063262939453,
142
+ "learning_rate": 1.9786324786324788e-05,
143
+ "loss": 2.7958,
144
+ "step": 400
145
+ },
146
+ {
147
+ "epoch": 2.336551724137931,
148
+ "grad_norm": 96.23841094970703,
149
+ "learning_rate": 1.9633699633699634e-05,
150
+ "loss": 2.4406,
151
+ "step": 425
152
+ },
153
+ {
154
+ "epoch": 2.47448275862069,
155
+ "grad_norm": 85.79228973388672,
156
+ "learning_rate": 1.948107448107448e-05,
157
+ "loss": 2.6577,
158
+ "step": 450
159
+ },
160
+ {
161
+ "epoch": 2.6124137931034483,
162
+ "grad_norm": 42.99919509887695,
163
+ "learning_rate": 1.932844932844933e-05,
164
+ "loss": 2.5698,
165
+ "step": 475
166
+ },
167
+ {
168
+ "epoch": 2.750344827586207,
169
+ "grad_norm": 82.11553192138672,
170
+ "learning_rate": 1.9175824175824178e-05,
171
+ "loss": 2.465,
172
+ "step": 500
173
+ },
174
+ {
175
+ "epoch": 2.8882758620689657,
176
+ "grad_norm": 64.50354766845703,
177
+ "learning_rate": 1.9023199023199025e-05,
178
+ "loss": 2.6201,
179
+ "step": 525
180
+ },
181
+ {
182
+ "epoch": 3.0,
183
+ "eval_accuracy": 0.7543859649122807,
184
+ "eval_loss": 0.5967164039611816,
185
+ "eval_macro_f1": 0.7556428199642172,
186
+ "eval_runtime": 1.0678,
187
+ "eval_samples_per_second": 587.187,
188
+ "eval_steps_per_second": 37.46,
189
+ "eval_weighted_f1": 0.7552112588758818,
190
+ "step": 546
191
+ },
192
+ {
193
+ "epoch": 3.0220689655172412,
194
+ "grad_norm": 57.78512954711914,
195
+ "learning_rate": 1.887057387057387e-05,
196
+ "loss": 2.2413,
197
+ "step": 550
198
+ },
199
+ {
200
+ "epoch": 3.16,
201
+ "grad_norm": 49.69960403442383,
202
+ "learning_rate": 1.8717948717948718e-05,
203
+ "loss": 2.0975,
204
+ "step": 575
205
+ },
206
+ {
207
+ "epoch": 3.2979310344827586,
208
+ "grad_norm": 70.23579406738281,
209
+ "learning_rate": 1.8565323565323568e-05,
210
+ "loss": 1.8645,
211
+ "step": 600
212
+ },
213
+ {
214
+ "epoch": 3.435862068965517,
215
+ "grad_norm": 139.99871826171875,
216
+ "learning_rate": 1.8412698412698415e-05,
217
+ "loss": 1.9666,
218
+ "step": 625
219
+ },
220
+ {
221
+ "epoch": 3.573793103448276,
222
+ "grad_norm": 41.8527717590332,
223
+ "learning_rate": 1.826007326007326e-05,
224
+ "loss": 1.8205,
225
+ "step": 650
226
+ },
227
+ {
228
+ "epoch": 3.7117241379310344,
229
+ "grad_norm": 109.82010650634766,
230
+ "learning_rate": 1.8107448107448108e-05,
231
+ "loss": 1.6387,
232
+ "step": 675
233
+ },
234
+ {
235
+ "epoch": 3.849655172413793,
236
+ "grad_norm": 107.30518341064453,
237
+ "learning_rate": 1.7954822954822958e-05,
238
+ "loss": 2.1496,
239
+ "step": 700
240
+ },
241
+ {
242
+ "epoch": 3.9875862068965517,
243
+ "grad_norm": 83.41844177246094,
244
+ "learning_rate": 1.78021978021978e-05,
245
+ "loss": 1.9329,
246
+ "step": 725
247
+ },
248
+ {
249
+ "epoch": 4.0,
250
+ "eval_accuracy": 0.7767145135566188,
251
+ "eval_loss": 0.5591686964035034,
252
+ "eval_macro_f1": 0.7763996474814833,
253
+ "eval_runtime": 1.1489,
254
+ "eval_samples_per_second": 545.719,
255
+ "eval_steps_per_second": 34.815,
256
+ "eval_weighted_f1": 0.7766707740817129,
257
+ "step": 728
258
+ },
259
+ {
260
+ "epoch": 4.121379310344827,
261
+ "grad_norm": 38.20743942260742,
262
+ "learning_rate": 1.764957264957265e-05,
263
+ "loss": 1.501,
264
+ "step": 750
265
+ },
266
+ {
267
+ "epoch": 4.259310344827586,
268
+ "grad_norm": 101.95227813720703,
269
+ "learning_rate": 1.74969474969475e-05,
270
+ "loss": 1.5836,
271
+ "step": 775
272
+ },
273
+ {
274
+ "epoch": 4.397241379310345,
275
+ "grad_norm": 48.2017936706543,
276
+ "learning_rate": 1.7344322344322345e-05,
277
+ "loss": 1.3153,
278
+ "step": 800
279
+ },
280
+ {
281
+ "epoch": 4.535172413793103,
282
+ "grad_norm": 56.81407928466797,
283
+ "learning_rate": 1.7191697191697195e-05,
284
+ "loss": 1.5429,
285
+ "step": 825
286
+ },
287
+ {
288
+ "epoch": 4.673103448275862,
289
+ "grad_norm": 101.91703796386719,
290
+ "learning_rate": 1.703907203907204e-05,
291
+ "loss": 1.3026,
292
+ "step": 850
293
+ },
294
+ {
295
+ "epoch": 4.811034482758621,
296
+ "grad_norm": 75.37382507324219,
297
+ "learning_rate": 1.688644688644689e-05,
298
+ "loss": 1.2796,
299
+ "step": 875
300
+ },
301
+ {
302
+ "epoch": 4.94896551724138,
303
+ "grad_norm": 123.11703491210938,
304
+ "learning_rate": 1.6733821733821735e-05,
305
+ "loss": 1.4378,
306
+ "step": 900
307
+ },
308
+ {
309
+ "epoch": 5.0,
310
+ "eval_accuracy": 0.7862838915470495,
311
+ "eval_loss": 0.7190022468566895,
312
+ "eval_macro_f1": 0.7867678171448537,
313
+ "eval_runtime": 1.1303,
314
+ "eval_samples_per_second": 554.703,
315
+ "eval_steps_per_second": 35.388,
316
+ "eval_weighted_f1": 0.7868702555186253,
317
+ "step": 910
318
+ },
319
+ {
320
+ "epoch": 5.082758620689655,
321
+ "grad_norm": 111.80149841308594,
322
+ "learning_rate": 1.6581196581196585e-05,
323
+ "loss": 1.0587,
324
+ "step": 925
325
+ },
326
+ {
327
+ "epoch": 5.220689655172414,
328
+ "grad_norm": 101.65257263183594,
329
+ "learning_rate": 1.642857142857143e-05,
330
+ "loss": 1.0306,
331
+ "step": 950
332
+ },
333
+ {
334
+ "epoch": 5.358620689655172,
335
+ "grad_norm": 58.97084045410156,
336
+ "learning_rate": 1.627594627594628e-05,
337
+ "loss": 1.1596,
338
+ "step": 975
339
+ },
340
+ {
341
+ "epoch": 5.496551724137931,
342
+ "grad_norm": 63.96665573120117,
343
+ "learning_rate": 1.6123321123321125e-05,
344
+ "loss": 1.2873,
345
+ "step": 1000
346
+ },
347
+ {
348
+ "epoch": 5.63448275862069,
349
+ "grad_norm": 104.3017349243164,
350
+ "learning_rate": 1.5970695970695972e-05,
351
+ "loss": 1.0208,
352
+ "step": 1025
353
+ },
354
+ {
355
+ "epoch": 5.772413793103448,
356
+ "grad_norm": 108.5999526977539,
357
+ "learning_rate": 1.581807081807082e-05,
358
+ "loss": 1.2335,
359
+ "step": 1050
360
+ },
361
+ {
362
+ "epoch": 5.910344827586207,
363
+ "grad_norm": 94.29696655273438,
364
+ "learning_rate": 1.5665445665445665e-05,
365
+ "loss": 1.2207,
366
+ "step": 1075
367
+ },
368
+ {
369
+ "epoch": 6.0,
370
+ "eval_accuracy": 0.7878787878787878,
371
+ "eval_loss": 0.6836919784545898,
372
+ "eval_macro_f1": 0.7879398797194143,
373
+ "eval_runtime": 1.1922,
374
+ "eval_samples_per_second": 525.905,
375
+ "eval_steps_per_second": 33.551,
376
+ "eval_weighted_f1": 0.7879111724300937,
377
+ "step": 1092
378
+ },
379
+ {
380
+ "epoch": 6.0441379310344825,
381
+ "grad_norm": 103.21680450439453,
382
+ "learning_rate": 1.5512820512820516e-05,
383
+ "loss": 0.9723,
384
+ "step": 1100
385
+ },
386
+ {
387
+ "epoch": 6.182068965517241,
388
+ "grad_norm": 104.1303482055664,
389
+ "learning_rate": 1.5360195360195362e-05,
390
+ "loss": 0.7962,
391
+ "step": 1125
392
+ },
393
+ {
394
+ "epoch": 6.32,
395
+ "grad_norm": 56.65522003173828,
396
+ "learning_rate": 1.5207570207570207e-05,
397
+ "loss": 0.7656,
398
+ "step": 1150
399
+ },
400
+ {
401
+ "epoch": 6.457931034482758,
402
+ "grad_norm": 136.6097869873047,
403
+ "learning_rate": 1.5054945054945056e-05,
404
+ "loss": 0.6219,
405
+ "step": 1175
406
+ },
407
+ {
408
+ "epoch": 6.595862068965517,
409
+ "grad_norm": 116.55268096923828,
410
+ "learning_rate": 1.4902319902319904e-05,
411
+ "loss": 0.7777,
412
+ "step": 1200
413
+ },
414
+ {
415
+ "epoch": 6.733793103448276,
416
+ "grad_norm": 61.935054779052734,
417
+ "learning_rate": 1.4749694749694752e-05,
418
+ "loss": 0.8592,
419
+ "step": 1225
420
+ },
421
+ {
422
+ "epoch": 6.871724137931034,
423
+ "grad_norm": 34.48295593261719,
424
+ "learning_rate": 1.4597069597069597e-05,
425
+ "loss": 0.9455,
426
+ "step": 1250
427
+ },
428
+ {
429
+ "epoch": 7.0,
430
+ "eval_accuracy": 0.7830940988835726,
431
+ "eval_loss": 0.7971011996269226,
432
+ "eval_macro_f1": 0.7841797417521805,
433
+ "eval_runtime": 1.1578,
434
+ "eval_samples_per_second": 541.551,
435
+ "eval_steps_per_second": 34.549,
436
+ "eval_weighted_f1": 0.7845438631046493,
437
+ "step": 1274
438
+ },
439
+ {
440
+ "epoch": 7.0055172413793105,
441
+ "grad_norm": 178.21438598632812,
442
+ "learning_rate": 1.4444444444444446e-05,
443
+ "loss": 0.7349,
444
+ "step": 1275
445
+ },
446
+ {
447
+ "epoch": 7.1434482758620685,
448
+ "grad_norm": 101.17195892333984,
449
+ "learning_rate": 1.4291819291819293e-05,
450
+ "loss": 0.6712,
451
+ "step": 1300
452
+ },
453
+ {
454
+ "epoch": 7.281379310344827,
455
+ "grad_norm": 60.77378845214844,
456
+ "learning_rate": 1.4139194139194141e-05,
457
+ "loss": 0.6681,
458
+ "step": 1325
459
+ },
460
+ {
461
+ "epoch": 7.419310344827586,
462
+ "grad_norm": 227.9224090576172,
463
+ "learning_rate": 1.3986568986568988e-05,
464
+ "loss": 0.5815,
465
+ "step": 1350
466
+ },
467
+ {
468
+ "epoch": 7.557241379310344,
469
+ "grad_norm": 93.1969223022461,
470
+ "learning_rate": 1.3833943833943834e-05,
471
+ "loss": 0.7168,
472
+ "step": 1375
473
+ },
474
+ {
475
+ "epoch": 7.695172413793103,
476
+ "grad_norm": 53.236114501953125,
477
+ "learning_rate": 1.3681318681318683e-05,
478
+ "loss": 0.8488,
479
+ "step": 1400
480
+ },
481
+ {
482
+ "epoch": 7.833103448275862,
483
+ "grad_norm": 84.6240005493164,
484
+ "learning_rate": 1.3528693528693531e-05,
485
+ "loss": 0.5962,
486
+ "step": 1425
487
+ },
488
+ {
489
+ "epoch": 7.971034482758621,
490
+ "grad_norm": 70.12305450439453,
491
+ "learning_rate": 1.3376068376068376e-05,
492
+ "loss": 0.6278,
493
+ "step": 1450
494
+ },
495
+ {
496
+ "epoch": 8.0,
497
+ "eval_accuracy": 0.7910685805422647,
498
+ "eval_loss": 0.9096579551696777,
499
+ "eval_macro_f1": 0.7890004378104845,
500
+ "eval_runtime": 1.0983,
501
+ "eval_samples_per_second": 570.88,
502
+ "eval_steps_per_second": 36.42,
503
+ "eval_weighted_f1": 0.7894828129259822,
504
+ "step": 1456
505
+ },
506
+ {
507
+ "epoch": 8.104827586206897,
508
+ "grad_norm": 181.92086791992188,
509
+ "learning_rate": 1.3223443223443224e-05,
510
+ "loss": 0.5441,
511
+ "step": 1475
512
+ },
513
+ {
514
+ "epoch": 8.242758620689655,
515
+ "grad_norm": 91.28218841552734,
516
+ "learning_rate": 1.3070818070818073e-05,
517
+ "loss": 0.6527,
518
+ "step": 1500
519
+ },
520
+ {
521
+ "epoch": 8.380689655172414,
522
+ "grad_norm": 211.7768096923828,
523
+ "learning_rate": 1.291819291819292e-05,
524
+ "loss": 0.4916,
525
+ "step": 1525
526
+ },
527
+ {
528
+ "epoch": 8.518620689655172,
529
+ "grad_norm": 32.13935089111328,
530
+ "learning_rate": 1.2765567765567766e-05,
531
+ "loss": 0.4391,
532
+ "step": 1550
533
+ },
534
+ {
535
+ "epoch": 8.65655172413793,
536
+ "grad_norm": 295.61663818359375,
537
+ "learning_rate": 1.2612942612942613e-05,
538
+ "loss": 0.5198,
539
+ "step": 1575
540
+ },
541
+ {
542
+ "epoch": 8.79448275862069,
543
+ "grad_norm": 32.63080596923828,
544
+ "learning_rate": 1.2460317460317461e-05,
545
+ "loss": 0.4442,
546
+ "step": 1600
547
+ },
548
+ {
549
+ "epoch": 8.932413793103448,
550
+ "grad_norm": 83.26895904541016,
551
+ "learning_rate": 1.230769230769231e-05,
552
+ "loss": 0.6965,
553
+ "step": 1625
554
+ },
555
+ {
556
+ "epoch": 9.0,
557
+ "eval_accuracy": 0.7974481658692185,
558
+ "eval_loss": 1.1046565771102905,
559
+ "eval_macro_f1": 0.7953705175062238,
560
+ "eval_runtime": 1.1067,
561
+ "eval_samples_per_second": 566.552,
562
+ "eval_steps_per_second": 36.144,
563
+ "eval_weighted_f1": 0.7957420601762267,
564
+ "step": 1638
565
+ },
566
+ {
567
+ "epoch": 9.066206896551725,
568
+ "grad_norm": 0.9721019268035889,
569
+ "learning_rate": 1.2155067155067155e-05,
570
+ "loss": 0.3916,
571
+ "step": 1650
572
+ },
573
+ {
574
+ "epoch": 9.204137931034483,
575
+ "grad_norm": 246.64122009277344,
576
+ "learning_rate": 1.2002442002442003e-05,
577
+ "loss": 0.5128,
578
+ "step": 1675
579
+ },
580
+ {
581
+ "epoch": 9.34206896551724,
582
+ "grad_norm": 66.1498794555664,
583
+ "learning_rate": 1.1849816849816852e-05,
584
+ "loss": 0.2748,
585
+ "step": 1700
586
+ },
587
+ {
588
+ "epoch": 9.48,
589
+ "grad_norm": 18.061092376708984,
590
+ "learning_rate": 1.1697191697191698e-05,
591
+ "loss": 0.421,
592
+ "step": 1725
593
+ },
594
+ {
595
+ "epoch": 9.617931034482758,
596
+ "grad_norm": 7.748656272888184,
597
+ "learning_rate": 1.1544566544566545e-05,
598
+ "loss": 0.3691,
599
+ "step": 1750
600
+ },
601
+ {
602
+ "epoch": 9.755862068965516,
603
+ "grad_norm": 139.55853271484375,
604
+ "learning_rate": 1.1391941391941392e-05,
605
+ "loss": 0.3909,
606
+ "step": 1775
607
+ },
608
+ {
609
+ "epoch": 9.893793103448276,
610
+ "grad_norm": 165.58914184570312,
611
+ "learning_rate": 1.123931623931624e-05,
612
+ "loss": 0.3689,
613
+ "step": 1800
614
+ },
615
+ {
616
+ "epoch": 10.0,
617
+ "eval_accuracy": 0.7926634768740032,
618
+ "eval_loss": 1.3370226621627808,
619
+ "eval_macro_f1": 0.7928866612894071,
620
+ "eval_runtime": 1.0674,
621
+ "eval_samples_per_second": 587.409,
622
+ "eval_steps_per_second": 37.474,
623
+ "eval_weighted_f1": 0.7930633611359907,
624
+ "step": 1820
625
+ },
626
+ {
627
+ "epoch": 10.027586206896551,
628
+ "grad_norm": 2.90893292427063,
629
+ "learning_rate": 1.1086691086691088e-05,
630
+ "loss": 0.2536,
631
+ "step": 1825
632
+ },
633
+ {
634
+ "epoch": 10.16551724137931,
635
+ "grad_norm": 69.7308120727539,
636
+ "learning_rate": 1.0934065934065933e-05,
637
+ "loss": 0.4142,
638
+ "step": 1850
639
+ },
640
+ {
641
+ "epoch": 10.303448275862069,
642
+ "grad_norm": 0.6135218739509583,
643
+ "learning_rate": 1.0781440781440782e-05,
644
+ "loss": 0.2497,
645
+ "step": 1875
646
+ },
647
+ {
648
+ "epoch": 10.441379310344828,
649
+ "grad_norm": 1.3770861625671387,
650
+ "learning_rate": 1.062881562881563e-05,
651
+ "loss": 0.2115,
652
+ "step": 1900
653
+ },
654
+ {
655
+ "epoch": 10.579310344827586,
656
+ "grad_norm": 284.72198486328125,
657
+ "learning_rate": 1.0476190476190477e-05,
658
+ "loss": 0.596,
659
+ "step": 1925
660
+ },
661
+ {
662
+ "epoch": 10.717241379310344,
663
+ "grad_norm": 2.9492316246032715,
664
+ "learning_rate": 1.0323565323565325e-05,
665
+ "loss": 0.4342,
666
+ "step": 1950
667
+ },
668
+ {
669
+ "epoch": 10.855172413793104,
670
+ "grad_norm": 68.05549621582031,
671
+ "learning_rate": 1.0170940170940172e-05,
672
+ "loss": 0.2668,
673
+ "step": 1975
674
+ },
675
+ {
676
+ "epoch": 10.993103448275862,
677
+ "grad_norm": 106.35601043701172,
678
+ "learning_rate": 1.0018315018315019e-05,
679
+ "loss": 0.4008,
680
+ "step": 2000
681
+ },
682
+ {
683
+ "epoch": 11.0,
684
+ "eval_accuracy": 0.773524720893142,
685
+ "eval_loss": 1.292651653289795,
686
+ "eval_macro_f1": 0.7737703583664469,
687
+ "eval_runtime": 1.2845,
688
+ "eval_samples_per_second": 488.141,
689
+ "eval_steps_per_second": 31.141,
690
+ "eval_weighted_f1": 0.7740535430140322,
691
+ "step": 2002
692
+ },
693
+ {
694
+ "epoch": 11.126896551724139,
695
+ "grad_norm": 12.272939682006836,
696
+ "learning_rate": 9.865689865689867e-06,
697
+ "loss": 0.1852,
698
+ "step": 2025
699
+ },
700
+ {
701
+ "epoch": 11.264827586206897,
702
+ "grad_norm": 0.17907734215259552,
703
+ "learning_rate": 9.713064713064714e-06,
704
+ "loss": 0.2774,
705
+ "step": 2050
706
+ },
707
+ {
708
+ "epoch": 11.402758620689655,
709
+ "grad_norm": 143.93467712402344,
710
+ "learning_rate": 9.560439560439562e-06,
711
+ "loss": 0.2405,
712
+ "step": 2075
713
+ },
714
+ {
715
+ "epoch": 11.540689655172415,
716
+ "grad_norm": 139.8834991455078,
717
+ "learning_rate": 9.407814407814409e-06,
718
+ "loss": 0.2925,
719
+ "step": 2100
720
+ },
721
+ {
722
+ "epoch": 11.678620689655173,
723
+ "grad_norm": 12.48167610168457,
724
+ "learning_rate": 9.255189255189256e-06,
725
+ "loss": 0.3291,
726
+ "step": 2125
727
+ },
728
+ {
729
+ "epoch": 11.81655172413793,
730
+ "grad_norm": 213.54396057128906,
731
+ "learning_rate": 9.102564102564104e-06,
732
+ "loss": 0.3811,
733
+ "step": 2150
734
+ },
735
+ {
736
+ "epoch": 11.95448275862069,
737
+ "grad_norm": 3.9226012229919434,
738
+ "learning_rate": 8.94993894993895e-06,
739
+ "loss": 0.272,
740
+ "step": 2175
741
+ },
742
+ {
743
+ "epoch": 12.0,
744
+ "eval_accuracy": 0.8070175438596491,
745
+ "eval_loss": 1.3421677350997925,
746
+ "eval_macro_f1": 0.8049096987990539,
747
+ "eval_runtime": 1.1904,
748
+ "eval_samples_per_second": 526.708,
749
+ "eval_steps_per_second": 33.602,
750
+ "eval_weighted_f1": 0.8049808148241474,
751
+ "step": 2184
752
+ },
753
+ {
754
+ "epoch": 12.088275862068965,
755
+ "grad_norm": 164.14903259277344,
756
+ "learning_rate": 8.797313797313797e-06,
757
+ "loss": 0.2303,
758
+ "step": 2200
759
+ },
760
+ {
761
+ "epoch": 12.226206896551725,
762
+ "grad_norm": 1.2080793380737305,
763
+ "learning_rate": 8.644688644688646e-06,
764
+ "loss": 0.2818,
765
+ "step": 2225
766
+ },
767
+ {
768
+ "epoch": 12.364137931034483,
769
+ "grad_norm": 0.2107601761817932,
770
+ "learning_rate": 8.492063492063492e-06,
771
+ "loss": 0.166,
772
+ "step": 2250
773
+ },
774
+ {
775
+ "epoch": 12.50206896551724,
776
+ "grad_norm": 597.9608154296875,
777
+ "learning_rate": 8.33943833943834e-06,
778
+ "loss": 0.2335,
779
+ "step": 2275
780
+ },
781
+ {
782
+ "epoch": 12.64,
783
+ "grad_norm": 0.06262794137001038,
784
+ "learning_rate": 8.186813186813188e-06,
785
+ "loss": 0.3141,
786
+ "step": 2300
787
+ },
788
+ {
789
+ "epoch": 12.777931034482759,
790
+ "grad_norm": 183.90464782714844,
791
+ "learning_rate": 8.034188034188036e-06,
792
+ "loss": 0.1393,
793
+ "step": 2325
794
+ },
795
+ {
796
+ "epoch": 12.915862068965517,
797
+ "grad_norm": 37.281959533691406,
798
+ "learning_rate": 7.881562881562883e-06,
799
+ "loss": 0.1174,
800
+ "step": 2350
801
+ },
802
+ {
803
+ "epoch": 13.0,
804
+ "eval_accuracy": 0.79585326953748,
805
+ "eval_loss": 1.623354434967041,
806
+ "eval_macro_f1": 0.7948055254552955,
807
+ "eval_runtime": 1.2964,
808
+ "eval_samples_per_second": 483.644,
809
+ "eval_steps_per_second": 30.854,
810
+ "eval_weighted_f1": 0.7949443345209599,
811
+ "step": 2366
812
+ },
813
+ {
814
+ "epoch": 13.049655172413793,
815
+ "grad_norm": 3.0382513999938965,
816
+ "learning_rate": 7.72893772893773e-06,
817
+ "loss": 0.1771,
818
+ "step": 2375
819
+ },
820
+ {
821
+ "epoch": 13.187586206896551,
822
+ "grad_norm": 0.5465341210365295,
823
+ "learning_rate": 7.576312576312577e-06,
824
+ "loss": 0.2488,
825
+ "step": 2400
826
+ },
827
+ {
828
+ "epoch": 13.32551724137931,
829
+ "grad_norm": 44.515506744384766,
830
+ "learning_rate": 7.423687423687424e-06,
831
+ "loss": 0.2077,
832
+ "step": 2425
833
+ },
834
+ {
835
+ "epoch": 13.463448275862069,
836
+ "grad_norm": 27.11043357849121,
837
+ "learning_rate": 7.271062271062271e-06,
838
+ "loss": 0.1455,
839
+ "step": 2450
840
+ },
841
+ {
842
+ "epoch": 13.601379310344827,
843
+ "grad_norm": 313.0473937988281,
844
+ "learning_rate": 7.1184371184371195e-06,
845
+ "loss": 0.1181,
846
+ "step": 2475
847
+ },
848
+ {
849
+ "epoch": 13.739310344827587,
850
+ "grad_norm": 0.022376621142029762,
851
+ "learning_rate": 6.965811965811966e-06,
852
+ "loss": 0.1992,
853
+ "step": 2500
854
+ },
855
+ {
856
+ "epoch": 13.877241379310345,
857
+ "grad_norm": 0.028782488778233528,
858
+ "learning_rate": 6.813186813186814e-06,
859
+ "loss": 0.0495,
860
+ "step": 2525
861
+ },
862
+ {
863
+ "epoch": 14.0,
864
+ "eval_accuracy": 0.8261562998405104,
865
+ "eval_loss": 1.483603596687317,
866
+ "eval_macro_f1": 0.8244676436475006,
867
+ "eval_runtime": 1.1393,
868
+ "eval_samples_per_second": 550.344,
869
+ "eval_steps_per_second": 35.11,
870
+ "eval_weighted_f1": 0.82470379163135,
871
+ "step": 2548
872
+ },
873
+ {
874
+ "epoch": 14.011034482758621,
875
+ "grad_norm": 167.15940856933594,
876
+ "learning_rate": 6.66056166056166e-06,
877
+ "loss": 0.1243,
878
+ "step": 2550
879
+ },
880
+ {
881
+ "epoch": 14.148965517241379,
882
+ "grad_norm": 0.06580586731433868,
883
+ "learning_rate": 6.507936507936509e-06,
884
+ "loss": 0.1277,
885
+ "step": 2575
886
+ },
887
+ {
888
+ "epoch": 14.286896551724137,
889
+ "grad_norm": 0.0872478038072586,
890
+ "learning_rate": 6.3553113553113555e-06,
891
+ "loss": 0.0469,
892
+ "step": 2600
893
+ },
894
+ {
895
+ "epoch": 14.424827586206897,
896
+ "grad_norm": 0.4317820966243744,
897
+ "learning_rate": 6.202686202686203e-06,
898
+ "loss": 0.274,
899
+ "step": 2625
900
+ },
901
+ {
902
+ "epoch": 14.562758620689655,
903
+ "grad_norm": 0.05667780339717865,
904
+ "learning_rate": 6.05006105006105e-06,
905
+ "loss": 0.2174,
906
+ "step": 2650
907
+ },
908
+ {
909
+ "epoch": 14.700689655172415,
910
+ "grad_norm": 30.911951065063477,
911
+ "learning_rate": 5.897435897435898e-06,
912
+ "loss": 0.1159,
913
+ "step": 2675
914
+ },
915
+ {
916
+ "epoch": 14.838620689655173,
917
+ "grad_norm": 0.06982552260160446,
918
+ "learning_rate": 5.744810744810745e-06,
919
+ "loss": 0.0219,
920
+ "step": 2700
921
+ },
922
+ {
923
+ "epoch": 14.97655172413793,
924
+ "grad_norm": 0.06446123123168945,
925
+ "learning_rate": 5.592185592185592e-06,
926
+ "loss": 0.2657,
927
+ "step": 2725
928
+ },
929
+ {
930
+ "epoch": 15.0,
931
+ "eval_accuracy": 0.8070175438596491,
932
+ "eval_loss": 1.6130914688110352,
933
+ "eval_macro_f1": 0.8066404339513583,
934
+ "eval_runtime": 1.21,
935
+ "eval_samples_per_second": 518.198,
936
+ "eval_steps_per_second": 33.059,
937
+ "eval_weighted_f1": 0.8065686880812932,
938
+ "step": 2730
939
+ },
940
+ {
941
+ "epoch": 15.110344827586207,
942
+ "grad_norm": 0.00978496577590704,
943
+ "learning_rate": 5.43956043956044e-06,
944
+ "loss": 0.0828,
945
+ "step": 2750
946
+ },
947
+ {
948
+ "epoch": 15.248275862068965,
949
+ "grad_norm": 0.05871890112757683,
950
+ "learning_rate": 5.2869352869352875e-06,
951
+ "loss": 0.1256,
952
+ "step": 2775
953
+ },
954
+ {
955
+ "epoch": 15.386206896551725,
956
+ "grad_norm": 0.015405900776386261,
957
+ "learning_rate": 5.134310134310134e-06,
958
+ "loss": 0.063,
959
+ "step": 2800
960
+ },
961
+ {
962
+ "epoch": 15.524137931034483,
963
+ "grad_norm": 417.1138916015625,
964
+ "learning_rate": 4.9816849816849826e-06,
965
+ "loss": 0.1314,
966
+ "step": 2825
967
+ },
968
+ {
969
+ "epoch": 15.662068965517241,
970
+ "grad_norm": 2.9514923095703125,
971
+ "learning_rate": 4.829059829059829e-06,
972
+ "loss": 0.0709,
973
+ "step": 2850
974
+ },
975
+ {
976
+ "epoch": 15.8,
977
+ "grad_norm": 106.02448272705078,
978
+ "learning_rate": 4.676434676434677e-06,
979
+ "loss": 0.1232,
980
+ "step": 2875
981
+ },
982
+ {
983
+ "epoch": 15.937931034482759,
984
+ "grad_norm": 102.89264678955078,
985
+ "learning_rate": 4.523809523809524e-06,
986
+ "loss": 0.1052,
987
+ "step": 2900
988
+ },
989
+ {
990
+ "epoch": 16.0,
991
+ "eval_accuracy": 0.8022328548644339,
992
+ "eval_loss": 1.6951606273651123,
993
+ "eval_macro_f1": 0.8008641234238222,
994
+ "eval_runtime": 1.2038,
995
+ "eval_samples_per_second": 520.843,
996
+ "eval_steps_per_second": 33.228,
997
+ "eval_weighted_f1": 0.8008428378737971,
998
+ "step": 2912
999
+ },
1000
+ {
1001
+ "epoch": 16.071724137931035,
1002
+ "grad_norm": 0.15309341251850128,
1003
+ "learning_rate": 4.371184371184372e-06,
1004
+ "loss": 0.1384,
1005
+ "step": 2925
1006
+ },
1007
+ {
1008
+ "epoch": 16.209655172413793,
1009
+ "grad_norm": 1.8501631021499634,
1010
+ "learning_rate": 4.218559218559219e-06,
1011
+ "loss": 0.2515,
1012
+ "step": 2950
1013
+ },
1014
+ {
1015
+ "epoch": 16.34758620689655,
1016
+ "grad_norm": 0.022675300016999245,
1017
+ "learning_rate": 4.065934065934066e-06,
1018
+ "loss": 0.0442,
1019
+ "step": 2975
1020
+ },
1021
+ {
1022
+ "epoch": 16.48551724137931,
1023
+ "grad_norm": 0.04241953790187836,
1024
+ "learning_rate": 3.913308913308914e-06,
1025
+ "loss": 0.0218,
1026
+ "step": 3000
1027
+ },
1028
+ {
1029
+ "epoch": 16.623448275862067,
1030
+ "grad_norm": 0.03193862363696098,
1031
+ "learning_rate": 3.760683760683761e-06,
1032
+ "loss": 0.0313,
1033
+ "step": 3025
1034
+ },
1035
+ {
1036
+ "epoch": 16.76137931034483,
1037
+ "grad_norm": 1.043158769607544,
1038
+ "learning_rate": 3.6080586080586083e-06,
1039
+ "loss": 0.101,
1040
+ "step": 3050
1041
+ },
1042
+ {
1043
+ "epoch": 16.899310344827587,
1044
+ "grad_norm": 0.015490840189158916,
1045
+ "learning_rate": 3.4554334554334555e-06,
1046
+ "loss": 0.1246,
1047
+ "step": 3075
1048
+ },
1049
+ {
1050
+ "epoch": 17.0,
1051
+ "eval_accuracy": 0.8038277511961722,
1052
+ "eval_loss": 1.728859305381775,
1053
+ "eval_macro_f1": 0.8024189355890387,
1054
+ "eval_runtime": 1.1977,
1055
+ "eval_samples_per_second": 523.508,
1056
+ "eval_steps_per_second": 33.398,
1057
+ "eval_weighted_f1": 0.802411980314856,
1058
+ "step": 3094
1059
+ }
1060
+ ],
1061
+ "logging_steps": 25,
1062
+ "max_steps": 3640,
1063
+ "num_input_tokens_seen": 0,
1064
+ "num_train_epochs": 20,
1065
+ "save_steps": 500,
1066
+ "stateful_callbacks": {
1067
+ "EarlyStoppingCallback": {
1068
+ "args": {
1069
+ "early_stopping_patience": 3,
1070
+ "early_stopping_threshold": 0.0
1071
+ },
1072
+ "attributes": {
1073
+ "early_stopping_patience_counter": 3
1074
+ }
1075
+ },
1076
+ "TrainerControl": {
1077
+ "args": {
1078
+ "should_epoch_stop": false,
1079
+ "should_evaluate": false,
1080
+ "should_log": false,
1081
+ "should_save": true,
1082
+ "should_training_stop": true
1083
+ },
1084
+ "attributes": {}
1085
+ }
1086
+ },
1087
+ "total_flos": 0.0,
1088
+ "train_batch_size": 8,
1089
+ "trial_name": null,
1090
+ "trial_params": null
1091
+ }
ood-vifactcheck-xlmr-source-source-01-s42-20260901T152349Z-27bda108/trainer/checkpoint-1197/special_tokens_map.json ADDED
@@ -0,0 +1,15 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "bos_token": "<s>",
3
+ "cls_token": "<s>",
4
+ "eos_token": "</s>",
5
+ "mask_token": {
6
+ "content": "<mask>",
7
+ "lstrip": true,
8
+ "normalized": false,
9
+ "rstrip": false,
10
+ "single_word": false
11
+ },
12
+ "pad_token": "<pad>",
13
+ "sep_token": "</s>",
14
+ "unk_token": "<unk>"
15
+ }
ood-vifactcheck-xlmr-source-source-01-s42-20260901T152349Z-27bda108/trainer/checkpoint-1197/tokenizer_config.json ADDED
@@ -0,0 +1,55 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "added_tokens_decoder": {
3
+ "0": {
4
+ "content": "<s>",
5
+ "lstrip": false,
6
+ "normalized": false,
7
+ "rstrip": false,
8
+ "single_word": false,
9
+ "special": true
10
+ },
11
+ "1": {
12
+ "content": "<pad>",
13
+ "lstrip": false,
14
+ "normalized": false,
15
+ "rstrip": false,
16
+ "single_word": false,
17
+ "special": true
18
+ },
19
+ "2": {
20
+ "content": "</s>",
21
+ "lstrip": false,
22
+ "normalized": false,
23
+ "rstrip": false,
24
+ "single_word": false,
25
+ "special": true
26
+ },
27
+ "3": {
28
+ "content": "<unk>",
29
+ "lstrip": false,
30
+ "normalized": false,
31
+ "rstrip": false,
32
+ "single_word": false,
33
+ "special": true
34
+ },
35
+ "250001": {
36
+ "content": "<mask>",
37
+ "lstrip": true,
38
+ "normalized": false,
39
+ "rstrip": false,
40
+ "single_word": false,
41
+ "special": true
42
+ }
43
+ },
44
+ "bos_token": "<s>",
45
+ "clean_up_tokenization_spaces": false,
46
+ "cls_token": "<s>",
47
+ "eos_token": "</s>",
48
+ "extra_special_tokens": {},
49
+ "mask_token": "<mask>",
50
+ "model_max_length": 512,
51
+ "pad_token": "<pad>",
52
+ "sep_token": "</s>",
53
+ "tokenizer_class": "XLMRobertaTokenizer",
54
+ "unk_token": "<unk>"
55
+ }
ood-vifactcheck-xlmr-source-source-01-s42-20260901T152349Z-27bda108/trainer/checkpoint-1197/trainer_state.json ADDED
@@ -0,0 +1,449 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_global_step": 1197,
3
+ "best_metric": 0.8295240383812666,
4
+ "best_model_checkpoint": "/root/VNEncoder/results/checkpoints/ood-vifactcheck-xlmr-source-source-01-s42-20260901T152349Z-27bda108/trainer/checkpoint-1197",
5
+ "epoch": 7.0,
6
+ "eval_steps": 500,
7
+ "global_step": 1197,
8
+ "is_hyper_param_search": false,
9
+ "is_local_process_zero": true,
10
+ "is_world_process_zero": true,
11
+ "log_history": [
12
+ {
13
+ "epoch": 0.14684287812041116,
14
+ "grad_norm": 8.276679992675781,
15
+ "learning_rate": 1.4035087719298246e-06,
16
+ "loss": 4.7457,
17
+ "step": 25
18
+ },
19
+ {
20
+ "epoch": 0.2936857562408223,
21
+ "grad_norm": 32.80095672607422,
22
+ "learning_rate": 2.865497076023392e-06,
23
+ "loss": 4.8125,
24
+ "step": 50
25
+ },
26
+ {
27
+ "epoch": 0.44052863436123346,
28
+ "grad_norm": 23.083709716796875,
29
+ "learning_rate": 4.3274853801169596e-06,
30
+ "loss": 4.6282,
31
+ "step": 75
32
+ },
33
+ {
34
+ "epoch": 0.5873715124816447,
35
+ "grad_norm": 25.51261329650879,
36
+ "learning_rate": 5.789473684210527e-06,
37
+ "loss": 4.6667,
38
+ "step": 100
39
+ },
40
+ {
41
+ "epoch": 0.7342143906020558,
42
+ "grad_norm": 41.84646224975586,
43
+ "learning_rate": 7.251461988304094e-06,
44
+ "loss": 4.593,
45
+ "step": 125
46
+ },
47
+ {
48
+ "epoch": 0.8810572687224669,
49
+ "grad_norm": 29.750471115112305,
50
+ "learning_rate": 8.713450292397661e-06,
51
+ "loss": 4.447,
52
+ "step": 150
53
+ },
54
+ {
55
+ "epoch": 1.0,
56
+ "eval_accuracy": 0.3736654804270463,
57
+ "eval_loss": 1.0819200277328491,
58
+ "eval_macro_f1": 0.29437285769412447,
59
+ "eval_runtime": 1.0765,
60
+ "eval_samples_per_second": 522.066,
61
+ "eval_steps_per_second": 33.442,
62
+ "eval_weighted_f1": 0.2922100579819657,
63
+ "step": 171
64
+ },
65
+ {
66
+ "epoch": 1.0234948604992657,
67
+ "grad_norm": 33.48249816894531,
68
+ "learning_rate": 1.017543859649123e-05,
69
+ "loss": 4.2602,
70
+ "step": 175
71
+ },
72
+ {
73
+ "epoch": 1.170337738619677,
74
+ "grad_norm": 30.328693389892578,
75
+ "learning_rate": 1.1637426900584797e-05,
76
+ "loss": 4.2628,
77
+ "step": 200
78
+ },
79
+ {
80
+ "epoch": 1.3171806167400881,
81
+ "grad_norm": 53.562705993652344,
82
+ "learning_rate": 1.3099415204678362e-05,
83
+ "loss": 3.9631,
84
+ "step": 225
85
+ },
86
+ {
87
+ "epoch": 1.4640234948604993,
88
+ "grad_norm": 64.7986831665039,
89
+ "learning_rate": 1.4561403508771931e-05,
90
+ "loss": 4.2405,
91
+ "step": 250
92
+ },
93
+ {
94
+ "epoch": 1.6108663729809103,
95
+ "grad_norm": 63.56324005126953,
96
+ "learning_rate": 1.60233918128655e-05,
97
+ "loss": 3.8987,
98
+ "step": 275
99
+ },
100
+ {
101
+ "epoch": 1.7577092511013217,
102
+ "grad_norm": 60.7152214050293,
103
+ "learning_rate": 1.7485380116959064e-05,
104
+ "loss": 3.6177,
105
+ "step": 300
106
+ },
107
+ {
108
+ "epoch": 1.9045521292217327,
109
+ "grad_norm": 60.10673141479492,
110
+ "learning_rate": 1.894736842105263e-05,
111
+ "loss": 3.4457,
112
+ "step": 325
113
+ },
114
+ {
115
+ "epoch": 2.0,
116
+ "eval_accuracy": 0.701067615658363,
117
+ "eval_loss": 0.6930170655250549,
118
+ "eval_macro_f1": 0.6986001077282614,
119
+ "eval_runtime": 1.1239,
120
+ "eval_samples_per_second": 500.04,
121
+ "eval_steps_per_second": 32.031,
122
+ "eval_weighted_f1": 0.6982019631794543,
123
+ "step": 342
124
+ },
125
+ {
126
+ "epoch": 2.0469897209985315,
127
+ "grad_norm": 32.58366012573242,
128
+ "learning_rate": 1.99545159194282e-05,
129
+ "loss": 3.0666,
130
+ "step": 350
131
+ },
132
+ {
133
+ "epoch": 2.193832599118943,
134
+ "grad_norm": 94.89712524414062,
135
+ "learning_rate": 1.9792072774528916e-05,
136
+ "loss": 3.0562,
137
+ "step": 375
138
+ },
139
+ {
140
+ "epoch": 2.340675477239354,
141
+ "grad_norm": 50.320068359375,
142
+ "learning_rate": 1.962962962962963e-05,
143
+ "loss": 2.7734,
144
+ "step": 400
145
+ },
146
+ {
147
+ "epoch": 2.4875183553597653,
148
+ "grad_norm": 93.8223648071289,
149
+ "learning_rate": 1.9467186484730346e-05,
150
+ "loss": 2.6324,
151
+ "step": 425
152
+ },
153
+ {
154
+ "epoch": 2.6343612334801763,
155
+ "grad_norm": 115.90371704101562,
156
+ "learning_rate": 1.930474333983106e-05,
157
+ "loss": 2.5846,
158
+ "step": 450
159
+ },
160
+ {
161
+ "epoch": 2.7812041116005872,
162
+ "grad_norm": 57.7183952331543,
163
+ "learning_rate": 1.9142300194931776e-05,
164
+ "loss": 2.7807,
165
+ "step": 475
166
+ },
167
+ {
168
+ "epoch": 2.9280469897209986,
169
+ "grad_norm": 52.94435501098633,
170
+ "learning_rate": 1.897985705003249e-05,
171
+ "loss": 2.4913,
172
+ "step": 500
173
+ },
174
+ {
175
+ "epoch": 3.0,
176
+ "eval_accuracy": 0.7740213523131673,
177
+ "eval_loss": 0.5719616413116455,
178
+ "eval_macro_f1": 0.7688477830945377,
179
+ "eval_runtime": 1.2147,
180
+ "eval_samples_per_second": 462.655,
181
+ "eval_steps_per_second": 29.636,
182
+ "eval_weighted_f1": 0.7688406479623778,
183
+ "step": 513
184
+ },
185
+ {
186
+ "epoch": 3.0704845814977975,
187
+ "grad_norm": 93.01853942871094,
188
+ "learning_rate": 1.8817413905133203e-05,
189
+ "loss": 2.4572,
190
+ "step": 525
191
+ },
192
+ {
193
+ "epoch": 3.2173274596182084,
194
+ "grad_norm": 63.457862854003906,
195
+ "learning_rate": 1.8654970760233918e-05,
196
+ "loss": 2.3328,
197
+ "step": 550
198
+ },
199
+ {
200
+ "epoch": 3.36417033773862,
201
+ "grad_norm": 54.87974548339844,
202
+ "learning_rate": 1.8492527615334633e-05,
203
+ "loss": 2.0633,
204
+ "step": 575
205
+ },
206
+ {
207
+ "epoch": 3.511013215859031,
208
+ "grad_norm": 84.1146469116211,
209
+ "learning_rate": 1.833008447043535e-05,
210
+ "loss": 2.1194,
211
+ "step": 600
212
+ },
213
+ {
214
+ "epoch": 3.657856093979442,
215
+ "grad_norm": 111.54753112792969,
216
+ "learning_rate": 1.8167641325536064e-05,
217
+ "loss": 2.2266,
218
+ "step": 625
219
+ },
220
+ {
221
+ "epoch": 3.804698972099853,
222
+ "grad_norm": 34.23565673828125,
223
+ "learning_rate": 1.800519818063678e-05,
224
+ "loss": 2.006,
225
+ "step": 650
226
+ },
227
+ {
228
+ "epoch": 3.951541850220264,
229
+ "grad_norm": 96.39556121826172,
230
+ "learning_rate": 1.7842755035737494e-05,
231
+ "loss": 2.0311,
232
+ "step": 675
233
+ },
234
+ {
235
+ "epoch": 4.0,
236
+ "eval_accuracy": 0.7900355871886121,
237
+ "eval_loss": 0.4983833432197571,
238
+ "eval_macro_f1": 0.7929140773256975,
239
+ "eval_runtime": 1.0853,
240
+ "eval_samples_per_second": 517.822,
241
+ "eval_steps_per_second": 33.17,
242
+ "eval_weighted_f1": 0.7928571005707712,
243
+ "step": 684
244
+ },
245
+ {
246
+ "epoch": 4.093979441997063,
247
+ "grad_norm": 71.5910415649414,
248
+ "learning_rate": 1.768031189083821e-05,
249
+ "loss": 2.0338,
250
+ "step": 700
251
+ },
252
+ {
253
+ "epoch": 4.240822320117474,
254
+ "grad_norm": 110.11136627197266,
255
+ "learning_rate": 1.751786874593892e-05,
256
+ "loss": 1.7973,
257
+ "step": 725
258
+ },
259
+ {
260
+ "epoch": 4.387665198237886,
261
+ "grad_norm": 70.80055236816406,
262
+ "learning_rate": 1.7355425601039636e-05,
263
+ "loss": 1.6429,
264
+ "step": 750
265
+ },
266
+ {
267
+ "epoch": 4.534508076358296,
268
+ "grad_norm": 84.0125961303711,
269
+ "learning_rate": 1.719298245614035e-05,
270
+ "loss": 1.5697,
271
+ "step": 775
272
+ },
273
+ {
274
+ "epoch": 4.681350954478708,
275
+ "grad_norm": 61.205753326416016,
276
+ "learning_rate": 1.7030539311241066e-05,
277
+ "loss": 1.503,
278
+ "step": 800
279
+ },
280
+ {
281
+ "epoch": 4.828193832599119,
282
+ "grad_norm": 90.11736297607422,
283
+ "learning_rate": 1.686809616634178e-05,
284
+ "loss": 1.5126,
285
+ "step": 825
286
+ },
287
+ {
288
+ "epoch": 4.975036710719531,
289
+ "grad_norm": 114.29306030273438,
290
+ "learning_rate": 1.6705653021442497e-05,
291
+ "loss": 1.6841,
292
+ "step": 850
293
+ },
294
+ {
295
+ "epoch": 5.0,
296
+ "eval_accuracy": 0.8185053380782918,
297
+ "eval_loss": 0.5628421306610107,
298
+ "eval_macro_f1": 0.8165680404186829,
299
+ "eval_runtime": 1.2562,
300
+ "eval_samples_per_second": 447.37,
301
+ "eval_steps_per_second": 28.657,
302
+ "eval_weighted_f1": 0.8158045472592881,
303
+ "step": 855
304
+ },
305
+ {
306
+ "epoch": 5.117474302496329,
307
+ "grad_norm": 73.93684387207031,
308
+ "learning_rate": 1.654320987654321e-05,
309
+ "loss": 1.3527,
310
+ "step": 875
311
+ },
312
+ {
313
+ "epoch": 5.26431718061674,
314
+ "grad_norm": 65.49266815185547,
315
+ "learning_rate": 1.6380766731643927e-05,
316
+ "loss": 1.4592,
317
+ "step": 900
318
+ },
319
+ {
320
+ "epoch": 5.411160058737151,
321
+ "grad_norm": 89.7956771850586,
322
+ "learning_rate": 1.621832358674464e-05,
323
+ "loss": 1.4047,
324
+ "step": 925
325
+ },
326
+ {
327
+ "epoch": 5.558002936857562,
328
+ "grad_norm": 201.89585876464844,
329
+ "learning_rate": 1.6055880441845354e-05,
330
+ "loss": 1.2312,
331
+ "step": 950
332
+ },
333
+ {
334
+ "epoch": 5.704845814977974,
335
+ "grad_norm": 67.38060760498047,
336
+ "learning_rate": 1.589343729694607e-05,
337
+ "loss": 1.3537,
338
+ "step": 975
339
+ },
340
+ {
341
+ "epoch": 5.851688693098385,
342
+ "grad_norm": 129.84413146972656,
343
+ "learning_rate": 1.5730994152046787e-05,
344
+ "loss": 1.4086,
345
+ "step": 1000
346
+ },
347
+ {
348
+ "epoch": 5.998531571218796,
349
+ "grad_norm": 103.0914535522461,
350
+ "learning_rate": 1.55685510071475e-05,
351
+ "loss": 1.3403,
352
+ "step": 1025
353
+ },
354
+ {
355
+ "epoch": 6.0,
356
+ "eval_accuracy": 0.8256227758007118,
357
+ "eval_loss": 0.566539466381073,
358
+ "eval_macro_f1": 0.8219632408216548,
359
+ "eval_runtime": 1.0559,
360
+ "eval_samples_per_second": 532.269,
361
+ "eval_steps_per_second": 34.096,
362
+ "eval_weighted_f1": 0.8219667601620793,
363
+ "step": 1026
364
+ },
365
+ {
366
+ "epoch": 6.140969162995595,
367
+ "grad_norm": 162.61659240722656,
368
+ "learning_rate": 1.5406107862248214e-05,
369
+ "loss": 1.1035,
370
+ "step": 1050
371
+ },
372
+ {
373
+ "epoch": 6.2878120411160054,
374
+ "grad_norm": 38.425254821777344,
375
+ "learning_rate": 1.524366471734893e-05,
376
+ "loss": 1.0021,
377
+ "step": 1075
378
+ },
379
+ {
380
+ "epoch": 6.434654919236417,
381
+ "grad_norm": 116.96482849121094,
382
+ "learning_rate": 1.5081221572449645e-05,
383
+ "loss": 0.9721,
384
+ "step": 1100
385
+ },
386
+ {
387
+ "epoch": 6.581497797356828,
388
+ "grad_norm": 20.544830322265625,
389
+ "learning_rate": 1.4918778427550358e-05,
390
+ "loss": 0.75,
391
+ "step": 1125
392
+ },
393
+ {
394
+ "epoch": 6.72834067547724,
395
+ "grad_norm": 50.387577056884766,
396
+ "learning_rate": 1.4756335282651073e-05,
397
+ "loss": 1.0023,
398
+ "step": 1150
399
+ },
400
+ {
401
+ "epoch": 6.87518355359765,
402
+ "grad_norm": 42.758583068847656,
403
+ "learning_rate": 1.4593892137751788e-05,
404
+ "loss": 0.9204,
405
+ "step": 1175
406
+ },
407
+ {
408
+ "epoch": 7.0,
409
+ "eval_accuracy": 0.8309608540925267,
410
+ "eval_loss": 0.5813337564468384,
411
+ "eval_macro_f1": 0.8295240383812666,
412
+ "eval_runtime": 1.02,
413
+ "eval_samples_per_second": 550.964,
414
+ "eval_steps_per_second": 35.293,
415
+ "eval_weighted_f1": 0.8292286443388538,
416
+ "step": 1197
417
+ }
418
+ ],
419
+ "logging_steps": 25,
420
+ "max_steps": 3420,
421
+ "num_input_tokens_seen": 0,
422
+ "num_train_epochs": 20,
423
+ "save_steps": 500,
424
+ "stateful_callbacks": {
425
+ "EarlyStoppingCallback": {
426
+ "args": {
427
+ "early_stopping_patience": 3,
428
+ "early_stopping_threshold": 0.0
429
+ },
430
+ "attributes": {
431
+ "early_stopping_patience_counter": 0
432
+ }
433
+ },
434
+ "TrainerControl": {
435
+ "args": {
436
+ "should_epoch_stop": false,
437
+ "should_evaluate": false,
438
+ "should_log": false,
439
+ "should_save": true,
440
+ "should_training_stop": false
441
+ },
442
+ "attributes": {}
443
+ }
444
+ },
445
+ "total_flos": 0.0,
446
+ "train_batch_size": 8,
447
+ "trial_name": null,
448
+ "trial_params": null
449
+ }
ood-vifactcheck-xlmr-source-source-01-s42-20260901T152349Z-27bda108/trainer/checkpoint-1710/special_tokens_map.json ADDED
@@ -0,0 +1,15 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "bos_token": "<s>",
3
+ "cls_token": "<s>",
4
+ "eos_token": "</s>",
5
+ "mask_token": {
6
+ "content": "<mask>",
7
+ "lstrip": true,
8
+ "normalized": false,
9
+ "rstrip": false,
10
+ "single_word": false
11
+ },
12
+ "pad_token": "<pad>",
13
+ "sep_token": "</s>",
14
+ "unk_token": "<unk>"
15
+ }
ood-vifactcheck-xlmr-source-source-01-s42-20260901T152349Z-27bda108/trainer/checkpoint-1710/tokenizer_config.json ADDED
@@ -0,0 +1,55 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "added_tokens_decoder": {
3
+ "0": {
4
+ "content": "<s>",
5
+ "lstrip": false,
6
+ "normalized": false,
7
+ "rstrip": false,
8
+ "single_word": false,
9
+ "special": true
10
+ },
11
+ "1": {
12
+ "content": "<pad>",
13
+ "lstrip": false,
14
+ "normalized": false,
15
+ "rstrip": false,
16
+ "single_word": false,
17
+ "special": true
18
+ },
19
+ "2": {
20
+ "content": "</s>",
21
+ "lstrip": false,
22
+ "normalized": false,
23
+ "rstrip": false,
24
+ "single_word": false,
25
+ "special": true
26
+ },
27
+ "3": {
28
+ "content": "<unk>",
29
+ "lstrip": false,
30
+ "normalized": false,
31
+ "rstrip": false,
32
+ "single_word": false,
33
+ "special": true
34
+ },
35
+ "250001": {
36
+ "content": "<mask>",
37
+ "lstrip": true,
38
+ "normalized": false,
39
+ "rstrip": false,
40
+ "single_word": false,
41
+ "special": true
42
+ }
43
+ },
44
+ "bos_token": "<s>",
45
+ "clean_up_tokenization_spaces": false,
46
+ "cls_token": "<s>",
47
+ "eos_token": "</s>",
48
+ "extra_special_tokens": {},
49
+ "mask_token": "<mask>",
50
+ "model_max_length": 512,
51
+ "pad_token": "<pad>",
52
+ "sep_token": "</s>",
53
+ "tokenizer_class": "XLMRobertaTokenizer",
54
+ "unk_token": "<unk>"
55
+ }
ood-vifactcheck-xlmr-source-source-01-s42-20260901T152349Z-27bda108/trainer/checkpoint-1710/trainer_state.json ADDED
@@ -0,0 +1,629 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_global_step": 1197,
3
+ "best_metric": 0.8295240383812666,
4
+ "best_model_checkpoint": "/root/VNEncoder/results/checkpoints/ood-vifactcheck-xlmr-source-source-01-s42-20260901T152349Z-27bda108/trainer/checkpoint-1197",
5
+ "epoch": 10.0,
6
+ "eval_steps": 500,
7
+ "global_step": 1710,
8
+ "is_hyper_param_search": false,
9
+ "is_local_process_zero": true,
10
+ "is_world_process_zero": true,
11
+ "log_history": [
12
+ {
13
+ "epoch": 0.14684287812041116,
14
+ "grad_norm": 8.276679992675781,
15
+ "learning_rate": 1.4035087719298246e-06,
16
+ "loss": 4.7457,
17
+ "step": 25
18
+ },
19
+ {
20
+ "epoch": 0.2936857562408223,
21
+ "grad_norm": 32.80095672607422,
22
+ "learning_rate": 2.865497076023392e-06,
23
+ "loss": 4.8125,
24
+ "step": 50
25
+ },
26
+ {
27
+ "epoch": 0.44052863436123346,
28
+ "grad_norm": 23.083709716796875,
29
+ "learning_rate": 4.3274853801169596e-06,
30
+ "loss": 4.6282,
31
+ "step": 75
32
+ },
33
+ {
34
+ "epoch": 0.5873715124816447,
35
+ "grad_norm": 25.51261329650879,
36
+ "learning_rate": 5.789473684210527e-06,
37
+ "loss": 4.6667,
38
+ "step": 100
39
+ },
40
+ {
41
+ "epoch": 0.7342143906020558,
42
+ "grad_norm": 41.84646224975586,
43
+ "learning_rate": 7.251461988304094e-06,
44
+ "loss": 4.593,
45
+ "step": 125
46
+ },
47
+ {
48
+ "epoch": 0.8810572687224669,
49
+ "grad_norm": 29.750471115112305,
50
+ "learning_rate": 8.713450292397661e-06,
51
+ "loss": 4.447,
52
+ "step": 150
53
+ },
54
+ {
55
+ "epoch": 1.0,
56
+ "eval_accuracy": 0.3736654804270463,
57
+ "eval_loss": 1.0819200277328491,
58
+ "eval_macro_f1": 0.29437285769412447,
59
+ "eval_runtime": 1.0765,
60
+ "eval_samples_per_second": 522.066,
61
+ "eval_steps_per_second": 33.442,
62
+ "eval_weighted_f1": 0.2922100579819657,
63
+ "step": 171
64
+ },
65
+ {
66
+ "epoch": 1.0234948604992657,
67
+ "grad_norm": 33.48249816894531,
68
+ "learning_rate": 1.017543859649123e-05,
69
+ "loss": 4.2602,
70
+ "step": 175
71
+ },
72
+ {
73
+ "epoch": 1.170337738619677,
74
+ "grad_norm": 30.328693389892578,
75
+ "learning_rate": 1.1637426900584797e-05,
76
+ "loss": 4.2628,
77
+ "step": 200
78
+ },
79
+ {
80
+ "epoch": 1.3171806167400881,
81
+ "grad_norm": 53.562705993652344,
82
+ "learning_rate": 1.3099415204678362e-05,
83
+ "loss": 3.9631,
84
+ "step": 225
85
+ },
86
+ {
87
+ "epoch": 1.4640234948604993,
88
+ "grad_norm": 64.7986831665039,
89
+ "learning_rate": 1.4561403508771931e-05,
90
+ "loss": 4.2405,
91
+ "step": 250
92
+ },
93
+ {
94
+ "epoch": 1.6108663729809103,
95
+ "grad_norm": 63.56324005126953,
96
+ "learning_rate": 1.60233918128655e-05,
97
+ "loss": 3.8987,
98
+ "step": 275
99
+ },
100
+ {
101
+ "epoch": 1.7577092511013217,
102
+ "grad_norm": 60.7152214050293,
103
+ "learning_rate": 1.7485380116959064e-05,
104
+ "loss": 3.6177,
105
+ "step": 300
106
+ },
107
+ {
108
+ "epoch": 1.9045521292217327,
109
+ "grad_norm": 60.10673141479492,
110
+ "learning_rate": 1.894736842105263e-05,
111
+ "loss": 3.4457,
112
+ "step": 325
113
+ },
114
+ {
115
+ "epoch": 2.0,
116
+ "eval_accuracy": 0.701067615658363,
117
+ "eval_loss": 0.6930170655250549,
118
+ "eval_macro_f1": 0.6986001077282614,
119
+ "eval_runtime": 1.1239,
120
+ "eval_samples_per_second": 500.04,
121
+ "eval_steps_per_second": 32.031,
122
+ "eval_weighted_f1": 0.6982019631794543,
123
+ "step": 342
124
+ },
125
+ {
126
+ "epoch": 2.0469897209985315,
127
+ "grad_norm": 32.58366012573242,
128
+ "learning_rate": 1.99545159194282e-05,
129
+ "loss": 3.0666,
130
+ "step": 350
131
+ },
132
+ {
133
+ "epoch": 2.193832599118943,
134
+ "grad_norm": 94.89712524414062,
135
+ "learning_rate": 1.9792072774528916e-05,
136
+ "loss": 3.0562,
137
+ "step": 375
138
+ },
139
+ {
140
+ "epoch": 2.340675477239354,
141
+ "grad_norm": 50.320068359375,
142
+ "learning_rate": 1.962962962962963e-05,
143
+ "loss": 2.7734,
144
+ "step": 400
145
+ },
146
+ {
147
+ "epoch": 2.4875183553597653,
148
+ "grad_norm": 93.8223648071289,
149
+ "learning_rate": 1.9467186484730346e-05,
150
+ "loss": 2.6324,
151
+ "step": 425
152
+ },
153
+ {
154
+ "epoch": 2.6343612334801763,
155
+ "grad_norm": 115.90371704101562,
156
+ "learning_rate": 1.930474333983106e-05,
157
+ "loss": 2.5846,
158
+ "step": 450
159
+ },
160
+ {
161
+ "epoch": 2.7812041116005872,
162
+ "grad_norm": 57.7183952331543,
163
+ "learning_rate": 1.9142300194931776e-05,
164
+ "loss": 2.7807,
165
+ "step": 475
166
+ },
167
+ {
168
+ "epoch": 2.9280469897209986,
169
+ "grad_norm": 52.94435501098633,
170
+ "learning_rate": 1.897985705003249e-05,
171
+ "loss": 2.4913,
172
+ "step": 500
173
+ },
174
+ {
175
+ "epoch": 3.0,
176
+ "eval_accuracy": 0.7740213523131673,
177
+ "eval_loss": 0.5719616413116455,
178
+ "eval_macro_f1": 0.7688477830945377,
179
+ "eval_runtime": 1.2147,
180
+ "eval_samples_per_second": 462.655,
181
+ "eval_steps_per_second": 29.636,
182
+ "eval_weighted_f1": 0.7688406479623778,
183
+ "step": 513
184
+ },
185
+ {
186
+ "epoch": 3.0704845814977975,
187
+ "grad_norm": 93.01853942871094,
188
+ "learning_rate": 1.8817413905133203e-05,
189
+ "loss": 2.4572,
190
+ "step": 525
191
+ },
192
+ {
193
+ "epoch": 3.2173274596182084,
194
+ "grad_norm": 63.457862854003906,
195
+ "learning_rate": 1.8654970760233918e-05,
196
+ "loss": 2.3328,
197
+ "step": 550
198
+ },
199
+ {
200
+ "epoch": 3.36417033773862,
201
+ "grad_norm": 54.87974548339844,
202
+ "learning_rate": 1.8492527615334633e-05,
203
+ "loss": 2.0633,
204
+ "step": 575
205
+ },
206
+ {
207
+ "epoch": 3.511013215859031,
208
+ "grad_norm": 84.1146469116211,
209
+ "learning_rate": 1.833008447043535e-05,
210
+ "loss": 2.1194,
211
+ "step": 600
212
+ },
213
+ {
214
+ "epoch": 3.657856093979442,
215
+ "grad_norm": 111.54753112792969,
216
+ "learning_rate": 1.8167641325536064e-05,
217
+ "loss": 2.2266,
218
+ "step": 625
219
+ },
220
+ {
221
+ "epoch": 3.804698972099853,
222
+ "grad_norm": 34.23565673828125,
223
+ "learning_rate": 1.800519818063678e-05,
224
+ "loss": 2.006,
225
+ "step": 650
226
+ },
227
+ {
228
+ "epoch": 3.951541850220264,
229
+ "grad_norm": 96.39556121826172,
230
+ "learning_rate": 1.7842755035737494e-05,
231
+ "loss": 2.0311,
232
+ "step": 675
233
+ },
234
+ {
235
+ "epoch": 4.0,
236
+ "eval_accuracy": 0.7900355871886121,
237
+ "eval_loss": 0.4983833432197571,
238
+ "eval_macro_f1": 0.7929140773256975,
239
+ "eval_runtime": 1.0853,
240
+ "eval_samples_per_second": 517.822,
241
+ "eval_steps_per_second": 33.17,
242
+ "eval_weighted_f1": 0.7928571005707712,
243
+ "step": 684
244
+ },
245
+ {
246
+ "epoch": 4.093979441997063,
247
+ "grad_norm": 71.5910415649414,
248
+ "learning_rate": 1.768031189083821e-05,
249
+ "loss": 2.0338,
250
+ "step": 700
251
+ },
252
+ {
253
+ "epoch": 4.240822320117474,
254
+ "grad_norm": 110.11136627197266,
255
+ "learning_rate": 1.751786874593892e-05,
256
+ "loss": 1.7973,
257
+ "step": 725
258
+ },
259
+ {
260
+ "epoch": 4.387665198237886,
261
+ "grad_norm": 70.80055236816406,
262
+ "learning_rate": 1.7355425601039636e-05,
263
+ "loss": 1.6429,
264
+ "step": 750
265
+ },
266
+ {
267
+ "epoch": 4.534508076358296,
268
+ "grad_norm": 84.0125961303711,
269
+ "learning_rate": 1.719298245614035e-05,
270
+ "loss": 1.5697,
271
+ "step": 775
272
+ },
273
+ {
274
+ "epoch": 4.681350954478708,
275
+ "grad_norm": 61.205753326416016,
276
+ "learning_rate": 1.7030539311241066e-05,
277
+ "loss": 1.503,
278
+ "step": 800
279
+ },
280
+ {
281
+ "epoch": 4.828193832599119,
282
+ "grad_norm": 90.11736297607422,
283
+ "learning_rate": 1.686809616634178e-05,
284
+ "loss": 1.5126,
285
+ "step": 825
286
+ },
287
+ {
288
+ "epoch": 4.975036710719531,
289
+ "grad_norm": 114.29306030273438,
290
+ "learning_rate": 1.6705653021442497e-05,
291
+ "loss": 1.6841,
292
+ "step": 850
293
+ },
294
+ {
295
+ "epoch": 5.0,
296
+ "eval_accuracy": 0.8185053380782918,
297
+ "eval_loss": 0.5628421306610107,
298
+ "eval_macro_f1": 0.8165680404186829,
299
+ "eval_runtime": 1.2562,
300
+ "eval_samples_per_second": 447.37,
301
+ "eval_steps_per_second": 28.657,
302
+ "eval_weighted_f1": 0.8158045472592881,
303
+ "step": 855
304
+ },
305
+ {
306
+ "epoch": 5.117474302496329,
307
+ "grad_norm": 73.93684387207031,
308
+ "learning_rate": 1.654320987654321e-05,
309
+ "loss": 1.3527,
310
+ "step": 875
311
+ },
312
+ {
313
+ "epoch": 5.26431718061674,
314
+ "grad_norm": 65.49266815185547,
315
+ "learning_rate": 1.6380766731643927e-05,
316
+ "loss": 1.4592,
317
+ "step": 900
318
+ },
319
+ {
320
+ "epoch": 5.411160058737151,
321
+ "grad_norm": 89.7956771850586,
322
+ "learning_rate": 1.621832358674464e-05,
323
+ "loss": 1.4047,
324
+ "step": 925
325
+ },
326
+ {
327
+ "epoch": 5.558002936857562,
328
+ "grad_norm": 201.89585876464844,
329
+ "learning_rate": 1.6055880441845354e-05,
330
+ "loss": 1.2312,
331
+ "step": 950
332
+ },
333
+ {
334
+ "epoch": 5.704845814977974,
335
+ "grad_norm": 67.38060760498047,
336
+ "learning_rate": 1.589343729694607e-05,
337
+ "loss": 1.3537,
338
+ "step": 975
339
+ },
340
+ {
341
+ "epoch": 5.851688693098385,
342
+ "grad_norm": 129.84413146972656,
343
+ "learning_rate": 1.5730994152046787e-05,
344
+ "loss": 1.4086,
345
+ "step": 1000
346
+ },
347
+ {
348
+ "epoch": 5.998531571218796,
349
+ "grad_norm": 103.0914535522461,
350
+ "learning_rate": 1.55685510071475e-05,
351
+ "loss": 1.3403,
352
+ "step": 1025
353
+ },
354
+ {
355
+ "epoch": 6.0,
356
+ "eval_accuracy": 0.8256227758007118,
357
+ "eval_loss": 0.566539466381073,
358
+ "eval_macro_f1": 0.8219632408216548,
359
+ "eval_runtime": 1.0559,
360
+ "eval_samples_per_second": 532.269,
361
+ "eval_steps_per_second": 34.096,
362
+ "eval_weighted_f1": 0.8219667601620793,
363
+ "step": 1026
364
+ },
365
+ {
366
+ "epoch": 6.140969162995595,
367
+ "grad_norm": 162.61659240722656,
368
+ "learning_rate": 1.5406107862248214e-05,
369
+ "loss": 1.1035,
370
+ "step": 1050
371
+ },
372
+ {
373
+ "epoch": 6.2878120411160054,
374
+ "grad_norm": 38.425254821777344,
375
+ "learning_rate": 1.524366471734893e-05,
376
+ "loss": 1.0021,
377
+ "step": 1075
378
+ },
379
+ {
380
+ "epoch": 6.434654919236417,
381
+ "grad_norm": 116.96482849121094,
382
+ "learning_rate": 1.5081221572449645e-05,
383
+ "loss": 0.9721,
384
+ "step": 1100
385
+ },
386
+ {
387
+ "epoch": 6.581497797356828,
388
+ "grad_norm": 20.544830322265625,
389
+ "learning_rate": 1.4918778427550358e-05,
390
+ "loss": 0.75,
391
+ "step": 1125
392
+ },
393
+ {
394
+ "epoch": 6.72834067547724,
395
+ "grad_norm": 50.387577056884766,
396
+ "learning_rate": 1.4756335282651073e-05,
397
+ "loss": 1.0023,
398
+ "step": 1150
399
+ },
400
+ {
401
+ "epoch": 6.87518355359765,
402
+ "grad_norm": 42.758583068847656,
403
+ "learning_rate": 1.4593892137751788e-05,
404
+ "loss": 0.9204,
405
+ "step": 1175
406
+ },
407
+ {
408
+ "epoch": 7.0,
409
+ "eval_accuracy": 0.8309608540925267,
410
+ "eval_loss": 0.5813337564468384,
411
+ "eval_macro_f1": 0.8295240383812666,
412
+ "eval_runtime": 1.02,
413
+ "eval_samples_per_second": 550.964,
414
+ "eval_steps_per_second": 35.293,
415
+ "eval_weighted_f1": 0.8292286443388538,
416
+ "step": 1197
417
+ },
418
+ {
419
+ "epoch": 7.0176211453744495,
420
+ "grad_norm": 169.1999053955078,
421
+ "learning_rate": 1.4431448992852503e-05,
422
+ "loss": 1.1155,
423
+ "step": 1200
424
+ },
425
+ {
426
+ "epoch": 7.164464023494861,
427
+ "grad_norm": 72.44261932373047,
428
+ "learning_rate": 1.4269005847953217e-05,
429
+ "loss": 0.6562,
430
+ "step": 1225
431
+ },
432
+ {
433
+ "epoch": 7.311306901615271,
434
+ "grad_norm": 70.4012451171875,
435
+ "learning_rate": 1.4106562703053932e-05,
436
+ "loss": 0.7291,
437
+ "step": 1250
438
+ },
439
+ {
440
+ "epoch": 7.458149779735683,
441
+ "grad_norm": 46.89229965209961,
442
+ "learning_rate": 1.3944119558154647e-05,
443
+ "loss": 0.6769,
444
+ "step": 1275
445
+ },
446
+ {
447
+ "epoch": 7.604992657856094,
448
+ "grad_norm": 80.89586639404297,
449
+ "learning_rate": 1.3781676413255362e-05,
450
+ "loss": 0.7291,
451
+ "step": 1300
452
+ },
453
+ {
454
+ "epoch": 7.751835535976505,
455
+ "grad_norm": 144.99386596679688,
456
+ "learning_rate": 1.3619233268356076e-05,
457
+ "loss": 0.7643,
458
+ "step": 1325
459
+ },
460
+ {
461
+ "epoch": 7.898678414096916,
462
+ "grad_norm": 59.30048370361328,
463
+ "learning_rate": 1.345679012345679e-05,
464
+ "loss": 0.6184,
465
+ "step": 1350
466
+ },
467
+ {
468
+ "epoch": 8.0,
469
+ "eval_accuracy": 0.8202846975088968,
470
+ "eval_loss": 0.6676938533782959,
471
+ "eval_macro_f1": 0.8193423338206208,
472
+ "eval_runtime": 1.2476,
473
+ "eval_samples_per_second": 450.467,
474
+ "eval_steps_per_second": 28.856,
475
+ "eval_weighted_f1": 0.81896413525029,
476
+ "step": 1368
477
+ },
478
+ {
479
+ "epoch": 8.041116005873715,
480
+ "grad_norm": 107.24308776855469,
481
+ "learning_rate": 1.3294346978557506e-05,
482
+ "loss": 0.7401,
483
+ "step": 1375
484
+ },
485
+ {
486
+ "epoch": 8.187958883994126,
487
+ "grad_norm": 8.43368148803711,
488
+ "learning_rate": 1.3131903833658221e-05,
489
+ "loss": 0.5523,
490
+ "step": 1400
491
+ },
492
+ {
493
+ "epoch": 8.334801762114537,
494
+ "grad_norm": 91.36034393310547,
495
+ "learning_rate": 1.2969460688758935e-05,
496
+ "loss": 0.5694,
497
+ "step": 1425
498
+ },
499
+ {
500
+ "epoch": 8.481644640234949,
501
+ "grad_norm": 76.99760437011719,
502
+ "learning_rate": 1.280701754385965e-05,
503
+ "loss": 0.5402,
504
+ "step": 1450
505
+ },
506
+ {
507
+ "epoch": 8.62848751835536,
508
+ "grad_norm": 19.6176815032959,
509
+ "learning_rate": 1.2644574398960365e-05,
510
+ "loss": 0.454,
511
+ "step": 1475
512
+ },
513
+ {
514
+ "epoch": 8.775330396475772,
515
+ "grad_norm": 89.50540924072266,
516
+ "learning_rate": 1.248213125406108e-05,
517
+ "loss": 0.7104,
518
+ "step": 1500
519
+ },
520
+ {
521
+ "epoch": 8.922173274596181,
522
+ "grad_norm": 54.60981369018555,
523
+ "learning_rate": 1.2319688109161793e-05,
524
+ "loss": 0.5546,
525
+ "step": 1525
526
+ },
527
+ {
528
+ "epoch": 9.0,
529
+ "eval_accuracy": 0.8185053380782918,
530
+ "eval_loss": 0.7808212041854858,
531
+ "eval_macro_f1": 0.8177603560755428,
532
+ "eval_runtime": 1.2467,
533
+ "eval_samples_per_second": 450.784,
534
+ "eval_steps_per_second": 28.876,
535
+ "eval_weighted_f1": 0.8177147662693375,
536
+ "step": 1539
537
+ },
538
+ {
539
+ "epoch": 9.06461086637298,
540
+ "grad_norm": 49.732154846191406,
541
+ "learning_rate": 1.2157244964262509e-05,
542
+ "loss": 0.6509,
543
+ "step": 1550
544
+ },
545
+ {
546
+ "epoch": 9.211453744493392,
547
+ "grad_norm": 120.88252258300781,
548
+ "learning_rate": 1.1994801819363224e-05,
549
+ "loss": 0.4404,
550
+ "step": 1575
551
+ },
552
+ {
553
+ "epoch": 9.358296622613803,
554
+ "grad_norm": 16.407711029052734,
555
+ "learning_rate": 1.1832358674463939e-05,
556
+ "loss": 0.2951,
557
+ "step": 1600
558
+ },
559
+ {
560
+ "epoch": 9.505139500734215,
561
+ "grad_norm": 194.71717834472656,
562
+ "learning_rate": 1.1669915529564652e-05,
563
+ "loss": 0.4514,
564
+ "step": 1625
565
+ },
566
+ {
567
+ "epoch": 9.651982378854626,
568
+ "grad_norm": 97.11231231689453,
569
+ "learning_rate": 1.1507472384665367e-05,
570
+ "loss": 0.3223,
571
+ "step": 1650
572
+ },
573
+ {
574
+ "epoch": 9.798825256975036,
575
+ "grad_norm": 35.31584167480469,
576
+ "learning_rate": 1.1345029239766083e-05,
577
+ "loss": 0.6108,
578
+ "step": 1675
579
+ },
580
+ {
581
+ "epoch": 9.945668135095447,
582
+ "grad_norm": 122.03335571289062,
583
+ "learning_rate": 1.11825860948668e-05,
584
+ "loss": 0.492,
585
+ "step": 1700
586
+ },
587
+ {
588
+ "epoch": 10.0,
589
+ "eval_accuracy": 0.8274021352313167,
590
+ "eval_loss": 0.8296598792076111,
591
+ "eval_macro_f1": 0.8281855148698446,
592
+ "eval_runtime": 1.2851,
593
+ "eval_samples_per_second": 437.305,
594
+ "eval_steps_per_second": 28.012,
595
+ "eval_weighted_f1": 0.8277867192540188,
596
+ "step": 1710
597
+ }
598
+ ],
599
+ "logging_steps": 25,
600
+ "max_steps": 3420,
601
+ "num_input_tokens_seen": 0,
602
+ "num_train_epochs": 20,
603
+ "save_steps": 500,
604
+ "stateful_callbacks": {
605
+ "EarlyStoppingCallback": {
606
+ "args": {
607
+ "early_stopping_patience": 3,
608
+ "early_stopping_threshold": 0.0
609
+ },
610
+ "attributes": {
611
+ "early_stopping_patience_counter": 3
612
+ }
613
+ },
614
+ "TrainerControl": {
615
+ "args": {
616
+ "should_epoch_stop": false,
617
+ "should_evaluate": false,
618
+ "should_log": false,
619
+ "should_save": true,
620
+ "should_training_stop": true
621
+ },
622
+ "attributes": {}
623
+ }
624
+ },
625
+ "total_flos": 0.0,
626
+ "train_batch_size": 8,
627
+ "trial_name": null,
628
+ "trial_params": null
629
+ }
ood-vifactcheck-xlmr-source-source-01-s43-20260901T170609Z-fbd63439/trainer/checkpoint-1197/special_tokens_map.json ADDED
@@ -0,0 +1,15 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "bos_token": "<s>",
3
+ "cls_token": "<s>",
4
+ "eos_token": "</s>",
5
+ "mask_token": {
6
+ "content": "<mask>",
7
+ "lstrip": true,
8
+ "normalized": false,
9
+ "rstrip": false,
10
+ "single_word": false
11
+ },
12
+ "pad_token": "<pad>",
13
+ "sep_token": "</s>",
14
+ "unk_token": "<unk>"
15
+ }
ood-vifactcheck-xlmr-source-source-01-s43-20260901T170609Z-fbd63439/trainer/checkpoint-1197/tokenizer_config.json ADDED
@@ -0,0 +1,55 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "added_tokens_decoder": {
3
+ "0": {
4
+ "content": "<s>",
5
+ "lstrip": false,
6
+ "normalized": false,
7
+ "rstrip": false,
8
+ "single_word": false,
9
+ "special": true
10
+ },
11
+ "1": {
12
+ "content": "<pad>",
13
+ "lstrip": false,
14
+ "normalized": false,
15
+ "rstrip": false,
16
+ "single_word": false,
17
+ "special": true
18
+ },
19
+ "2": {
20
+ "content": "</s>",
21
+ "lstrip": false,
22
+ "normalized": false,
23
+ "rstrip": false,
24
+ "single_word": false,
25
+ "special": true
26
+ },
27
+ "3": {
28
+ "content": "<unk>",
29
+ "lstrip": false,
30
+ "normalized": false,
31
+ "rstrip": false,
32
+ "single_word": false,
33
+ "special": true
34
+ },
35
+ "250001": {
36
+ "content": "<mask>",
37
+ "lstrip": true,
38
+ "normalized": false,
39
+ "rstrip": false,
40
+ "single_word": false,
41
+ "special": true
42
+ }
43
+ },
44
+ "bos_token": "<s>",
45
+ "clean_up_tokenization_spaces": false,
46
+ "cls_token": "<s>",
47
+ "eos_token": "</s>",
48
+ "extra_special_tokens": {},
49
+ "mask_token": "<mask>",
50
+ "model_max_length": 512,
51
+ "pad_token": "<pad>",
52
+ "sep_token": "</s>",
53
+ "tokenizer_class": "XLMRobertaTokenizer",
54
+ "unk_token": "<unk>"
55
+ }
ood-vifactcheck-xlmr-source-source-01-s43-20260901T170609Z-fbd63439/trainer/checkpoint-1197/trainer_state.json ADDED
@@ -0,0 +1,449 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_global_step": 1197,
3
+ "best_metric": 0.8120160172108976,
4
+ "best_model_checkpoint": "/root/VNEncoder/results/checkpoints/ood-vifactcheck-xlmr-source-source-01-s43-20260901T170609Z-fbd63439/trainer/checkpoint-1197",
5
+ "epoch": 7.0,
6
+ "eval_steps": 500,
7
+ "global_step": 1197,
8
+ "is_hyper_param_search": false,
9
+ "is_local_process_zero": true,
10
+ "is_world_process_zero": true,
11
+ "log_history": [
12
+ {
13
+ "epoch": 0.14684287812041116,
14
+ "grad_norm": 19.898252487182617,
15
+ "learning_rate": 1.4035087719298246e-06,
16
+ "loss": 4.8259,
17
+ "step": 25
18
+ },
19
+ {
20
+ "epoch": 0.2936857562408223,
21
+ "grad_norm": 28.20109748840332,
22
+ "learning_rate": 2.865497076023392e-06,
23
+ "loss": 4.6946,
24
+ "step": 50
25
+ },
26
+ {
27
+ "epoch": 0.44052863436123346,
28
+ "grad_norm": 19.910593032836914,
29
+ "learning_rate": 4.3274853801169596e-06,
30
+ "loss": 4.5263,
31
+ "step": 75
32
+ },
33
+ {
34
+ "epoch": 0.5873715124816447,
35
+ "grad_norm": 23.893070220947266,
36
+ "learning_rate": 5.789473684210527e-06,
37
+ "loss": 4.4799,
38
+ "step": 100
39
+ },
40
+ {
41
+ "epoch": 0.7342143906020558,
42
+ "grad_norm": 21.93558120727539,
43
+ "learning_rate": 7.251461988304094e-06,
44
+ "loss": 4.5204,
45
+ "step": 125
46
+ },
47
+ {
48
+ "epoch": 0.8810572687224669,
49
+ "grad_norm": 20.283721923828125,
50
+ "learning_rate": 8.713450292397661e-06,
51
+ "loss": 4.3808,
52
+ "step": 150
53
+ },
54
+ {
55
+ "epoch": 1.0,
56
+ "eval_accuracy": 0.3274021352313167,
57
+ "eval_loss": 1.1216636896133423,
58
+ "eval_macro_f1": 0.1680264699233441,
59
+ "eval_runtime": 1.1499,
60
+ "eval_samples_per_second": 488.741,
61
+ "eval_steps_per_second": 31.307,
62
+ "eval_weighted_f1": 0.16437457594901667,
63
+ "step": 171
64
+ },
65
+ {
66
+ "epoch": 1.0234948604992657,
67
+ "grad_norm": 18.82137680053711,
68
+ "learning_rate": 1.017543859649123e-05,
69
+ "loss": 4.3132,
70
+ "step": 175
71
+ },
72
+ {
73
+ "epoch": 1.170337738619677,
74
+ "grad_norm": 39.67106628417969,
75
+ "learning_rate": 1.1637426900584797e-05,
76
+ "loss": 4.3691,
77
+ "step": 200
78
+ },
79
+ {
80
+ "epoch": 1.3171806167400881,
81
+ "grad_norm": 25.474512100219727,
82
+ "learning_rate": 1.3099415204678362e-05,
83
+ "loss": 4.3744,
84
+ "step": 225
85
+ },
86
+ {
87
+ "epoch": 1.4640234948604993,
88
+ "grad_norm": 29.966018676757812,
89
+ "learning_rate": 1.4561403508771931e-05,
90
+ "loss": 4.1576,
91
+ "step": 250
92
+ },
93
+ {
94
+ "epoch": 1.6108663729809103,
95
+ "grad_norm": 38.67800521850586,
96
+ "learning_rate": 1.60233918128655e-05,
97
+ "loss": 3.9165,
98
+ "step": 275
99
+ },
100
+ {
101
+ "epoch": 1.7577092511013217,
102
+ "grad_norm": 120.15225219726562,
103
+ "learning_rate": 1.7485380116959064e-05,
104
+ "loss": 3.5492,
105
+ "step": 300
106
+ },
107
+ {
108
+ "epoch": 1.9045521292217327,
109
+ "grad_norm": 145.66622924804688,
110
+ "learning_rate": 1.894736842105263e-05,
111
+ "loss": 3.2797,
112
+ "step": 325
113
+ },
114
+ {
115
+ "epoch": 2.0,
116
+ "eval_accuracy": 0.6814946619217082,
117
+ "eval_loss": 0.8266801238059998,
118
+ "eval_macro_f1": 0.6695599664073812,
119
+ "eval_runtime": 1.0467,
120
+ "eval_samples_per_second": 536.94,
121
+ "eval_steps_per_second": 34.395,
122
+ "eval_weighted_f1": 0.6707705283670244,
123
+ "step": 342
124
+ },
125
+ {
126
+ "epoch": 2.0469897209985315,
127
+ "grad_norm": 61.46903610229492,
128
+ "learning_rate": 1.99545159194282e-05,
129
+ "loss": 3.1801,
130
+ "step": 350
131
+ },
132
+ {
133
+ "epoch": 2.193832599118943,
134
+ "grad_norm": 94.04988861083984,
135
+ "learning_rate": 1.9792072774528916e-05,
136
+ "loss": 2.9369,
137
+ "step": 375
138
+ },
139
+ {
140
+ "epoch": 2.340675477239354,
141
+ "grad_norm": 131.471923828125,
142
+ "learning_rate": 1.962962962962963e-05,
143
+ "loss": 2.8236,
144
+ "step": 400
145
+ },
146
+ {
147
+ "epoch": 2.4875183553597653,
148
+ "grad_norm": 176.6563262939453,
149
+ "learning_rate": 1.9467186484730346e-05,
150
+ "loss": 2.9867,
151
+ "step": 425
152
+ },
153
+ {
154
+ "epoch": 2.6343612334801763,
155
+ "grad_norm": 59.89003372192383,
156
+ "learning_rate": 1.930474333983106e-05,
157
+ "loss": 2.9233,
158
+ "step": 450
159
+ },
160
+ {
161
+ "epoch": 2.7812041116005872,
162
+ "grad_norm": 68.92916870117188,
163
+ "learning_rate": 1.9142300194931776e-05,
164
+ "loss": 2.8271,
165
+ "step": 475
166
+ },
167
+ {
168
+ "epoch": 2.9280469897209986,
169
+ "grad_norm": 103.18282318115234,
170
+ "learning_rate": 1.897985705003249e-05,
171
+ "loss": 2.7263,
172
+ "step": 500
173
+ },
174
+ {
175
+ "epoch": 3.0,
176
+ "eval_accuracy": 0.7686832740213523,
177
+ "eval_loss": 0.6141501069068909,
178
+ "eval_macro_f1": 0.7704908189767045,
179
+ "eval_runtime": 1.1127,
180
+ "eval_samples_per_second": 505.085,
181
+ "eval_steps_per_second": 32.354,
182
+ "eval_weighted_f1": 0.7697335190679598,
183
+ "step": 513
184
+ },
185
+ {
186
+ "epoch": 3.0704845814977975,
187
+ "grad_norm": 106.83403778076172,
188
+ "learning_rate": 1.8817413905133203e-05,
189
+ "loss": 2.5354,
190
+ "step": 525
191
+ },
192
+ {
193
+ "epoch": 3.2173274596182084,
194
+ "grad_norm": 103.26848602294922,
195
+ "learning_rate": 1.8654970760233918e-05,
196
+ "loss": 2.4149,
197
+ "step": 550
198
+ },
199
+ {
200
+ "epoch": 3.36417033773862,
201
+ "grad_norm": 84.16307067871094,
202
+ "learning_rate": 1.8492527615334633e-05,
203
+ "loss": 2.1616,
204
+ "step": 575
205
+ },
206
+ {
207
+ "epoch": 3.511013215859031,
208
+ "grad_norm": 51.95928955078125,
209
+ "learning_rate": 1.833008447043535e-05,
210
+ "loss": 2.2616,
211
+ "step": 600
212
+ },
213
+ {
214
+ "epoch": 3.657856093979442,
215
+ "grad_norm": 48.234153747558594,
216
+ "learning_rate": 1.8167641325536064e-05,
217
+ "loss": 1.916,
218
+ "step": 625
219
+ },
220
+ {
221
+ "epoch": 3.804698972099853,
222
+ "grad_norm": 97.79098510742188,
223
+ "learning_rate": 1.800519818063678e-05,
224
+ "loss": 2.1566,
225
+ "step": 650
226
+ },
227
+ {
228
+ "epoch": 3.951541850220264,
229
+ "grad_norm": 83.2856674194336,
230
+ "learning_rate": 1.7842755035737494e-05,
231
+ "loss": 2.0654,
232
+ "step": 675
233
+ },
234
+ {
235
+ "epoch": 4.0,
236
+ "eval_accuracy": 0.791814946619217,
237
+ "eval_loss": 0.5504909753799438,
238
+ "eval_macro_f1": 0.7894521890156985,
239
+ "eval_runtime": 1.1111,
240
+ "eval_samples_per_second": 505.812,
241
+ "eval_steps_per_second": 32.401,
242
+ "eval_weighted_f1": 0.7894752915277294,
243
+ "step": 684
244
+ },
245
+ {
246
+ "epoch": 4.093979441997063,
247
+ "grad_norm": 51.651371002197266,
248
+ "learning_rate": 1.768031189083821e-05,
249
+ "loss": 1.7594,
250
+ "step": 700
251
+ },
252
+ {
253
+ "epoch": 4.240822320117474,
254
+ "grad_norm": 96.34539794921875,
255
+ "learning_rate": 1.751786874593892e-05,
256
+ "loss": 1.7873,
257
+ "step": 725
258
+ },
259
+ {
260
+ "epoch": 4.387665198237886,
261
+ "grad_norm": 113.27198791503906,
262
+ "learning_rate": 1.7355425601039636e-05,
263
+ "loss": 1.6555,
264
+ "step": 750
265
+ },
266
+ {
267
+ "epoch": 4.534508076358296,
268
+ "grad_norm": 52.46745681762695,
269
+ "learning_rate": 1.719298245614035e-05,
270
+ "loss": 1.5451,
271
+ "step": 775
272
+ },
273
+ {
274
+ "epoch": 4.681350954478708,
275
+ "grad_norm": 81.5442886352539,
276
+ "learning_rate": 1.7030539311241066e-05,
277
+ "loss": 1.5964,
278
+ "step": 800
279
+ },
280
+ {
281
+ "epoch": 4.828193832599119,
282
+ "grad_norm": 155.65516662597656,
283
+ "learning_rate": 1.686809616634178e-05,
284
+ "loss": 1.7224,
285
+ "step": 825
286
+ },
287
+ {
288
+ "epoch": 4.975036710719531,
289
+ "grad_norm": 48.24635314941406,
290
+ "learning_rate": 1.6705653021442497e-05,
291
+ "loss": 1.7126,
292
+ "step": 850
293
+ },
294
+ {
295
+ "epoch": 5.0,
296
+ "eval_accuracy": 0.8078291814946619,
297
+ "eval_loss": 0.5076857805252075,
298
+ "eval_macro_f1": 0.8081435252166961,
299
+ "eval_runtime": 1.0205,
300
+ "eval_samples_per_second": 550.698,
301
+ "eval_steps_per_second": 35.276,
302
+ "eval_weighted_f1": 0.8077148227916391,
303
+ "step": 855
304
+ },
305
+ {
306
+ "epoch": 5.117474302496329,
307
+ "grad_norm": 90.79500579833984,
308
+ "learning_rate": 1.654320987654321e-05,
309
+ "loss": 1.4157,
310
+ "step": 875
311
+ },
312
+ {
313
+ "epoch": 5.26431718061674,
314
+ "grad_norm": 74.45726013183594,
315
+ "learning_rate": 1.6380766731643927e-05,
316
+ "loss": 1.3348,
317
+ "step": 900
318
+ },
319
+ {
320
+ "epoch": 5.411160058737151,
321
+ "grad_norm": 72.0616455078125,
322
+ "learning_rate": 1.621832358674464e-05,
323
+ "loss": 1.1797,
324
+ "step": 925
325
+ },
326
+ {
327
+ "epoch": 5.558002936857562,
328
+ "grad_norm": 83.30821228027344,
329
+ "learning_rate": 1.6055880441845354e-05,
330
+ "loss": 1.0697,
331
+ "step": 950
332
+ },
333
+ {
334
+ "epoch": 5.704845814977974,
335
+ "grad_norm": 72.64716339111328,
336
+ "learning_rate": 1.589343729694607e-05,
337
+ "loss": 1.3893,
338
+ "step": 975
339
+ },
340
+ {
341
+ "epoch": 5.851688693098385,
342
+ "grad_norm": 99.64823150634766,
343
+ "learning_rate": 1.5730994152046787e-05,
344
+ "loss": 1.3584,
345
+ "step": 1000
346
+ },
347
+ {
348
+ "epoch": 5.998531571218796,
349
+ "grad_norm": 97.09939575195312,
350
+ "learning_rate": 1.55685510071475e-05,
351
+ "loss": 1.5412,
352
+ "step": 1025
353
+ },
354
+ {
355
+ "epoch": 6.0,
356
+ "eval_accuracy": 0.8078291814946619,
357
+ "eval_loss": 0.5981741547584534,
358
+ "eval_macro_f1": 0.8021303737721648,
359
+ "eval_runtime": 1.1109,
360
+ "eval_samples_per_second": 505.911,
361
+ "eval_steps_per_second": 32.407,
362
+ "eval_weighted_f1": 0.802081094663556,
363
+ "step": 1026
364
+ },
365
+ {
366
+ "epoch": 6.140969162995595,
367
+ "grad_norm": 95.87456512451172,
368
+ "learning_rate": 1.5406107862248214e-05,
369
+ "loss": 1.0143,
370
+ "step": 1050
371
+ },
372
+ {
373
+ "epoch": 6.2878120411160054,
374
+ "grad_norm": 68.2560806274414,
375
+ "learning_rate": 1.524366471734893e-05,
376
+ "loss": 0.8491,
377
+ "step": 1075
378
+ },
379
+ {
380
+ "epoch": 6.434654919236417,
381
+ "grad_norm": 77.93789672851562,
382
+ "learning_rate": 1.5081221572449645e-05,
383
+ "loss": 1.0485,
384
+ "step": 1100
385
+ },
386
+ {
387
+ "epoch": 6.581497797356828,
388
+ "grad_norm": 16.324548721313477,
389
+ "learning_rate": 1.4918778427550358e-05,
390
+ "loss": 0.935,
391
+ "step": 1125
392
+ },
393
+ {
394
+ "epoch": 6.72834067547724,
395
+ "grad_norm": 65.43302917480469,
396
+ "learning_rate": 1.4756335282651073e-05,
397
+ "loss": 1.0821,
398
+ "step": 1150
399
+ },
400
+ {
401
+ "epoch": 6.87518355359765,
402
+ "grad_norm": 124.80599975585938,
403
+ "learning_rate": 1.4593892137751788e-05,
404
+ "loss": 0.9082,
405
+ "step": 1175
406
+ },
407
+ {
408
+ "epoch": 7.0,
409
+ "eval_accuracy": 0.8113879003558719,
410
+ "eval_loss": 0.6220943331718445,
411
+ "eval_macro_f1": 0.8120160172108976,
412
+ "eval_runtime": 1.0895,
413
+ "eval_samples_per_second": 515.83,
414
+ "eval_steps_per_second": 33.042,
415
+ "eval_weighted_f1": 0.8116930414813697,
416
+ "step": 1197
417
+ }
418
+ ],
419
+ "logging_steps": 25,
420
+ "max_steps": 3420,
421
+ "num_input_tokens_seen": 0,
422
+ "num_train_epochs": 20,
423
+ "save_steps": 500,
424
+ "stateful_callbacks": {
425
+ "EarlyStoppingCallback": {
426
+ "args": {
427
+ "early_stopping_patience": 3,
428
+ "early_stopping_threshold": 0.0
429
+ },
430
+ "attributes": {
431
+ "early_stopping_patience_counter": 0
432
+ }
433
+ },
434
+ "TrainerControl": {
435
+ "args": {
436
+ "should_epoch_stop": false,
437
+ "should_evaluate": false,
438
+ "should_log": false,
439
+ "should_save": true,
440
+ "should_training_stop": false
441
+ },
442
+ "attributes": {}
443
+ }
444
+ },
445
+ "total_flos": 0.0,
446
+ "train_batch_size": 8,
447
+ "trial_name": null,
448
+ "trial_params": null
449
+ }
ood-vifactcheck-xlmr-source-source-01-s43-20260901T170609Z-fbd63439/trainer/checkpoint-1710/special_tokens_map.json ADDED
@@ -0,0 +1,15 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "bos_token": "<s>",
3
+ "cls_token": "<s>",
4
+ "eos_token": "</s>",
5
+ "mask_token": {
6
+ "content": "<mask>",
7
+ "lstrip": true,
8
+ "normalized": false,
9
+ "rstrip": false,
10
+ "single_word": false
11
+ },
12
+ "pad_token": "<pad>",
13
+ "sep_token": "</s>",
14
+ "unk_token": "<unk>"
15
+ }
ood-vifactcheck-xlmr-source-source-01-s43-20260901T170609Z-fbd63439/trainer/checkpoint-1710/tokenizer_config.json ADDED
@@ -0,0 +1,55 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "added_tokens_decoder": {
3
+ "0": {
4
+ "content": "<s>",
5
+ "lstrip": false,
6
+ "normalized": false,
7
+ "rstrip": false,
8
+ "single_word": false,
9
+ "special": true
10
+ },
11
+ "1": {
12
+ "content": "<pad>",
13
+ "lstrip": false,
14
+ "normalized": false,
15
+ "rstrip": false,
16
+ "single_word": false,
17
+ "special": true
18
+ },
19
+ "2": {
20
+ "content": "</s>",
21
+ "lstrip": false,
22
+ "normalized": false,
23
+ "rstrip": false,
24
+ "single_word": false,
25
+ "special": true
26
+ },
27
+ "3": {
28
+ "content": "<unk>",
29
+ "lstrip": false,
30
+ "normalized": false,
31
+ "rstrip": false,
32
+ "single_word": false,
33
+ "special": true
34
+ },
35
+ "250001": {
36
+ "content": "<mask>",
37
+ "lstrip": true,
38
+ "normalized": false,
39
+ "rstrip": false,
40
+ "single_word": false,
41
+ "special": true
42
+ }
43
+ },
44
+ "bos_token": "<s>",
45
+ "clean_up_tokenization_spaces": false,
46
+ "cls_token": "<s>",
47
+ "eos_token": "</s>",
48
+ "extra_special_tokens": {},
49
+ "mask_token": "<mask>",
50
+ "model_max_length": 512,
51
+ "pad_token": "<pad>",
52
+ "sep_token": "</s>",
53
+ "tokenizer_class": "XLMRobertaTokenizer",
54
+ "unk_token": "<unk>"
55
+ }
ood-vifactcheck-xlmr-source-source-01-s43-20260901T170609Z-fbd63439/trainer/checkpoint-1710/trainer_state.json ADDED
@@ -0,0 +1,629 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_global_step": 1197,
3
+ "best_metric": 0.8120160172108976,
4
+ "best_model_checkpoint": "/root/VNEncoder/results/checkpoints/ood-vifactcheck-xlmr-source-source-01-s43-20260901T170609Z-fbd63439/trainer/checkpoint-1197",
5
+ "epoch": 10.0,
6
+ "eval_steps": 500,
7
+ "global_step": 1710,
8
+ "is_hyper_param_search": false,
9
+ "is_local_process_zero": true,
10
+ "is_world_process_zero": true,
11
+ "log_history": [
12
+ {
13
+ "epoch": 0.14684287812041116,
14
+ "grad_norm": 19.898252487182617,
15
+ "learning_rate": 1.4035087719298246e-06,
16
+ "loss": 4.8259,
17
+ "step": 25
18
+ },
19
+ {
20
+ "epoch": 0.2936857562408223,
21
+ "grad_norm": 28.20109748840332,
22
+ "learning_rate": 2.865497076023392e-06,
23
+ "loss": 4.6946,
24
+ "step": 50
25
+ },
26
+ {
27
+ "epoch": 0.44052863436123346,
28
+ "grad_norm": 19.910593032836914,
29
+ "learning_rate": 4.3274853801169596e-06,
30
+ "loss": 4.5263,
31
+ "step": 75
32
+ },
33
+ {
34
+ "epoch": 0.5873715124816447,
35
+ "grad_norm": 23.893070220947266,
36
+ "learning_rate": 5.789473684210527e-06,
37
+ "loss": 4.4799,
38
+ "step": 100
39
+ },
40
+ {
41
+ "epoch": 0.7342143906020558,
42
+ "grad_norm": 21.93558120727539,
43
+ "learning_rate": 7.251461988304094e-06,
44
+ "loss": 4.5204,
45
+ "step": 125
46
+ },
47
+ {
48
+ "epoch": 0.8810572687224669,
49
+ "grad_norm": 20.283721923828125,
50
+ "learning_rate": 8.713450292397661e-06,
51
+ "loss": 4.3808,
52
+ "step": 150
53
+ },
54
+ {
55
+ "epoch": 1.0,
56
+ "eval_accuracy": 0.3274021352313167,
57
+ "eval_loss": 1.1216636896133423,
58
+ "eval_macro_f1": 0.1680264699233441,
59
+ "eval_runtime": 1.1499,
60
+ "eval_samples_per_second": 488.741,
61
+ "eval_steps_per_second": 31.307,
62
+ "eval_weighted_f1": 0.16437457594901667,
63
+ "step": 171
64
+ },
65
+ {
66
+ "epoch": 1.0234948604992657,
67
+ "grad_norm": 18.82137680053711,
68
+ "learning_rate": 1.017543859649123e-05,
69
+ "loss": 4.3132,
70
+ "step": 175
71
+ },
72
+ {
73
+ "epoch": 1.170337738619677,
74
+ "grad_norm": 39.67106628417969,
75
+ "learning_rate": 1.1637426900584797e-05,
76
+ "loss": 4.3691,
77
+ "step": 200
78
+ },
79
+ {
80
+ "epoch": 1.3171806167400881,
81
+ "grad_norm": 25.474512100219727,
82
+ "learning_rate": 1.3099415204678362e-05,
83
+ "loss": 4.3744,
84
+ "step": 225
85
+ },
86
+ {
87
+ "epoch": 1.4640234948604993,
88
+ "grad_norm": 29.966018676757812,
89
+ "learning_rate": 1.4561403508771931e-05,
90
+ "loss": 4.1576,
91
+ "step": 250
92
+ },
93
+ {
94
+ "epoch": 1.6108663729809103,
95
+ "grad_norm": 38.67800521850586,
96
+ "learning_rate": 1.60233918128655e-05,
97
+ "loss": 3.9165,
98
+ "step": 275
99
+ },
100
+ {
101
+ "epoch": 1.7577092511013217,
102
+ "grad_norm": 120.15225219726562,
103
+ "learning_rate": 1.7485380116959064e-05,
104
+ "loss": 3.5492,
105
+ "step": 300
106
+ },
107
+ {
108
+ "epoch": 1.9045521292217327,
109
+ "grad_norm": 145.66622924804688,
110
+ "learning_rate": 1.894736842105263e-05,
111
+ "loss": 3.2797,
112
+ "step": 325
113
+ },
114
+ {
115
+ "epoch": 2.0,
116
+ "eval_accuracy": 0.6814946619217082,
117
+ "eval_loss": 0.8266801238059998,
118
+ "eval_macro_f1": 0.6695599664073812,
119
+ "eval_runtime": 1.0467,
120
+ "eval_samples_per_second": 536.94,
121
+ "eval_steps_per_second": 34.395,
122
+ "eval_weighted_f1": 0.6707705283670244,
123
+ "step": 342
124
+ },
125
+ {
126
+ "epoch": 2.0469897209985315,
127
+ "grad_norm": 61.46903610229492,
128
+ "learning_rate": 1.99545159194282e-05,
129
+ "loss": 3.1801,
130
+ "step": 350
131
+ },
132
+ {
133
+ "epoch": 2.193832599118943,
134
+ "grad_norm": 94.04988861083984,
135
+ "learning_rate": 1.9792072774528916e-05,
136
+ "loss": 2.9369,
137
+ "step": 375
138
+ },
139
+ {
140
+ "epoch": 2.340675477239354,
141
+ "grad_norm": 131.471923828125,
142
+ "learning_rate": 1.962962962962963e-05,
143
+ "loss": 2.8236,
144
+ "step": 400
145
+ },
146
+ {
147
+ "epoch": 2.4875183553597653,
148
+ "grad_norm": 176.6563262939453,
149
+ "learning_rate": 1.9467186484730346e-05,
150
+ "loss": 2.9867,
151
+ "step": 425
152
+ },
153
+ {
154
+ "epoch": 2.6343612334801763,
155
+ "grad_norm": 59.89003372192383,
156
+ "learning_rate": 1.930474333983106e-05,
157
+ "loss": 2.9233,
158
+ "step": 450
159
+ },
160
+ {
161
+ "epoch": 2.7812041116005872,
162
+ "grad_norm": 68.92916870117188,
163
+ "learning_rate": 1.9142300194931776e-05,
164
+ "loss": 2.8271,
165
+ "step": 475
166
+ },
167
+ {
168
+ "epoch": 2.9280469897209986,
169
+ "grad_norm": 103.18282318115234,
170
+ "learning_rate": 1.897985705003249e-05,
171
+ "loss": 2.7263,
172
+ "step": 500
173
+ },
174
+ {
175
+ "epoch": 3.0,
176
+ "eval_accuracy": 0.7686832740213523,
177
+ "eval_loss": 0.6141501069068909,
178
+ "eval_macro_f1": 0.7704908189767045,
179
+ "eval_runtime": 1.1127,
180
+ "eval_samples_per_second": 505.085,
181
+ "eval_steps_per_second": 32.354,
182
+ "eval_weighted_f1": 0.7697335190679598,
183
+ "step": 513
184
+ },
185
+ {
186
+ "epoch": 3.0704845814977975,
187
+ "grad_norm": 106.83403778076172,
188
+ "learning_rate": 1.8817413905133203e-05,
189
+ "loss": 2.5354,
190
+ "step": 525
191
+ },
192
+ {
193
+ "epoch": 3.2173274596182084,
194
+ "grad_norm": 103.26848602294922,
195
+ "learning_rate": 1.8654970760233918e-05,
196
+ "loss": 2.4149,
197
+ "step": 550
198
+ },
199
+ {
200
+ "epoch": 3.36417033773862,
201
+ "grad_norm": 84.16307067871094,
202
+ "learning_rate": 1.8492527615334633e-05,
203
+ "loss": 2.1616,
204
+ "step": 575
205
+ },
206
+ {
207
+ "epoch": 3.511013215859031,
208
+ "grad_norm": 51.95928955078125,
209
+ "learning_rate": 1.833008447043535e-05,
210
+ "loss": 2.2616,
211
+ "step": 600
212
+ },
213
+ {
214
+ "epoch": 3.657856093979442,
215
+ "grad_norm": 48.234153747558594,
216
+ "learning_rate": 1.8167641325536064e-05,
217
+ "loss": 1.916,
218
+ "step": 625
219
+ },
220
+ {
221
+ "epoch": 3.804698972099853,
222
+ "grad_norm": 97.79098510742188,
223
+ "learning_rate": 1.800519818063678e-05,
224
+ "loss": 2.1566,
225
+ "step": 650
226
+ },
227
+ {
228
+ "epoch": 3.951541850220264,
229
+ "grad_norm": 83.2856674194336,
230
+ "learning_rate": 1.7842755035737494e-05,
231
+ "loss": 2.0654,
232
+ "step": 675
233
+ },
234
+ {
235
+ "epoch": 4.0,
236
+ "eval_accuracy": 0.791814946619217,
237
+ "eval_loss": 0.5504909753799438,
238
+ "eval_macro_f1": 0.7894521890156985,
239
+ "eval_runtime": 1.1111,
240
+ "eval_samples_per_second": 505.812,
241
+ "eval_steps_per_second": 32.401,
242
+ "eval_weighted_f1": 0.7894752915277294,
243
+ "step": 684
244
+ },
245
+ {
246
+ "epoch": 4.093979441997063,
247
+ "grad_norm": 51.651371002197266,
248
+ "learning_rate": 1.768031189083821e-05,
249
+ "loss": 1.7594,
250
+ "step": 700
251
+ },
252
+ {
253
+ "epoch": 4.240822320117474,
254
+ "grad_norm": 96.34539794921875,
255
+ "learning_rate": 1.751786874593892e-05,
256
+ "loss": 1.7873,
257
+ "step": 725
258
+ },
259
+ {
260
+ "epoch": 4.387665198237886,
261
+ "grad_norm": 113.27198791503906,
262
+ "learning_rate": 1.7355425601039636e-05,
263
+ "loss": 1.6555,
264
+ "step": 750
265
+ },
266
+ {
267
+ "epoch": 4.534508076358296,
268
+ "grad_norm": 52.46745681762695,
269
+ "learning_rate": 1.719298245614035e-05,
270
+ "loss": 1.5451,
271
+ "step": 775
272
+ },
273
+ {
274
+ "epoch": 4.681350954478708,
275
+ "grad_norm": 81.5442886352539,
276
+ "learning_rate": 1.7030539311241066e-05,
277
+ "loss": 1.5964,
278
+ "step": 800
279
+ },
280
+ {
281
+ "epoch": 4.828193832599119,
282
+ "grad_norm": 155.65516662597656,
283
+ "learning_rate": 1.686809616634178e-05,
284
+ "loss": 1.7224,
285
+ "step": 825
286
+ },
287
+ {
288
+ "epoch": 4.975036710719531,
289
+ "grad_norm": 48.24635314941406,
290
+ "learning_rate": 1.6705653021442497e-05,
291
+ "loss": 1.7126,
292
+ "step": 850
293
+ },
294
+ {
295
+ "epoch": 5.0,
296
+ "eval_accuracy": 0.8078291814946619,
297
+ "eval_loss": 0.5076857805252075,
298
+ "eval_macro_f1": 0.8081435252166961,
299
+ "eval_runtime": 1.0205,
300
+ "eval_samples_per_second": 550.698,
301
+ "eval_steps_per_second": 35.276,
302
+ "eval_weighted_f1": 0.8077148227916391,
303
+ "step": 855
304
+ },
305
+ {
306
+ "epoch": 5.117474302496329,
307
+ "grad_norm": 90.79500579833984,
308
+ "learning_rate": 1.654320987654321e-05,
309
+ "loss": 1.4157,
310
+ "step": 875
311
+ },
312
+ {
313
+ "epoch": 5.26431718061674,
314
+ "grad_norm": 74.45726013183594,
315
+ "learning_rate": 1.6380766731643927e-05,
316
+ "loss": 1.3348,
317
+ "step": 900
318
+ },
319
+ {
320
+ "epoch": 5.411160058737151,
321
+ "grad_norm": 72.0616455078125,
322
+ "learning_rate": 1.621832358674464e-05,
323
+ "loss": 1.1797,
324
+ "step": 925
325
+ },
326
+ {
327
+ "epoch": 5.558002936857562,
328
+ "grad_norm": 83.30821228027344,
329
+ "learning_rate": 1.6055880441845354e-05,
330
+ "loss": 1.0697,
331
+ "step": 950
332
+ },
333
+ {
334
+ "epoch": 5.704845814977974,
335
+ "grad_norm": 72.64716339111328,
336
+ "learning_rate": 1.589343729694607e-05,
337
+ "loss": 1.3893,
338
+ "step": 975
339
+ },
340
+ {
341
+ "epoch": 5.851688693098385,
342
+ "grad_norm": 99.64823150634766,
343
+ "learning_rate": 1.5730994152046787e-05,
344
+ "loss": 1.3584,
345
+ "step": 1000
346
+ },
347
+ {
348
+ "epoch": 5.998531571218796,
349
+ "grad_norm": 97.09939575195312,
350
+ "learning_rate": 1.55685510071475e-05,
351
+ "loss": 1.5412,
352
+ "step": 1025
353
+ },
354
+ {
355
+ "epoch": 6.0,
356
+ "eval_accuracy": 0.8078291814946619,
357
+ "eval_loss": 0.5981741547584534,
358
+ "eval_macro_f1": 0.8021303737721648,
359
+ "eval_runtime": 1.1109,
360
+ "eval_samples_per_second": 505.911,
361
+ "eval_steps_per_second": 32.407,
362
+ "eval_weighted_f1": 0.802081094663556,
363
+ "step": 1026
364
+ },
365
+ {
366
+ "epoch": 6.140969162995595,
367
+ "grad_norm": 95.87456512451172,
368
+ "learning_rate": 1.5406107862248214e-05,
369
+ "loss": 1.0143,
370
+ "step": 1050
371
+ },
372
+ {
373
+ "epoch": 6.2878120411160054,
374
+ "grad_norm": 68.2560806274414,
375
+ "learning_rate": 1.524366471734893e-05,
376
+ "loss": 0.8491,
377
+ "step": 1075
378
+ },
379
+ {
380
+ "epoch": 6.434654919236417,
381
+ "grad_norm": 77.93789672851562,
382
+ "learning_rate": 1.5081221572449645e-05,
383
+ "loss": 1.0485,
384
+ "step": 1100
385
+ },
386
+ {
387
+ "epoch": 6.581497797356828,
388
+ "grad_norm": 16.324548721313477,
389
+ "learning_rate": 1.4918778427550358e-05,
390
+ "loss": 0.935,
391
+ "step": 1125
392
+ },
393
+ {
394
+ "epoch": 6.72834067547724,
395
+ "grad_norm": 65.43302917480469,
396
+ "learning_rate": 1.4756335282651073e-05,
397
+ "loss": 1.0821,
398
+ "step": 1150
399
+ },
400
+ {
401
+ "epoch": 6.87518355359765,
402
+ "grad_norm": 124.80599975585938,
403
+ "learning_rate": 1.4593892137751788e-05,
404
+ "loss": 0.9082,
405
+ "step": 1175
406
+ },
407
+ {
408
+ "epoch": 7.0,
409
+ "eval_accuracy": 0.8113879003558719,
410
+ "eval_loss": 0.6220943331718445,
411
+ "eval_macro_f1": 0.8120160172108976,
412
+ "eval_runtime": 1.0895,
413
+ "eval_samples_per_second": 515.83,
414
+ "eval_steps_per_second": 33.042,
415
+ "eval_weighted_f1": 0.8116930414813697,
416
+ "step": 1197
417
+ },
418
+ {
419
+ "epoch": 7.0176211453744495,
420
+ "grad_norm": 71.75617980957031,
421
+ "learning_rate": 1.4431448992852503e-05,
422
+ "loss": 1.1683,
423
+ "step": 1200
424
+ },
425
+ {
426
+ "epoch": 7.164464023494861,
427
+ "grad_norm": 99.44718170166016,
428
+ "learning_rate": 1.4269005847953217e-05,
429
+ "loss": 0.7722,
430
+ "step": 1225
431
+ },
432
+ {
433
+ "epoch": 7.311306901615271,
434
+ "grad_norm": 85.54470825195312,
435
+ "learning_rate": 1.4106562703053932e-05,
436
+ "loss": 0.7766,
437
+ "step": 1250
438
+ },
439
+ {
440
+ "epoch": 7.458149779735683,
441
+ "grad_norm": 21.999500274658203,
442
+ "learning_rate": 1.3944119558154647e-05,
443
+ "loss": 0.6444,
444
+ "step": 1275
445
+ },
446
+ {
447
+ "epoch": 7.604992657856094,
448
+ "grad_norm": 89.1247787475586,
449
+ "learning_rate": 1.3781676413255362e-05,
450
+ "loss": 0.6769,
451
+ "step": 1300
452
+ },
453
+ {
454
+ "epoch": 7.751835535976505,
455
+ "grad_norm": 89.86659240722656,
456
+ "learning_rate": 1.3619233268356076e-05,
457
+ "loss": 1.0367,
458
+ "step": 1325
459
+ },
460
+ {
461
+ "epoch": 7.898678414096916,
462
+ "grad_norm": 135.19757080078125,
463
+ "learning_rate": 1.345679012345679e-05,
464
+ "loss": 0.5959,
465
+ "step": 1350
466
+ },
467
+ {
468
+ "epoch": 8.0,
469
+ "eval_accuracy": 0.7882562277580071,
470
+ "eval_loss": 0.7940157055854797,
471
+ "eval_macro_f1": 0.7912761397548417,
472
+ "eval_runtime": 1.0699,
473
+ "eval_samples_per_second": 525.305,
474
+ "eval_steps_per_second": 33.649,
475
+ "eval_weighted_f1": 0.79101162234856,
476
+ "step": 1368
477
+ },
478
+ {
479
+ "epoch": 8.041116005873715,
480
+ "grad_norm": 123.22237396240234,
481
+ "learning_rate": 1.3294346978557506e-05,
482
+ "loss": 0.8356,
483
+ "step": 1375
484
+ },
485
+ {
486
+ "epoch": 8.187958883994126,
487
+ "grad_norm": 68.84619903564453,
488
+ "learning_rate": 1.3131903833658221e-05,
489
+ "loss": 0.4695,
490
+ "step": 1400
491
+ },
492
+ {
493
+ "epoch": 8.334801762114537,
494
+ "grad_norm": 154.56805419921875,
495
+ "learning_rate": 1.2969460688758935e-05,
496
+ "loss": 0.6034,
497
+ "step": 1425
498
+ },
499
+ {
500
+ "epoch": 8.481644640234949,
501
+ "grad_norm": 86.3011245727539,
502
+ "learning_rate": 1.280701754385965e-05,
503
+ "loss": 0.5087,
504
+ "step": 1450
505
+ },
506
+ {
507
+ "epoch": 8.62848751835536,
508
+ "grad_norm": 24.14583969116211,
509
+ "learning_rate": 1.2644574398960365e-05,
510
+ "loss": 0.422,
511
+ "step": 1475
512
+ },
513
+ {
514
+ "epoch": 8.775330396475772,
515
+ "grad_norm": 310.94207763671875,
516
+ "learning_rate": 1.248213125406108e-05,
517
+ "loss": 0.6256,
518
+ "step": 1500
519
+ },
520
+ {
521
+ "epoch": 8.922173274596181,
522
+ "grad_norm": 139.9342041015625,
523
+ "learning_rate": 1.2319688109161793e-05,
524
+ "loss": 0.6883,
525
+ "step": 1525
526
+ },
527
+ {
528
+ "epoch": 9.0,
529
+ "eval_accuracy": 0.802491103202847,
530
+ "eval_loss": 0.9116710424423218,
531
+ "eval_macro_f1": 0.8018396686159844,
532
+ "eval_runtime": 1.0243,
533
+ "eval_samples_per_second": 548.666,
534
+ "eval_steps_per_second": 35.146,
535
+ "eval_weighted_f1": 0.8012830768350294,
536
+ "step": 1539
537
+ },
538
+ {
539
+ "epoch": 9.06461086637298,
540
+ "grad_norm": 46.106658935546875,
541
+ "learning_rate": 1.2157244964262509e-05,
542
+ "loss": 0.5617,
543
+ "step": 1550
544
+ },
545
+ {
546
+ "epoch": 9.211453744493392,
547
+ "grad_norm": 26.815168380737305,
548
+ "learning_rate": 1.1994801819363224e-05,
549
+ "loss": 0.4145,
550
+ "step": 1575
551
+ },
552
+ {
553
+ "epoch": 9.358296622613803,
554
+ "grad_norm": 113.18115997314453,
555
+ "learning_rate": 1.1832358674463939e-05,
556
+ "loss": 0.6228,
557
+ "step": 1600
558
+ },
559
+ {
560
+ "epoch": 9.505139500734215,
561
+ "grad_norm": 15.208633422851562,
562
+ "learning_rate": 1.1669915529564652e-05,
563
+ "loss": 0.4904,
564
+ "step": 1625
565
+ },
566
+ {
567
+ "epoch": 9.651982378854626,
568
+ "grad_norm": 120.86727142333984,
569
+ "learning_rate": 1.1507472384665367e-05,
570
+ "loss": 0.3946,
571
+ "step": 1650
572
+ },
573
+ {
574
+ "epoch": 9.798825256975036,
575
+ "grad_norm": 340.96685791015625,
576
+ "learning_rate": 1.1345029239766083e-05,
577
+ "loss": 0.6081,
578
+ "step": 1675
579
+ },
580
+ {
581
+ "epoch": 9.945668135095447,
582
+ "grad_norm": 213.39102172851562,
583
+ "learning_rate": 1.11825860948668e-05,
584
+ "loss": 0.3996,
585
+ "step": 1700
586
+ },
587
+ {
588
+ "epoch": 10.0,
589
+ "eval_accuracy": 0.8096085409252669,
590
+ "eval_loss": 1.0028072595596313,
591
+ "eval_macro_f1": 0.8115669703691483,
592
+ "eval_runtime": 1.128,
593
+ "eval_samples_per_second": 498.24,
594
+ "eval_steps_per_second": 31.916,
595
+ "eval_weighted_f1": 0.8112526763459069,
596
+ "step": 1710
597
+ }
598
+ ],
599
+ "logging_steps": 25,
600
+ "max_steps": 3420,
601
+ "num_input_tokens_seen": 0,
602
+ "num_train_epochs": 20,
603
+ "save_steps": 500,
604
+ "stateful_callbacks": {
605
+ "EarlyStoppingCallback": {
606
+ "args": {
607
+ "early_stopping_patience": 3,
608
+ "early_stopping_threshold": 0.0
609
+ },
610
+ "attributes": {
611
+ "early_stopping_patience_counter": 3
612
+ }
613
+ },
614
+ "TrainerControl": {
615
+ "args": {
616
+ "should_epoch_stop": false,
617
+ "should_evaluate": false,
618
+ "should_log": false,
619
+ "should_save": true,
620
+ "should_training_stop": true
621
+ },
622
+ "attributes": {}
623
+ }
624
+ },
625
+ "total_flos": 0.0,
626
+ "train_batch_size": 8,
627
+ "trial_name": null,
628
+ "trial_params": null
629
+ }
ood-vifactcheck-xlmr-source-source-02-s42-20260901T153352Z-27bda108/trainer/checkpoint-1448/special_tokens_map.json ADDED
@@ -0,0 +1,15 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "bos_token": "<s>",
3
+ "cls_token": "<s>",
4
+ "eos_token": "</s>",
5
+ "mask_token": {
6
+ "content": "<mask>",
7
+ "lstrip": true,
8
+ "normalized": false,
9
+ "rstrip": false,
10
+ "single_word": false
11
+ },
12
+ "pad_token": "<pad>",
13
+ "sep_token": "</s>",
14
+ "unk_token": "<unk>"
15
+ }
ood-vifactcheck-xlmr-source-source-02-s42-20260901T153352Z-27bda108/trainer/checkpoint-1448/tokenizer_config.json ADDED
@@ -0,0 +1,55 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "added_tokens_decoder": {
3
+ "0": {
4
+ "content": "<s>",
5
+ "lstrip": false,
6
+ "normalized": false,
7
+ "rstrip": false,
8
+ "single_word": false,
9
+ "special": true
10
+ },
11
+ "1": {
12
+ "content": "<pad>",
13
+ "lstrip": false,
14
+ "normalized": false,
15
+ "rstrip": false,
16
+ "single_word": false,
17
+ "special": true
18
+ },
19
+ "2": {
20
+ "content": "</s>",
21
+ "lstrip": false,
22
+ "normalized": false,
23
+ "rstrip": false,
24
+ "single_word": false,
25
+ "special": true
26
+ },
27
+ "3": {
28
+ "content": "<unk>",
29
+ "lstrip": false,
30
+ "normalized": false,
31
+ "rstrip": false,
32
+ "single_word": false,
33
+ "special": true
34
+ },
35
+ "250001": {
36
+ "content": "<mask>",
37
+ "lstrip": true,
38
+ "normalized": false,
39
+ "rstrip": false,
40
+ "single_word": false,
41
+ "special": true
42
+ }
43
+ },
44
+ "bos_token": "<s>",
45
+ "clean_up_tokenization_spaces": false,
46
+ "cls_token": "<s>",
47
+ "eos_token": "</s>",
48
+ "extra_special_tokens": {},
49
+ "mask_token": "<mask>",
50
+ "model_max_length": 512,
51
+ "pad_token": "<pad>",
52
+ "sep_token": "</s>",
53
+ "tokenizer_class": "XLMRobertaTokenizer",
54
+ "unk_token": "<unk>"
55
+ }
ood-vifactcheck-xlmr-source-source-02-s42-20260901T153352Z-27bda108/trainer/checkpoint-1448/trainer_state.json ADDED
@@ -0,0 +1,530 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_global_step": 905,
3
+ "best_metric": 0.8079147423599644,
4
+ "best_model_checkpoint": "/root/VNEncoder/results/checkpoints/ood-vifactcheck-xlmr-source-source-02-s42-20260901T153352Z-27bda108/trainer/checkpoint-905",
5
+ "epoch": 8.0,
6
+ "eval_steps": 500,
7
+ "global_step": 1448,
8
+ "is_hyper_param_search": false,
9
+ "is_local_process_zero": true,
10
+ "is_world_process_zero": true,
11
+ "log_history": [
12
+ {
13
+ "epoch": 0.13812154696132597,
14
+ "grad_norm": 12.728955268859863,
15
+ "learning_rate": 1.3259668508287293e-06,
16
+ "loss": 4.8048,
17
+ "step": 25
18
+ },
19
+ {
20
+ "epoch": 0.27624309392265195,
21
+ "grad_norm": 56.0223274230957,
22
+ "learning_rate": 2.707182320441989e-06,
23
+ "loss": 4.711,
24
+ "step": 50
25
+ },
26
+ {
27
+ "epoch": 0.4143646408839779,
28
+ "grad_norm": 39.932518005371094,
29
+ "learning_rate": 4.088397790055249e-06,
30
+ "loss": 4.5371,
31
+ "step": 75
32
+ },
33
+ {
34
+ "epoch": 0.5524861878453039,
35
+ "grad_norm": 42.93894577026367,
36
+ "learning_rate": 5.469613259668509e-06,
37
+ "loss": 4.678,
38
+ "step": 100
39
+ },
40
+ {
41
+ "epoch": 0.6906077348066298,
42
+ "grad_norm": 37.85771179199219,
43
+ "learning_rate": 6.850828729281769e-06,
44
+ "loss": 4.662,
45
+ "step": 125
46
+ },
47
+ {
48
+ "epoch": 0.8287292817679558,
49
+ "grad_norm": 48.18323516845703,
50
+ "learning_rate": 8.232044198895029e-06,
51
+ "loss": 4.4032,
52
+ "step": 150
53
+ },
54
+ {
55
+ "epoch": 0.9668508287292817,
56
+ "grad_norm": 99.214599609375,
57
+ "learning_rate": 9.613259668508288e-06,
58
+ "loss": 4.2769,
59
+ "step": 175
60
+ },
61
+ {
62
+ "epoch": 1.0,
63
+ "eval_accuracy": 0.5567010309278351,
64
+ "eval_loss": 0.9564104080200195,
65
+ "eval_macro_f1": 0.531338415603882,
66
+ "eval_runtime": 1.078,
67
+ "eval_samples_per_second": 539.897,
68
+ "eval_steps_per_second": 34.323,
69
+ "eval_weighted_f1": 0.5352146874056312,
70
+ "step": 181
71
+ },
72
+ {
73
+ "epoch": 1.1049723756906078,
74
+ "grad_norm": 68.16238403320312,
75
+ "learning_rate": 1.0994475138121547e-05,
76
+ "loss": 4.0766,
77
+ "step": 200
78
+ },
79
+ {
80
+ "epoch": 1.2430939226519337,
81
+ "grad_norm": 36.474945068359375,
82
+ "learning_rate": 1.2375690607734806e-05,
83
+ "loss": 4.0206,
84
+ "step": 225
85
+ },
86
+ {
87
+ "epoch": 1.3812154696132597,
88
+ "grad_norm": 41.88302230834961,
89
+ "learning_rate": 1.3756906077348067e-05,
90
+ "loss": 4.0765,
91
+ "step": 250
92
+ },
93
+ {
94
+ "epoch": 1.5193370165745856,
95
+ "grad_norm": 73.10688781738281,
96
+ "learning_rate": 1.5138121546961326e-05,
97
+ "loss": 3.8742,
98
+ "step": 275
99
+ },
100
+ {
101
+ "epoch": 1.6574585635359116,
102
+ "grad_norm": 57.27144241333008,
103
+ "learning_rate": 1.6519337016574586e-05,
104
+ "loss": 3.7189,
105
+ "step": 300
106
+ },
107
+ {
108
+ "epoch": 1.7955801104972375,
109
+ "grad_norm": 80.90764617919922,
110
+ "learning_rate": 1.7900552486187846e-05,
111
+ "loss": 3.3838,
112
+ "step": 325
113
+ },
114
+ {
115
+ "epoch": 1.9337016574585635,
116
+ "grad_norm": 100.98841094970703,
117
+ "learning_rate": 1.9281767955801107e-05,
118
+ "loss": 3.2471,
119
+ "step": 350
120
+ },
121
+ {
122
+ "epoch": 2.0,
123
+ "eval_accuracy": 0.7096219931271478,
124
+ "eval_loss": 0.7144930362701416,
125
+ "eval_macro_f1": 0.708526703261887,
126
+ "eval_runtime": 1.087,
127
+ "eval_samples_per_second": 535.442,
128
+ "eval_steps_per_second": 34.04,
129
+ "eval_weighted_f1": 0.7085504083724671,
130
+ "step": 362
131
+ },
132
+ {
133
+ "epoch": 2.0718232044198897,
134
+ "grad_norm": 61.62918472290039,
135
+ "learning_rate": 1.9926335174953962e-05,
136
+ "loss": 2.9424,
137
+ "step": 375
138
+ },
139
+ {
140
+ "epoch": 2.2099447513812156,
141
+ "grad_norm": 66.43571472167969,
142
+ "learning_rate": 1.9772866789441377e-05,
143
+ "loss": 2.7574,
144
+ "step": 400
145
+ },
146
+ {
147
+ "epoch": 2.3480662983425415,
148
+ "grad_norm": 124.33316802978516,
149
+ "learning_rate": 1.9619398403928794e-05,
150
+ "loss": 2.8673,
151
+ "step": 425
152
+ },
153
+ {
154
+ "epoch": 2.4861878453038675,
155
+ "grad_norm": 52.74990463256836,
156
+ "learning_rate": 1.946593001841621e-05,
157
+ "loss": 2.7332,
158
+ "step": 450
159
+ },
160
+ {
161
+ "epoch": 2.6243093922651934,
162
+ "grad_norm": 120.99431610107422,
163
+ "learning_rate": 1.9312461632903623e-05,
164
+ "loss": 2.5593,
165
+ "step": 475
166
+ },
167
+ {
168
+ "epoch": 2.7624309392265194,
169
+ "grad_norm": 85.82913970947266,
170
+ "learning_rate": 1.915899324739104e-05,
171
+ "loss": 2.4126,
172
+ "step": 500
173
+ },
174
+ {
175
+ "epoch": 2.9005524861878453,
176
+ "grad_norm": 105.4771499633789,
177
+ "learning_rate": 1.9005524861878455e-05,
178
+ "loss": 2.4222,
179
+ "step": 525
180
+ },
181
+ {
182
+ "epoch": 3.0,
183
+ "eval_accuracy": 0.7663230240549829,
184
+ "eval_loss": 0.5739344358444214,
185
+ "eval_macro_f1": 0.7566652570915214,
186
+ "eval_runtime": 1.0416,
187
+ "eval_samples_per_second": 558.739,
188
+ "eval_steps_per_second": 35.521,
189
+ "eval_weighted_f1": 0.758295393308871,
190
+ "step": 543
191
+ },
192
+ {
193
+ "epoch": 3.0386740331491713,
194
+ "grad_norm": 104.53805541992188,
195
+ "learning_rate": 1.885205647636587e-05,
196
+ "loss": 2.2542,
197
+ "step": 550
198
+ },
199
+ {
200
+ "epoch": 3.1767955801104972,
201
+ "grad_norm": 119.23262786865234,
202
+ "learning_rate": 1.8698588090853287e-05,
203
+ "loss": 2.2797,
204
+ "step": 575
205
+ },
206
+ {
207
+ "epoch": 3.314917127071823,
208
+ "grad_norm": 67.51237487792969,
209
+ "learning_rate": 1.85451197053407e-05,
210
+ "loss": 1.8071,
211
+ "step": 600
212
+ },
213
+ {
214
+ "epoch": 3.453038674033149,
215
+ "grad_norm": 48.214908599853516,
216
+ "learning_rate": 1.8391651319828116e-05,
217
+ "loss": 1.9983,
218
+ "step": 625
219
+ },
220
+ {
221
+ "epoch": 3.591160220994475,
222
+ "grad_norm": 67.250732421875,
223
+ "learning_rate": 1.823818293431553e-05,
224
+ "loss": 1.8777,
225
+ "step": 650
226
+ },
227
+ {
228
+ "epoch": 3.729281767955801,
229
+ "grad_norm": 115.72201538085938,
230
+ "learning_rate": 1.8084714548802948e-05,
231
+ "loss": 1.9912,
232
+ "step": 675
233
+ },
234
+ {
235
+ "epoch": 3.867403314917127,
236
+ "grad_norm": 47.928062438964844,
237
+ "learning_rate": 1.7931246163290362e-05,
238
+ "loss": 1.9777,
239
+ "step": 700
240
+ },
241
+ {
242
+ "epoch": 4.0,
243
+ "eval_accuracy": 0.788659793814433,
244
+ "eval_loss": 0.5242534875869751,
245
+ "eval_macro_f1": 0.787816514118535,
246
+ "eval_runtime": 1.1404,
247
+ "eval_samples_per_second": 510.334,
248
+ "eval_steps_per_second": 32.444,
249
+ "eval_weighted_f1": 0.7883303300363687,
250
+ "step": 724
251
+ },
252
+ {
253
+ "epoch": 4.005524861878453,
254
+ "grad_norm": 58.05070495605469,
255
+ "learning_rate": 1.7777777777777777e-05,
256
+ "loss": 1.9077,
257
+ "step": 725
258
+ },
259
+ {
260
+ "epoch": 4.143646408839779,
261
+ "grad_norm": 108.5123062133789,
262
+ "learning_rate": 1.7624309392265194e-05,
263
+ "loss": 1.3115,
264
+ "step": 750
265
+ },
266
+ {
267
+ "epoch": 4.281767955801105,
268
+ "grad_norm": 139.55747985839844,
269
+ "learning_rate": 1.747084100675261e-05,
270
+ "loss": 1.5837,
271
+ "step": 775
272
+ },
273
+ {
274
+ "epoch": 4.419889502762431,
275
+ "grad_norm": 84.59967803955078,
276
+ "learning_rate": 1.7317372621240026e-05,
277
+ "loss": 1.5453,
278
+ "step": 800
279
+ },
280
+ {
281
+ "epoch": 4.558011049723757,
282
+ "grad_norm": 60.05119323730469,
283
+ "learning_rate": 1.716390423572744e-05,
284
+ "loss": 1.7393,
285
+ "step": 825
286
+ },
287
+ {
288
+ "epoch": 4.696132596685083,
289
+ "grad_norm": 85.76810455322266,
290
+ "learning_rate": 1.701043585021486e-05,
291
+ "loss": 1.5069,
292
+ "step": 850
293
+ },
294
+ {
295
+ "epoch": 4.834254143646409,
296
+ "grad_norm": 140.66427612304688,
297
+ "learning_rate": 1.6856967464702273e-05,
298
+ "loss": 1.6284,
299
+ "step": 875
300
+ },
301
+ {
302
+ "epoch": 4.972375690607735,
303
+ "grad_norm": 116.09174346923828,
304
+ "learning_rate": 1.670349907918969e-05,
305
+ "loss": 1.4866,
306
+ "step": 900
307
+ },
308
+ {
309
+ "epoch": 5.0,
310
+ "eval_accuracy": 0.8092783505154639,
311
+ "eval_loss": 0.5569475889205933,
312
+ "eval_macro_f1": 0.8079147423599644,
313
+ "eval_runtime": 1.1105,
314
+ "eval_samples_per_second": 524.085,
315
+ "eval_steps_per_second": 33.318,
316
+ "eval_weighted_f1": 0.8088169111415616,
317
+ "step": 905
318
+ },
319
+ {
320
+ "epoch": 5.110497237569061,
321
+ "grad_norm": 75.29962921142578,
322
+ "learning_rate": 1.6550030693677105e-05,
323
+ "loss": 1.174,
324
+ "step": 925
325
+ },
326
+ {
327
+ "epoch": 5.248618784530387,
328
+ "grad_norm": 86.1469497680664,
329
+ "learning_rate": 1.639656230816452e-05,
330
+ "loss": 1.1958,
331
+ "step": 950
332
+ },
333
+ {
334
+ "epoch": 5.386740331491713,
335
+ "grad_norm": 151.677978515625,
336
+ "learning_rate": 1.6243093922651937e-05,
337
+ "loss": 1.0569,
338
+ "step": 975
339
+ },
340
+ {
341
+ "epoch": 5.524861878453039,
342
+ "grad_norm": 50.731685638427734,
343
+ "learning_rate": 1.608962553713935e-05,
344
+ "loss": 1.1108,
345
+ "step": 1000
346
+ },
347
+ {
348
+ "epoch": 5.662983425414365,
349
+ "grad_norm": 33.633480072021484,
350
+ "learning_rate": 1.5936157151626766e-05,
351
+ "loss": 1.013,
352
+ "step": 1025
353
+ },
354
+ {
355
+ "epoch": 5.801104972375691,
356
+ "grad_norm": 105.0028076171875,
357
+ "learning_rate": 1.5782688766114183e-05,
358
+ "loss": 1.2053,
359
+ "step": 1050
360
+ },
361
+ {
362
+ "epoch": 5.939226519337017,
363
+ "grad_norm": 95.74874114990234,
364
+ "learning_rate": 1.5629220380601598e-05,
365
+ "loss": 1.1072,
366
+ "step": 1075
367
+ },
368
+ {
369
+ "epoch": 6.0,
370
+ "eval_accuracy": 0.802405498281787,
371
+ "eval_loss": 0.6753490567207336,
372
+ "eval_macro_f1": 0.8016931014910001,
373
+ "eval_runtime": 1.1063,
374
+ "eval_samples_per_second": 526.071,
375
+ "eval_steps_per_second": 33.444,
376
+ "eval_weighted_f1": 0.8022709706811875,
377
+ "step": 1086
378
+ },
379
+ {
380
+ "epoch": 6.077348066298343,
381
+ "grad_norm": 141.87071228027344,
382
+ "learning_rate": 1.5475751995089012e-05,
383
+ "loss": 1.1294,
384
+ "step": 1100
385
+ },
386
+ {
387
+ "epoch": 6.2154696132596685,
388
+ "grad_norm": 55.73676300048828,
389
+ "learning_rate": 1.532228360957643e-05,
390
+ "loss": 0.9051,
391
+ "step": 1125
392
+ },
393
+ {
394
+ "epoch": 6.3535911602209945,
395
+ "grad_norm": 53.82463836669922,
396
+ "learning_rate": 1.5168815224063844e-05,
397
+ "loss": 0.9204,
398
+ "step": 1150
399
+ },
400
+ {
401
+ "epoch": 6.49171270718232,
402
+ "grad_norm": 113.02780151367188,
403
+ "learning_rate": 1.5015346838551258e-05,
404
+ "loss": 0.8417,
405
+ "step": 1175
406
+ },
407
+ {
408
+ "epoch": 6.629834254143646,
409
+ "grad_norm": 96.89856719970703,
410
+ "learning_rate": 1.4861878453038676e-05,
411
+ "loss": 0.7389,
412
+ "step": 1200
413
+ },
414
+ {
415
+ "epoch": 6.767955801104972,
416
+ "grad_norm": 66.34088134765625,
417
+ "learning_rate": 1.470841006752609e-05,
418
+ "loss": 0.8814,
419
+ "step": 1225
420
+ },
421
+ {
422
+ "epoch": 6.906077348066298,
423
+ "grad_norm": 86.4268569946289,
424
+ "learning_rate": 1.4554941682013506e-05,
425
+ "loss": 0.7623,
426
+ "step": 1250
427
+ },
428
+ {
429
+ "epoch": 7.0,
430
+ "eval_accuracy": 0.7938144329896907,
431
+ "eval_loss": 0.7672864198684692,
432
+ "eval_macro_f1": 0.791636724497827,
433
+ "eval_runtime": 1.2192,
434
+ "eval_samples_per_second": 477.353,
435
+ "eval_steps_per_second": 30.347,
436
+ "eval_weighted_f1": 0.7921651883437386,
437
+ "step": 1267
438
+ },
439
+ {
440
+ "epoch": 7.044198895027624,
441
+ "grad_norm": 10.062307357788086,
442
+ "learning_rate": 1.4401473296500922e-05,
443
+ "loss": 1.2103,
444
+ "step": 1275
445
+ },
446
+ {
447
+ "epoch": 7.18232044198895,
448
+ "grad_norm": 176.4008331298828,
449
+ "learning_rate": 1.4248004910988338e-05,
450
+ "loss": 0.7116,
451
+ "step": 1300
452
+ },
453
+ {
454
+ "epoch": 7.320441988950276,
455
+ "grad_norm": 62.07518005371094,
456
+ "learning_rate": 1.4094536525475753e-05,
457
+ "loss": 0.6019,
458
+ "step": 1325
459
+ },
460
+ {
461
+ "epoch": 7.458563535911602,
462
+ "grad_norm": 127.64053344726562,
463
+ "learning_rate": 1.394106813996317e-05,
464
+ "loss": 0.5221,
465
+ "step": 1350
466
+ },
467
+ {
468
+ "epoch": 7.596685082872928,
469
+ "grad_norm": 73.96366882324219,
470
+ "learning_rate": 1.3787599754450585e-05,
471
+ "loss": 0.6994,
472
+ "step": 1375
473
+ },
474
+ {
475
+ "epoch": 7.734806629834254,
476
+ "grad_norm": 109.97692108154297,
477
+ "learning_rate": 1.3634131368937999e-05,
478
+ "loss": 0.5523,
479
+ "step": 1400
480
+ },
481
+ {
482
+ "epoch": 7.87292817679558,
483
+ "grad_norm": 101.4031753540039,
484
+ "learning_rate": 1.3480662983425417e-05,
485
+ "loss": 0.5238,
486
+ "step": 1425
487
+ },
488
+ {
489
+ "epoch": 8.0,
490
+ "eval_accuracy": 0.7989690721649485,
491
+ "eval_loss": 0.8612092137336731,
492
+ "eval_macro_f1": 0.7970887738341478,
493
+ "eval_runtime": 1.0604,
494
+ "eval_samples_per_second": 548.846,
495
+ "eval_steps_per_second": 34.892,
496
+ "eval_weighted_f1": 0.7979701647109763,
497
+ "step": 1448
498
+ }
499
+ ],
500
+ "logging_steps": 25,
501
+ "max_steps": 3620,
502
+ "num_input_tokens_seen": 0,
503
+ "num_train_epochs": 20,
504
+ "save_steps": 500,
505
+ "stateful_callbacks": {
506
+ "EarlyStoppingCallback": {
507
+ "args": {
508
+ "early_stopping_patience": 3,
509
+ "early_stopping_threshold": 0.0
510
+ },
511
+ "attributes": {
512
+ "early_stopping_patience_counter": 3
513
+ }
514
+ },
515
+ "TrainerControl": {
516
+ "args": {
517
+ "should_epoch_stop": false,
518
+ "should_evaluate": false,
519
+ "should_log": false,
520
+ "should_save": true,
521
+ "should_training_stop": true
522
+ },
523
+ "attributes": {}
524
+ }
525
+ },
526
+ "total_flos": 0.0,
527
+ "train_batch_size": 8,
528
+ "trial_name": null,
529
+ "trial_params": null
530
+ }
ood-vifactcheck-xlmr-source-source-02-s42-20260901T153352Z-27bda108/trainer/checkpoint-905/special_tokens_map.json ADDED
@@ -0,0 +1,15 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "bos_token": "<s>",
3
+ "cls_token": "<s>",
4
+ "eos_token": "</s>",
5
+ "mask_token": {
6
+ "content": "<mask>",
7
+ "lstrip": true,
8
+ "normalized": false,
9
+ "rstrip": false,
10
+ "single_word": false
11
+ },
12
+ "pad_token": "<pad>",
13
+ "sep_token": "</s>",
14
+ "unk_token": "<unk>"
15
+ }
ood-vifactcheck-xlmr-source-source-02-s42-20260901T153352Z-27bda108/trainer/checkpoint-905/tokenizer_config.json ADDED
@@ -0,0 +1,55 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "added_tokens_decoder": {
3
+ "0": {
4
+ "content": "<s>",
5
+ "lstrip": false,
6
+ "normalized": false,
7
+ "rstrip": false,
8
+ "single_word": false,
9
+ "special": true
10
+ },
11
+ "1": {
12
+ "content": "<pad>",
13
+ "lstrip": false,
14
+ "normalized": false,
15
+ "rstrip": false,
16
+ "single_word": false,
17
+ "special": true
18
+ },
19
+ "2": {
20
+ "content": "</s>",
21
+ "lstrip": false,
22
+ "normalized": false,
23
+ "rstrip": false,
24
+ "single_word": false,
25
+ "special": true
26
+ },
27
+ "3": {
28
+ "content": "<unk>",
29
+ "lstrip": false,
30
+ "normalized": false,
31
+ "rstrip": false,
32
+ "single_word": false,
33
+ "special": true
34
+ },
35
+ "250001": {
36
+ "content": "<mask>",
37
+ "lstrip": true,
38
+ "normalized": false,
39
+ "rstrip": false,
40
+ "single_word": false,
41
+ "special": true
42
+ }
43
+ },
44
+ "bos_token": "<s>",
45
+ "clean_up_tokenization_spaces": false,
46
+ "cls_token": "<s>",
47
+ "eos_token": "</s>",
48
+ "extra_special_tokens": {},
49
+ "mask_token": "<mask>",
50
+ "model_max_length": 512,
51
+ "pad_token": "<pad>",
52
+ "sep_token": "</s>",
53
+ "tokenizer_class": "XLMRobertaTokenizer",
54
+ "unk_token": "<unk>"
55
+ }
ood-vifactcheck-xlmr-source-source-02-s42-20260901T153352Z-27bda108/trainer/checkpoint-905/trainer_state.json ADDED
@@ -0,0 +1,350 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_global_step": 905,
3
+ "best_metric": 0.8079147423599644,
4
+ "best_model_checkpoint": "/root/VNEncoder/results/checkpoints/ood-vifactcheck-xlmr-source-source-02-s42-20260901T153352Z-27bda108/trainer/checkpoint-905",
5
+ "epoch": 5.0,
6
+ "eval_steps": 500,
7
+ "global_step": 905,
8
+ "is_hyper_param_search": false,
9
+ "is_local_process_zero": true,
10
+ "is_world_process_zero": true,
11
+ "log_history": [
12
+ {
13
+ "epoch": 0.13812154696132597,
14
+ "grad_norm": 12.728955268859863,
15
+ "learning_rate": 1.3259668508287293e-06,
16
+ "loss": 4.8048,
17
+ "step": 25
18
+ },
19
+ {
20
+ "epoch": 0.27624309392265195,
21
+ "grad_norm": 56.0223274230957,
22
+ "learning_rate": 2.707182320441989e-06,
23
+ "loss": 4.711,
24
+ "step": 50
25
+ },
26
+ {
27
+ "epoch": 0.4143646408839779,
28
+ "grad_norm": 39.932518005371094,
29
+ "learning_rate": 4.088397790055249e-06,
30
+ "loss": 4.5371,
31
+ "step": 75
32
+ },
33
+ {
34
+ "epoch": 0.5524861878453039,
35
+ "grad_norm": 42.93894577026367,
36
+ "learning_rate": 5.469613259668509e-06,
37
+ "loss": 4.678,
38
+ "step": 100
39
+ },
40
+ {
41
+ "epoch": 0.6906077348066298,
42
+ "grad_norm": 37.85771179199219,
43
+ "learning_rate": 6.850828729281769e-06,
44
+ "loss": 4.662,
45
+ "step": 125
46
+ },
47
+ {
48
+ "epoch": 0.8287292817679558,
49
+ "grad_norm": 48.18323516845703,
50
+ "learning_rate": 8.232044198895029e-06,
51
+ "loss": 4.4032,
52
+ "step": 150
53
+ },
54
+ {
55
+ "epoch": 0.9668508287292817,
56
+ "grad_norm": 99.214599609375,
57
+ "learning_rate": 9.613259668508288e-06,
58
+ "loss": 4.2769,
59
+ "step": 175
60
+ },
61
+ {
62
+ "epoch": 1.0,
63
+ "eval_accuracy": 0.5567010309278351,
64
+ "eval_loss": 0.9564104080200195,
65
+ "eval_macro_f1": 0.531338415603882,
66
+ "eval_runtime": 1.078,
67
+ "eval_samples_per_second": 539.897,
68
+ "eval_steps_per_second": 34.323,
69
+ "eval_weighted_f1": 0.5352146874056312,
70
+ "step": 181
71
+ },
72
+ {
73
+ "epoch": 1.1049723756906078,
74
+ "grad_norm": 68.16238403320312,
75
+ "learning_rate": 1.0994475138121547e-05,
76
+ "loss": 4.0766,
77
+ "step": 200
78
+ },
79
+ {
80
+ "epoch": 1.2430939226519337,
81
+ "grad_norm": 36.474945068359375,
82
+ "learning_rate": 1.2375690607734806e-05,
83
+ "loss": 4.0206,
84
+ "step": 225
85
+ },
86
+ {
87
+ "epoch": 1.3812154696132597,
88
+ "grad_norm": 41.88302230834961,
89
+ "learning_rate": 1.3756906077348067e-05,
90
+ "loss": 4.0765,
91
+ "step": 250
92
+ },
93
+ {
94
+ "epoch": 1.5193370165745856,
95
+ "grad_norm": 73.10688781738281,
96
+ "learning_rate": 1.5138121546961326e-05,
97
+ "loss": 3.8742,
98
+ "step": 275
99
+ },
100
+ {
101
+ "epoch": 1.6574585635359116,
102
+ "grad_norm": 57.27144241333008,
103
+ "learning_rate": 1.6519337016574586e-05,
104
+ "loss": 3.7189,
105
+ "step": 300
106
+ },
107
+ {
108
+ "epoch": 1.7955801104972375,
109
+ "grad_norm": 80.90764617919922,
110
+ "learning_rate": 1.7900552486187846e-05,
111
+ "loss": 3.3838,
112
+ "step": 325
113
+ },
114
+ {
115
+ "epoch": 1.9337016574585635,
116
+ "grad_norm": 100.98841094970703,
117
+ "learning_rate": 1.9281767955801107e-05,
118
+ "loss": 3.2471,
119
+ "step": 350
120
+ },
121
+ {
122
+ "epoch": 2.0,
123
+ "eval_accuracy": 0.7096219931271478,
124
+ "eval_loss": 0.7144930362701416,
125
+ "eval_macro_f1": 0.708526703261887,
126
+ "eval_runtime": 1.087,
127
+ "eval_samples_per_second": 535.442,
128
+ "eval_steps_per_second": 34.04,
129
+ "eval_weighted_f1": 0.7085504083724671,
130
+ "step": 362
131
+ },
132
+ {
133
+ "epoch": 2.0718232044198897,
134
+ "grad_norm": 61.62918472290039,
135
+ "learning_rate": 1.9926335174953962e-05,
136
+ "loss": 2.9424,
137
+ "step": 375
138
+ },
139
+ {
140
+ "epoch": 2.2099447513812156,
141
+ "grad_norm": 66.43571472167969,
142
+ "learning_rate": 1.9772866789441377e-05,
143
+ "loss": 2.7574,
144
+ "step": 400
145
+ },
146
+ {
147
+ "epoch": 2.3480662983425415,
148
+ "grad_norm": 124.33316802978516,
149
+ "learning_rate": 1.9619398403928794e-05,
150
+ "loss": 2.8673,
151
+ "step": 425
152
+ },
153
+ {
154
+ "epoch": 2.4861878453038675,
155
+ "grad_norm": 52.74990463256836,
156
+ "learning_rate": 1.946593001841621e-05,
157
+ "loss": 2.7332,
158
+ "step": 450
159
+ },
160
+ {
161
+ "epoch": 2.6243093922651934,
162
+ "grad_norm": 120.99431610107422,
163
+ "learning_rate": 1.9312461632903623e-05,
164
+ "loss": 2.5593,
165
+ "step": 475
166
+ },
167
+ {
168
+ "epoch": 2.7624309392265194,
169
+ "grad_norm": 85.82913970947266,
170
+ "learning_rate": 1.915899324739104e-05,
171
+ "loss": 2.4126,
172
+ "step": 500
173
+ },
174
+ {
175
+ "epoch": 2.9005524861878453,
176
+ "grad_norm": 105.4771499633789,
177
+ "learning_rate": 1.9005524861878455e-05,
178
+ "loss": 2.4222,
179
+ "step": 525
180
+ },
181
+ {
182
+ "epoch": 3.0,
183
+ "eval_accuracy": 0.7663230240549829,
184
+ "eval_loss": 0.5739344358444214,
185
+ "eval_macro_f1": 0.7566652570915214,
186
+ "eval_runtime": 1.0416,
187
+ "eval_samples_per_second": 558.739,
188
+ "eval_steps_per_second": 35.521,
189
+ "eval_weighted_f1": 0.758295393308871,
190
+ "step": 543
191
+ },
192
+ {
193
+ "epoch": 3.0386740331491713,
194
+ "grad_norm": 104.53805541992188,
195
+ "learning_rate": 1.885205647636587e-05,
196
+ "loss": 2.2542,
197
+ "step": 550
198
+ },
199
+ {
200
+ "epoch": 3.1767955801104972,
201
+ "grad_norm": 119.23262786865234,
202
+ "learning_rate": 1.8698588090853287e-05,
203
+ "loss": 2.2797,
204
+ "step": 575
205
+ },
206
+ {
207
+ "epoch": 3.314917127071823,
208
+ "grad_norm": 67.51237487792969,
209
+ "learning_rate": 1.85451197053407e-05,
210
+ "loss": 1.8071,
211
+ "step": 600
212
+ },
213
+ {
214
+ "epoch": 3.453038674033149,
215
+ "grad_norm": 48.214908599853516,
216
+ "learning_rate": 1.8391651319828116e-05,
217
+ "loss": 1.9983,
218
+ "step": 625
219
+ },
220
+ {
221
+ "epoch": 3.591160220994475,
222
+ "grad_norm": 67.250732421875,
223
+ "learning_rate": 1.823818293431553e-05,
224
+ "loss": 1.8777,
225
+ "step": 650
226
+ },
227
+ {
228
+ "epoch": 3.729281767955801,
229
+ "grad_norm": 115.72201538085938,
230
+ "learning_rate": 1.8084714548802948e-05,
231
+ "loss": 1.9912,
232
+ "step": 675
233
+ },
234
+ {
235
+ "epoch": 3.867403314917127,
236
+ "grad_norm": 47.928062438964844,
237
+ "learning_rate": 1.7931246163290362e-05,
238
+ "loss": 1.9777,
239
+ "step": 700
240
+ },
241
+ {
242
+ "epoch": 4.0,
243
+ "eval_accuracy": 0.788659793814433,
244
+ "eval_loss": 0.5242534875869751,
245
+ "eval_macro_f1": 0.787816514118535,
246
+ "eval_runtime": 1.1404,
247
+ "eval_samples_per_second": 510.334,
248
+ "eval_steps_per_second": 32.444,
249
+ "eval_weighted_f1": 0.7883303300363687,
250
+ "step": 724
251
+ },
252
+ {
253
+ "epoch": 4.005524861878453,
254
+ "grad_norm": 58.05070495605469,
255
+ "learning_rate": 1.7777777777777777e-05,
256
+ "loss": 1.9077,
257
+ "step": 725
258
+ },
259
+ {
260
+ "epoch": 4.143646408839779,
261
+ "grad_norm": 108.5123062133789,
262
+ "learning_rate": 1.7624309392265194e-05,
263
+ "loss": 1.3115,
264
+ "step": 750
265
+ },
266
+ {
267
+ "epoch": 4.281767955801105,
268
+ "grad_norm": 139.55747985839844,
269
+ "learning_rate": 1.747084100675261e-05,
270
+ "loss": 1.5837,
271
+ "step": 775
272
+ },
273
+ {
274
+ "epoch": 4.419889502762431,
275
+ "grad_norm": 84.59967803955078,
276
+ "learning_rate": 1.7317372621240026e-05,
277
+ "loss": 1.5453,
278
+ "step": 800
279
+ },
280
+ {
281
+ "epoch": 4.558011049723757,
282
+ "grad_norm": 60.05119323730469,
283
+ "learning_rate": 1.716390423572744e-05,
284
+ "loss": 1.7393,
285
+ "step": 825
286
+ },
287
+ {
288
+ "epoch": 4.696132596685083,
289
+ "grad_norm": 85.76810455322266,
290
+ "learning_rate": 1.701043585021486e-05,
291
+ "loss": 1.5069,
292
+ "step": 850
293
+ },
294
+ {
295
+ "epoch": 4.834254143646409,
296
+ "grad_norm": 140.66427612304688,
297
+ "learning_rate": 1.6856967464702273e-05,
298
+ "loss": 1.6284,
299
+ "step": 875
300
+ },
301
+ {
302
+ "epoch": 4.972375690607735,
303
+ "grad_norm": 116.09174346923828,
304
+ "learning_rate": 1.670349907918969e-05,
305
+ "loss": 1.4866,
306
+ "step": 900
307
+ },
308
+ {
309
+ "epoch": 5.0,
310
+ "eval_accuracy": 0.8092783505154639,
311
+ "eval_loss": 0.5569475889205933,
312
+ "eval_macro_f1": 0.8079147423599644,
313
+ "eval_runtime": 1.1105,
314
+ "eval_samples_per_second": 524.085,
315
+ "eval_steps_per_second": 33.318,
316
+ "eval_weighted_f1": 0.8088169111415616,
317
+ "step": 905
318
+ }
319
+ ],
320
+ "logging_steps": 25,
321
+ "max_steps": 3620,
322
+ "num_input_tokens_seen": 0,
323
+ "num_train_epochs": 20,
324
+ "save_steps": 500,
325
+ "stateful_callbacks": {
326
+ "EarlyStoppingCallback": {
327
+ "args": {
328
+ "early_stopping_patience": 3,
329
+ "early_stopping_threshold": 0.0
330
+ },
331
+ "attributes": {
332
+ "early_stopping_patience_counter": 0
333
+ }
334
+ },
335
+ "TrainerControl": {
336
+ "args": {
337
+ "should_epoch_stop": false,
338
+ "should_evaluate": false,
339
+ "should_log": false,
340
+ "should_save": true,
341
+ "should_training_stop": false
342
+ },
343
+ "attributes": {}
344
+ }
345
+ },
346
+ "total_flos": 0.0,
347
+ "train_batch_size": 8,
348
+ "trial_name": null,
349
+ "trial_params": null
350
+ }
ood-vifactcheck-xlmr-source-source-02-s43-20260901T171619Z-fbd63439/trainer/checkpoint-1267/special_tokens_map.json ADDED
@@ -0,0 +1,15 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "bos_token": "<s>",
3
+ "cls_token": "<s>",
4
+ "eos_token": "</s>",
5
+ "mask_token": {
6
+ "content": "<mask>",
7
+ "lstrip": true,
8
+ "normalized": false,
9
+ "rstrip": false,
10
+ "single_word": false
11
+ },
12
+ "pad_token": "<pad>",
13
+ "sep_token": "</s>",
14
+ "unk_token": "<unk>"
15
+ }