Add files using upload-large-folder tool
Browse files- ood-vifactcheck-phobert-topic-domain-04-s43-20260814T141531Z-bdd2c308/trainer/checkpoint-2444/added_tokens.json +3 -0
- ood-vifactcheck-phobert-topic-domain-04-s43-20260814T141531Z-bdd2c308/trainer/checkpoint-2444/bpe.codes +0 -0
- ood-vifactcheck-phobert-topic-domain-04-s43-20260814T141531Z-bdd2c308/trainer/checkpoint-2444/trainer_state.json +865 -0
- ood-vifactcheck-phobert-topic-domain-04-s43-20260814T141531Z-bdd2c308/trainer/checkpoint-2444/vocab.txt +0 -0
- ood-vifactcheck-phobert-topic-domain-05-s43-20260814T143211Z-bdd2c308/trainer/checkpoint-1225/added_tokens.json +3 -0
- ood-vifactcheck-phobert-topic-domain-05-s43-20260814T143211Z-bdd2c308/trainer/checkpoint-1225/bpe.codes +0 -0
- ood-vifactcheck-phobert-topic-domain-05-s43-20260814T143211Z-bdd2c308/trainer/checkpoint-1225/special_tokens_map.json +9 -0
- ood-vifactcheck-phobert-topic-domain-05-s43-20260814T143211Z-bdd2c308/trainer/checkpoint-1225/tokenizer_config.json +55 -0
- ood-vifactcheck-phobert-topic-domain-05-s43-20260814T143211Z-bdd2c308/trainer/checkpoint-1225/trainer_state.json +463 -0
- ood-vifactcheck-phobert-topic-domain-05-s43-20260814T143211Z-bdd2c308/trainer/checkpoint-1225/vocab.txt +0 -0
- ood-vifactcheck-phobert-topic-domain-05-s43-20260814T143211Z-bdd2c308/trainer/checkpoint-700/added_tokens.json +3 -0
- ood-vifactcheck-phobert-topic-domain-05-s43-20260814T143211Z-bdd2c308/trainer/checkpoint-700/bpe.codes +0 -0
- ood-vifactcheck-phobert-topic-domain-05-s43-20260814T143211Z-bdd2c308/trainer/checkpoint-700/special_tokens_map.json +9 -0
- ood-vifactcheck-phobert-topic-domain-05-s43-20260814T143211Z-bdd2c308/trainer/checkpoint-700/tokenizer_config.json +55 -0
- ood-vifactcheck-phobert-topic-domain-05-s43-20260814T143211Z-bdd2c308/trainer/checkpoint-700/trainer_state.json +283 -0
- ood-vifactcheck-phobert-topic-domain-05-s43-20260814T143211Z-bdd2c308/trainer/checkpoint-700/vocab.txt +0 -0
ood-vifactcheck-phobert-topic-domain-04-s43-20260814T141531Z-bdd2c308/trainer/checkpoint-2444/added_tokens.json
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"<mask>": 64000
|
| 3 |
+
}
|
ood-vifactcheck-phobert-topic-domain-04-s43-20260814T141531Z-bdd2c308/trainer/checkpoint-2444/bpe.codes
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-vifactcheck-phobert-topic-domain-04-s43-20260814T141531Z-bdd2c308/trainer/checkpoint-2444/trainer_state.json
ADDED
|
@@ -0,0 +1,865 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": 1880,
|
| 3 |
+
"best_metric": 0.851402261482772,
|
| 4 |
+
"best_model_checkpoint": "/root/VNEncoder/results/checkpoints/ood-vifactcheck-phobert-topic-domain-04-s43-20260814T141531Z-bdd2c308/trainer/checkpoint-1880",
|
| 5 |
+
"epoch": 13.0,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 2444,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"epoch": 0.13351134846461948,
|
| 14 |
+
"grad_norm": 7.934043884277344,
|
| 15 |
+
"learning_rate": 1.276595744680851e-06,
|
| 16 |
+
"loss": 4.4356,
|
| 17 |
+
"step": 25
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.26702269692923897,
|
| 21 |
+
"grad_norm": 6.544791221618652,
|
| 22 |
+
"learning_rate": 2.6063829787234047e-06,
|
| 23 |
+
"loss": 4.4338,
|
| 24 |
+
"step": 50
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.40053404539385845,
|
| 28 |
+
"grad_norm": 9.024622917175293,
|
| 29 |
+
"learning_rate": 3.936170212765958e-06,
|
| 30 |
+
"loss": 4.3864,
|
| 31 |
+
"step": 75
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 0.5340453938584779,
|
| 35 |
+
"grad_norm": 10.13355827331543,
|
| 36 |
+
"learning_rate": 5.265957446808511e-06,
|
| 37 |
+
"loss": 4.3939,
|
| 38 |
+
"step": 100
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.6675567423230975,
|
| 42 |
+
"grad_norm": 8.044421195983887,
|
| 43 |
+
"learning_rate": 6.595744680851064e-06,
|
| 44 |
+
"loss": 4.32,
|
| 45 |
+
"step": 125
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.8010680907877169,
|
| 49 |
+
"grad_norm": 9.940009117126465,
|
| 50 |
+
"learning_rate": 7.925531914893617e-06,
|
| 51 |
+
"loss": 4.2966,
|
| 52 |
+
"step": 150
|
| 53 |
+
},
|
| 54 |
+
{
|
| 55 |
+
"epoch": 0.9345794392523364,
|
| 56 |
+
"grad_norm": 14.215585708618164,
|
| 57 |
+
"learning_rate": 9.255319148936171e-06,
|
| 58 |
+
"loss": 4.1807,
|
| 59 |
+
"step": 175
|
| 60 |
+
},
|
| 61 |
+
{
|
| 62 |
+
"epoch": 1.0,
|
| 63 |
+
"eval_accuracy": 0.5856079404466501,
|
| 64 |
+
"eval_loss": 0.9507163166999817,
|
| 65 |
+
"eval_macro_f1": 0.5844170151302073,
|
| 66 |
+
"eval_runtime": 1.5566,
|
| 67 |
+
"eval_samples_per_second": 517.812,
|
| 68 |
+
"eval_steps_per_second": 32.765,
|
| 69 |
+
"eval_weighted_f1": 0.5851611471295068,
|
| 70 |
+
"step": 188
|
| 71 |
+
},
|
| 72 |
+
{
|
| 73 |
+
"epoch": 1.0640854472630175,
|
| 74 |
+
"grad_norm": 21.562387466430664,
|
| 75 |
+
"learning_rate": 1.0585106382978725e-05,
|
| 76 |
+
"loss": 3.8094,
|
| 77 |
+
"step": 200
|
| 78 |
+
},
|
| 79 |
+
{
|
| 80 |
+
"epoch": 1.1975967957276368,
|
| 81 |
+
"grad_norm": 25.575475692749023,
|
| 82 |
+
"learning_rate": 1.1914893617021277e-05,
|
| 83 |
+
"loss": 3.5508,
|
| 84 |
+
"step": 225
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"epoch": 1.3311081441922563,
|
| 88 |
+
"grad_norm": 34.92906951904297,
|
| 89 |
+
"learning_rate": 1.3244680851063832e-05,
|
| 90 |
+
"loss": 3.2303,
|
| 91 |
+
"step": 250
|
| 92 |
+
},
|
| 93 |
+
{
|
| 94 |
+
"epoch": 1.4646194926568759,
|
| 95 |
+
"grad_norm": 30.78358268737793,
|
| 96 |
+
"learning_rate": 1.4574468085106384e-05,
|
| 97 |
+
"loss": 2.6556,
|
| 98 |
+
"step": 275
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 1.5981308411214954,
|
| 102 |
+
"grad_norm": 39.51386260986328,
|
| 103 |
+
"learning_rate": 1.590425531914894e-05,
|
| 104 |
+
"loss": 2.4965,
|
| 105 |
+
"step": 300
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 1.7316421895861147,
|
| 109 |
+
"grad_norm": 27.575973510742188,
|
| 110 |
+
"learning_rate": 1.723404255319149e-05,
|
| 111 |
+
"loss": 2.532,
|
| 112 |
+
"step": 325
|
| 113 |
+
},
|
| 114 |
+
{
|
| 115 |
+
"epoch": 1.8651535380507345,
|
| 116 |
+
"grad_norm": 28.04973793029785,
|
| 117 |
+
"learning_rate": 1.8563829787234043e-05,
|
| 118 |
+
"loss": 2.4271,
|
| 119 |
+
"step": 350
|
| 120 |
+
},
|
| 121 |
+
{
|
| 122 |
+
"epoch": 1.9986648865153538,
|
| 123 |
+
"grad_norm": 48.404117584228516,
|
| 124 |
+
"learning_rate": 1.9893617021276595e-05,
|
| 125 |
+
"loss": 2.3178,
|
| 126 |
+
"step": 375
|
| 127 |
+
},
|
| 128 |
+
{
|
| 129 |
+
"epoch": 2.0,
|
| 130 |
+
"eval_accuracy": 0.7977667493796526,
|
| 131 |
+
"eval_loss": 0.5268439650535583,
|
| 132 |
+
"eval_macro_f1": 0.7988966469814484,
|
| 133 |
+
"eval_runtime": 1.5368,
|
| 134 |
+
"eval_samples_per_second": 524.456,
|
| 135 |
+
"eval_steps_per_second": 33.185,
|
| 136 |
+
"eval_weighted_f1": 0.7987501272263708,
|
| 137 |
+
"step": 376
|
| 138 |
+
},
|
| 139 |
+
{
|
| 140 |
+
"epoch": 2.128170894526035,
|
| 141 |
+
"grad_norm": 22.407825469970703,
|
| 142 |
+
"learning_rate": 1.986406619385343e-05,
|
| 143 |
+
"loss": 1.6391,
|
| 144 |
+
"step": 400
|
| 145 |
+
},
|
| 146 |
+
{
|
| 147 |
+
"epoch": 2.2616822429906542,
|
| 148 |
+
"grad_norm": 16.890195846557617,
|
| 149 |
+
"learning_rate": 1.971631205673759e-05,
|
| 150 |
+
"loss": 1.6852,
|
| 151 |
+
"step": 425
|
| 152 |
+
},
|
| 153 |
+
{
|
| 154 |
+
"epoch": 2.3951935914552736,
|
| 155 |
+
"grad_norm": 55.55033493041992,
|
| 156 |
+
"learning_rate": 1.956855791962175e-05,
|
| 157 |
+
"loss": 1.7144,
|
| 158 |
+
"step": 450
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 2.5287049399198933,
|
| 162 |
+
"grad_norm": 16.002870559692383,
|
| 163 |
+
"learning_rate": 1.9420803782505913e-05,
|
| 164 |
+
"loss": 1.5319,
|
| 165 |
+
"step": 475
|
| 166 |
+
},
|
| 167 |
+
{
|
| 168 |
+
"epoch": 2.6622162883845126,
|
| 169 |
+
"grad_norm": 60.28468704223633,
|
| 170 |
+
"learning_rate": 1.927304964539007e-05,
|
| 171 |
+
"loss": 1.8092,
|
| 172 |
+
"step": 500
|
| 173 |
+
},
|
| 174 |
+
{
|
| 175 |
+
"epoch": 2.795727636849132,
|
| 176 |
+
"grad_norm": 47.960731506347656,
|
| 177 |
+
"learning_rate": 1.9125295508274234e-05,
|
| 178 |
+
"loss": 1.5918,
|
| 179 |
+
"step": 525
|
| 180 |
+
},
|
| 181 |
+
{
|
| 182 |
+
"epoch": 2.9292389853137517,
|
| 183 |
+
"grad_norm": 26.226947784423828,
|
| 184 |
+
"learning_rate": 1.8977541371158394e-05,
|
| 185 |
+
"loss": 1.4631,
|
| 186 |
+
"step": 550
|
| 187 |
+
},
|
| 188 |
+
{
|
| 189 |
+
"epoch": 3.0,
|
| 190 |
+
"eval_accuracy": 0.8188585607940446,
|
| 191 |
+
"eval_loss": 0.46551620960235596,
|
| 192 |
+
"eval_macro_f1": 0.8174860615537313,
|
| 193 |
+
"eval_runtime": 1.4874,
|
| 194 |
+
"eval_samples_per_second": 541.886,
|
| 195 |
+
"eval_steps_per_second": 34.288,
|
| 196 |
+
"eval_weighted_f1": 0.8174557710897836,
|
| 197 |
+
"step": 564
|
| 198 |
+
},
|
| 199 |
+
{
|
| 200 |
+
"epoch": 3.0587449933244324,
|
| 201 |
+
"grad_norm": 79.6458969116211,
|
| 202 |
+
"learning_rate": 1.8829787234042557e-05,
|
| 203 |
+
"loss": 1.4074,
|
| 204 |
+
"step": 575
|
| 205 |
+
},
|
| 206 |
+
{
|
| 207 |
+
"epoch": 3.192256341789052,
|
| 208 |
+
"grad_norm": 29.355688095092773,
|
| 209 |
+
"learning_rate": 1.8682033096926714e-05,
|
| 210 |
+
"loss": 1.0495,
|
| 211 |
+
"step": 600
|
| 212 |
+
},
|
| 213 |
+
{
|
| 214 |
+
"epoch": 3.3257676902536715,
|
| 215 |
+
"grad_norm": 29.34524154663086,
|
| 216 |
+
"learning_rate": 1.8534278959810878e-05,
|
| 217 |
+
"loss": 1.0116,
|
| 218 |
+
"step": 625
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 3.4592790387182912,
|
| 222 |
+
"grad_norm": 51.03464126586914,
|
| 223 |
+
"learning_rate": 1.8386524822695038e-05,
|
| 224 |
+
"loss": 1.1958,
|
| 225 |
+
"step": 650
|
| 226 |
+
},
|
| 227 |
+
{
|
| 228 |
+
"epoch": 3.5927903871829105,
|
| 229 |
+
"grad_norm": 50.30847930908203,
|
| 230 |
+
"learning_rate": 1.8238770685579198e-05,
|
| 231 |
+
"loss": 0.9959,
|
| 232 |
+
"step": 675
|
| 233 |
+
},
|
| 234 |
+
{
|
| 235 |
+
"epoch": 3.72630173564753,
|
| 236 |
+
"grad_norm": 23.34055519104004,
|
| 237 |
+
"learning_rate": 1.8091016548463358e-05,
|
| 238 |
+
"loss": 1.0551,
|
| 239 |
+
"step": 700
|
| 240 |
+
},
|
| 241 |
+
{
|
| 242 |
+
"epoch": 3.8598130841121496,
|
| 243 |
+
"grad_norm": 35.14168167114258,
|
| 244 |
+
"learning_rate": 1.7943262411347518e-05,
|
| 245 |
+
"loss": 1.211,
|
| 246 |
+
"step": 725
|
| 247 |
+
},
|
| 248 |
+
{
|
| 249 |
+
"epoch": 3.993324432576769,
|
| 250 |
+
"grad_norm": 36.98830795288086,
|
| 251 |
+
"learning_rate": 1.779550827423168e-05,
|
| 252 |
+
"loss": 1.1453,
|
| 253 |
+
"step": 750
|
| 254 |
+
},
|
| 255 |
+
{
|
| 256 |
+
"epoch": 4.0,
|
| 257 |
+
"eval_accuracy": 0.8275434243176178,
|
| 258 |
+
"eval_loss": 0.5185574293136597,
|
| 259 |
+
"eval_macro_f1": 0.8285568830183712,
|
| 260 |
+
"eval_runtime": 1.5009,
|
| 261 |
+
"eval_samples_per_second": 537.009,
|
| 262 |
+
"eval_steps_per_second": 33.98,
|
| 263 |
+
"eval_weighted_f1": 0.8283670301172862,
|
| 264 |
+
"step": 752
|
| 265 |
+
},
|
| 266 |
+
{
|
| 267 |
+
"epoch": 4.12283044058745,
|
| 268 |
+
"grad_norm": 45.3942985534668,
|
| 269 |
+
"learning_rate": 1.7647754137115842e-05,
|
| 270 |
+
"loss": 0.7742,
|
| 271 |
+
"step": 775
|
| 272 |
+
},
|
| 273 |
+
{
|
| 274 |
+
"epoch": 4.25634178905207,
|
| 275 |
+
"grad_norm": 59.520057678222656,
|
| 276 |
+
"learning_rate": 1.7500000000000002e-05,
|
| 277 |
+
"loss": 0.6921,
|
| 278 |
+
"step": 800
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 4.389853137516689,
|
| 282 |
+
"grad_norm": 53.73417663574219,
|
| 283 |
+
"learning_rate": 1.7352245862884162e-05,
|
| 284 |
+
"loss": 0.7898,
|
| 285 |
+
"step": 825
|
| 286 |
+
},
|
| 287 |
+
{
|
| 288 |
+
"epoch": 4.5233644859813085,
|
| 289 |
+
"grad_norm": 96.21471405029297,
|
| 290 |
+
"learning_rate": 1.7204491725768322e-05,
|
| 291 |
+
"loss": 0.8286,
|
| 292 |
+
"step": 850
|
| 293 |
+
},
|
| 294 |
+
{
|
| 295 |
+
"epoch": 4.656875834445928,
|
| 296 |
+
"grad_norm": 29.843088150024414,
|
| 297 |
+
"learning_rate": 1.7056737588652482e-05,
|
| 298 |
+
"loss": 0.9201,
|
| 299 |
+
"step": 875
|
| 300 |
+
},
|
| 301 |
+
{
|
| 302 |
+
"epoch": 4.790387182910547,
|
| 303 |
+
"grad_norm": 89.1033935546875,
|
| 304 |
+
"learning_rate": 1.6908983451536646e-05,
|
| 305 |
+
"loss": 0.6124,
|
| 306 |
+
"step": 900
|
| 307 |
+
},
|
| 308 |
+
{
|
| 309 |
+
"epoch": 4.923898531375167,
|
| 310 |
+
"grad_norm": 18.528675079345703,
|
| 311 |
+
"learning_rate": 1.6761229314420806e-05,
|
| 312 |
+
"loss": 0.9164,
|
| 313 |
+
"step": 925
|
| 314 |
+
},
|
| 315 |
+
{
|
| 316 |
+
"epoch": 5.0,
|
| 317 |
+
"eval_accuracy": 0.8238213399503722,
|
| 318 |
+
"eval_loss": 0.5889860987663269,
|
| 319 |
+
"eval_macro_f1": 0.8242997973798257,
|
| 320 |
+
"eval_runtime": 1.5115,
|
| 321 |
+
"eval_samples_per_second": 533.252,
|
| 322 |
+
"eval_steps_per_second": 33.742,
|
| 323 |
+
"eval_weighted_f1": 0.8241527532134966,
|
| 324 |
+
"step": 940
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"epoch": 5.0534045393858475,
|
| 328 |
+
"grad_norm": 9.284804344177246,
|
| 329 |
+
"learning_rate": 1.6613475177304966e-05,
|
| 330 |
+
"loss": 0.6795,
|
| 331 |
+
"step": 950
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 5.186915887850467,
|
| 335 |
+
"grad_norm": 62.4229850769043,
|
| 336 |
+
"learning_rate": 1.6465721040189126e-05,
|
| 337 |
+
"loss": 0.5714,
|
| 338 |
+
"step": 975
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 5.320427236315087,
|
| 342 |
+
"grad_norm": 96.04256439208984,
|
| 343 |
+
"learning_rate": 1.6317966903073286e-05,
|
| 344 |
+
"loss": 0.5529,
|
| 345 |
+
"step": 1000
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 5.453938584779706,
|
| 349 |
+
"grad_norm": 20.270235061645508,
|
| 350 |
+
"learning_rate": 1.6170212765957446e-05,
|
| 351 |
+
"loss": 0.6248,
|
| 352 |
+
"step": 1025
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 5.587449933244326,
|
| 356 |
+
"grad_norm": 3.7398602962493896,
|
| 357 |
+
"learning_rate": 1.602245862884161e-05,
|
| 358 |
+
"loss": 0.4304,
|
| 359 |
+
"step": 1050
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 5.720961281708945,
|
| 363 |
+
"grad_norm": 44.74562454223633,
|
| 364 |
+
"learning_rate": 1.587470449172577e-05,
|
| 365 |
+
"loss": 0.535,
|
| 366 |
+
"step": 1075
|
| 367 |
+
},
|
| 368 |
+
{
|
| 369 |
+
"epoch": 5.854472630173564,
|
| 370 |
+
"grad_norm": 62.393531799316406,
|
| 371 |
+
"learning_rate": 1.572695035460993e-05,
|
| 372 |
+
"loss": 0.7312,
|
| 373 |
+
"step": 1100
|
| 374 |
+
},
|
| 375 |
+
{
|
| 376 |
+
"epoch": 5.9879839786381845,
|
| 377 |
+
"grad_norm": 85.06871032714844,
|
| 378 |
+
"learning_rate": 1.557919621749409e-05,
|
| 379 |
+
"loss": 0.6614,
|
| 380 |
+
"step": 1125
|
| 381 |
+
},
|
| 382 |
+
{
|
| 383 |
+
"epoch": 6.0,
|
| 384 |
+
"eval_accuracy": 0.8362282878411911,
|
| 385 |
+
"eval_loss": 0.6402100920677185,
|
| 386 |
+
"eval_macro_f1": 0.8373454860413431,
|
| 387 |
+
"eval_runtime": 1.5361,
|
| 388 |
+
"eval_samples_per_second": 524.721,
|
| 389 |
+
"eval_steps_per_second": 33.202,
|
| 390 |
+
"eval_weighted_f1": 0.8370620771745654,
|
| 391 |
+
"step": 1128
|
| 392 |
+
},
|
| 393 |
+
{
|
| 394 |
+
"epoch": 6.117489986648865,
|
| 395 |
+
"grad_norm": 58.2725830078125,
|
| 396 |
+
"learning_rate": 1.543144208037825e-05,
|
| 397 |
+
"loss": 0.4208,
|
| 398 |
+
"step": 1150
|
| 399 |
+
},
|
| 400 |
+
{
|
| 401 |
+
"epoch": 6.251001335113485,
|
| 402 |
+
"grad_norm": 264.6571960449219,
|
| 403 |
+
"learning_rate": 1.5283687943262414e-05,
|
| 404 |
+
"loss": 0.4828,
|
| 405 |
+
"step": 1175
|
| 406 |
+
},
|
| 407 |
+
{
|
| 408 |
+
"epoch": 6.384512683578104,
|
| 409 |
+
"grad_norm": 0.8864717483520508,
|
| 410 |
+
"learning_rate": 1.5135933806146574e-05,
|
| 411 |
+
"loss": 0.2963,
|
| 412 |
+
"step": 1200
|
| 413 |
+
},
|
| 414 |
+
{
|
| 415 |
+
"epoch": 6.518024032042724,
|
| 416 |
+
"grad_norm": 283.2027587890625,
|
| 417 |
+
"learning_rate": 1.4988179669030734e-05,
|
| 418 |
+
"loss": 0.4004,
|
| 419 |
+
"step": 1225
|
| 420 |
+
},
|
| 421 |
+
{
|
| 422 |
+
"epoch": 6.651535380507343,
|
| 423 |
+
"grad_norm": 105.5600814819336,
|
| 424 |
+
"learning_rate": 1.4840425531914894e-05,
|
| 425 |
+
"loss": 0.4072,
|
| 426 |
+
"step": 1250
|
| 427 |
+
},
|
| 428 |
+
{
|
| 429 |
+
"epoch": 6.785046728971962,
|
| 430 |
+
"grad_norm": 41.844276428222656,
|
| 431 |
+
"learning_rate": 1.4692671394799056e-05,
|
| 432 |
+
"loss": 0.6366,
|
| 433 |
+
"step": 1275
|
| 434 |
+
},
|
| 435 |
+
{
|
| 436 |
+
"epoch": 6.9185580774365825,
|
| 437 |
+
"grad_norm": 0.7505985498428345,
|
| 438 |
+
"learning_rate": 1.4544917257683215e-05,
|
| 439 |
+
"loss": 0.3737,
|
| 440 |
+
"step": 1300
|
| 441 |
+
},
|
| 442 |
+
{
|
| 443 |
+
"epoch": 7.0,
|
| 444 |
+
"eval_accuracy": 0.8362282878411911,
|
| 445 |
+
"eval_loss": 0.7384033203125,
|
| 446 |
+
"eval_macro_f1": 0.8366120880359323,
|
| 447 |
+
"eval_runtime": 1.5457,
|
| 448 |
+
"eval_samples_per_second": 521.453,
|
| 449 |
+
"eval_steps_per_second": 32.995,
|
| 450 |
+
"eval_weighted_f1": 0.8364901116757201,
|
| 451 |
+
"step": 1316
|
| 452 |
+
},
|
| 453 |
+
{
|
| 454 |
+
"epoch": 7.048064085447263,
|
| 455 |
+
"grad_norm": 2.613600015640259,
|
| 456 |
+
"learning_rate": 1.4397163120567377e-05,
|
| 457 |
+
"loss": 0.4115,
|
| 458 |
+
"step": 1325
|
| 459 |
+
},
|
| 460 |
+
{
|
| 461 |
+
"epoch": 7.181575433911883,
|
| 462 |
+
"grad_norm": 46.876792907714844,
|
| 463 |
+
"learning_rate": 1.4249408983451538e-05,
|
| 464 |
+
"loss": 0.5657,
|
| 465 |
+
"step": 1350
|
| 466 |
+
},
|
| 467 |
+
{
|
| 468 |
+
"epoch": 7.315086782376502,
|
| 469 |
+
"grad_norm": 52.67552185058594,
|
| 470 |
+
"learning_rate": 1.41016548463357e-05,
|
| 471 |
+
"loss": 0.3544,
|
| 472 |
+
"step": 1375
|
| 473 |
+
},
|
| 474 |
+
{
|
| 475 |
+
"epoch": 7.4485981308411215,
|
| 476 |
+
"grad_norm": 2.136875629425049,
|
| 477 |
+
"learning_rate": 1.3953900709219859e-05,
|
| 478 |
+
"loss": 0.5325,
|
| 479 |
+
"step": 1400
|
| 480 |
+
},
|
| 481 |
+
{
|
| 482 |
+
"epoch": 7.582109479305741,
|
| 483 |
+
"grad_norm": 10.860918998718262,
|
| 484 |
+
"learning_rate": 1.380614657210402e-05,
|
| 485 |
+
"loss": 0.3777,
|
| 486 |
+
"step": 1425
|
| 487 |
+
},
|
| 488 |
+
{
|
| 489 |
+
"epoch": 7.71562082777036,
|
| 490 |
+
"grad_norm": 1.0959819555282593,
|
| 491 |
+
"learning_rate": 1.365839243498818e-05,
|
| 492 |
+
"loss": 0.3069,
|
| 493 |
+
"step": 1450
|
| 494 |
+
},
|
| 495 |
+
{
|
| 496 |
+
"epoch": 7.8491321762349795,
|
| 497 |
+
"grad_norm": 24.337228775024414,
|
| 498 |
+
"learning_rate": 1.351063829787234e-05,
|
| 499 |
+
"loss": 0.4703,
|
| 500 |
+
"step": 1475
|
| 501 |
+
},
|
| 502 |
+
{
|
| 503 |
+
"epoch": 7.9826435246996,
|
| 504 |
+
"grad_norm": 0.31375980377197266,
|
| 505 |
+
"learning_rate": 1.3362884160756503e-05,
|
| 506 |
+
"loss": 0.3374,
|
| 507 |
+
"step": 1500
|
| 508 |
+
},
|
| 509 |
+
{
|
| 510 |
+
"epoch": 8.0,
|
| 511 |
+
"eval_accuracy": 0.8337468982630273,
|
| 512 |
+
"eval_loss": 0.7685278654098511,
|
| 513 |
+
"eval_macro_f1": 0.8344114186003894,
|
| 514 |
+
"eval_runtime": 1.4795,
|
| 515 |
+
"eval_samples_per_second": 544.786,
|
| 516 |
+
"eval_steps_per_second": 34.472,
|
| 517 |
+
"eval_weighted_f1": 0.8343047182350389,
|
| 518 |
+
"step": 1504
|
| 519 |
+
},
|
| 520 |
+
{
|
| 521 |
+
"epoch": 8.11214953271028,
|
| 522 |
+
"grad_norm": 4.257840633392334,
|
| 523 |
+
"learning_rate": 1.3215130023640663e-05,
|
| 524 |
+
"loss": 0.2527,
|
| 525 |
+
"step": 1525
|
| 526 |
+
},
|
| 527 |
+
{
|
| 528 |
+
"epoch": 8.2456608811749,
|
| 529 |
+
"grad_norm": 209.26927185058594,
|
| 530 |
+
"learning_rate": 1.3067375886524825e-05,
|
| 531 |
+
"loss": 0.4283,
|
| 532 |
+
"step": 1550
|
| 533 |
+
},
|
| 534 |
+
{
|
| 535 |
+
"epoch": 8.379172229639519,
|
| 536 |
+
"grad_norm": 0.261762410402298,
|
| 537 |
+
"learning_rate": 1.2919621749408983e-05,
|
| 538 |
+
"loss": 0.364,
|
| 539 |
+
"step": 1575
|
| 540 |
+
},
|
| 541 |
+
{
|
| 542 |
+
"epoch": 8.51268357810414,
|
| 543 |
+
"grad_norm": 0.21697907149791718,
|
| 544 |
+
"learning_rate": 1.2771867612293145e-05,
|
| 545 |
+
"loss": 0.3595,
|
| 546 |
+
"step": 1600
|
| 547 |
+
},
|
| 548 |
+
{
|
| 549 |
+
"epoch": 8.646194926568759,
|
| 550 |
+
"grad_norm": 3.3832175731658936,
|
| 551 |
+
"learning_rate": 1.2624113475177307e-05,
|
| 552 |
+
"loss": 0.234,
|
| 553 |
+
"step": 1625
|
| 554 |
+
},
|
| 555 |
+
{
|
| 556 |
+
"epoch": 8.779706275033378,
|
| 557 |
+
"grad_norm": 7.727792263031006,
|
| 558 |
+
"learning_rate": 1.2476359338061469e-05,
|
| 559 |
+
"loss": 0.2908,
|
| 560 |
+
"step": 1650
|
| 561 |
+
},
|
| 562 |
+
{
|
| 563 |
+
"epoch": 8.913217623497998,
|
| 564 |
+
"grad_norm": 1.895972728729248,
|
| 565 |
+
"learning_rate": 1.2328605200945627e-05,
|
| 566 |
+
"loss": 0.2292,
|
| 567 |
+
"step": 1675
|
| 568 |
+
},
|
| 569 |
+
{
|
| 570 |
+
"epoch": 9.0,
|
| 571 |
+
"eval_accuracy": 0.8399503722084367,
|
| 572 |
+
"eval_loss": 0.8085663914680481,
|
| 573 |
+
"eval_macro_f1": 0.8400026638585129,
|
| 574 |
+
"eval_runtime": 1.4607,
|
| 575 |
+
"eval_samples_per_second": 551.777,
|
| 576 |
+
"eval_steps_per_second": 34.914,
|
| 577 |
+
"eval_weighted_f1": 0.8397936836652928,
|
| 578 |
+
"step": 1692
|
| 579 |
+
},
|
| 580 |
+
{
|
| 581 |
+
"epoch": 9.042723631508679,
|
| 582 |
+
"grad_norm": 0.21228604018688202,
|
| 583 |
+
"learning_rate": 1.2180851063829789e-05,
|
| 584 |
+
"loss": 0.296,
|
| 585 |
+
"step": 1700
|
| 586 |
+
},
|
| 587 |
+
{
|
| 588 |
+
"epoch": 9.176234979973298,
|
| 589 |
+
"grad_norm": 15.617260932922363,
|
| 590 |
+
"learning_rate": 1.2033096926713949e-05,
|
| 591 |
+
"loss": 0.3964,
|
| 592 |
+
"step": 1725
|
| 593 |
+
},
|
| 594 |
+
{
|
| 595 |
+
"epoch": 9.309746328437917,
|
| 596 |
+
"grad_norm": 0.32938575744628906,
|
| 597 |
+
"learning_rate": 1.1885342789598109e-05,
|
| 598 |
+
"loss": 0.1054,
|
| 599 |
+
"step": 1750
|
| 600 |
+
},
|
| 601 |
+
{
|
| 602 |
+
"epoch": 9.443257676902537,
|
| 603 |
+
"grad_norm": 0.19549989700317383,
|
| 604 |
+
"learning_rate": 1.1737588652482271e-05,
|
| 605 |
+
"loss": 0.1513,
|
| 606 |
+
"step": 1775
|
| 607 |
+
},
|
| 608 |
+
{
|
| 609 |
+
"epoch": 9.576769025367156,
|
| 610 |
+
"grad_norm": 0.27149826288223267,
|
| 611 |
+
"learning_rate": 1.1589834515366431e-05,
|
| 612 |
+
"loss": 0.1925,
|
| 613 |
+
"step": 1800
|
| 614 |
+
},
|
| 615 |
+
{
|
| 616 |
+
"epoch": 9.710280373831775,
|
| 617 |
+
"grad_norm": 16.880571365356445,
|
| 618 |
+
"learning_rate": 1.1442080378250593e-05,
|
| 619 |
+
"loss": 0.2079,
|
| 620 |
+
"step": 1825
|
| 621 |
+
},
|
| 622 |
+
{
|
| 623 |
+
"epoch": 9.843791722296395,
|
| 624 |
+
"grad_norm": 0.19067148864269257,
|
| 625 |
+
"learning_rate": 1.1294326241134751e-05,
|
| 626 |
+
"loss": 0.1876,
|
| 627 |
+
"step": 1850
|
| 628 |
+
},
|
| 629 |
+
{
|
| 630 |
+
"epoch": 9.977303070761014,
|
| 631 |
+
"grad_norm": 10.91338062286377,
|
| 632 |
+
"learning_rate": 1.1146572104018913e-05,
|
| 633 |
+
"loss": 0.3566,
|
| 634 |
+
"step": 1875
|
| 635 |
+
},
|
| 636 |
+
{
|
| 637 |
+
"epoch": 10.0,
|
| 638 |
+
"eval_accuracy": 0.8511166253101737,
|
| 639 |
+
"eval_loss": 0.7843964099884033,
|
| 640 |
+
"eval_macro_f1": 0.851402261482772,
|
| 641 |
+
"eval_runtime": 1.5441,
|
| 642 |
+
"eval_samples_per_second": 521.989,
|
| 643 |
+
"eval_steps_per_second": 33.029,
|
| 644 |
+
"eval_weighted_f1": 0.8511831879998023,
|
| 645 |
+
"step": 1880
|
| 646 |
+
},
|
| 647 |
+
{
|
| 648 |
+
"epoch": 10.106809078771695,
|
| 649 |
+
"grad_norm": 0.7412228584289551,
|
| 650 |
+
"learning_rate": 1.0998817966903075e-05,
|
| 651 |
+
"loss": 0.2339,
|
| 652 |
+
"step": 1900
|
| 653 |
+
},
|
| 654 |
+
{
|
| 655 |
+
"epoch": 10.240320427236314,
|
| 656 |
+
"grad_norm": 0.16804783046245575,
|
| 657 |
+
"learning_rate": 1.0851063829787233e-05,
|
| 658 |
+
"loss": 0.1806,
|
| 659 |
+
"step": 1925
|
| 660 |
+
},
|
| 661 |
+
{
|
| 662 |
+
"epoch": 10.373831775700934,
|
| 663 |
+
"grad_norm": 78.7177734375,
|
| 664 |
+
"learning_rate": 1.0703309692671395e-05,
|
| 665 |
+
"loss": 0.1535,
|
| 666 |
+
"step": 1950
|
| 667 |
+
},
|
| 668 |
+
{
|
| 669 |
+
"epoch": 10.507343124165555,
|
| 670 |
+
"grad_norm": 0.1393023133277893,
|
| 671 |
+
"learning_rate": 1.0555555555555557e-05,
|
| 672 |
+
"loss": 0.1048,
|
| 673 |
+
"step": 1975
|
| 674 |
+
},
|
| 675 |
+
{
|
| 676 |
+
"epoch": 10.640854472630174,
|
| 677 |
+
"grad_norm": 0.15626592934131622,
|
| 678 |
+
"learning_rate": 1.0407801418439717e-05,
|
| 679 |
+
"loss": 0.1934,
|
| 680 |
+
"step": 2000
|
| 681 |
+
},
|
| 682 |
+
{
|
| 683 |
+
"epoch": 10.774365821094793,
|
| 684 |
+
"grad_norm": 1.1121844053268433,
|
| 685 |
+
"learning_rate": 1.0260047281323877e-05,
|
| 686 |
+
"loss": 0.1597,
|
| 687 |
+
"step": 2025
|
| 688 |
+
},
|
| 689 |
+
{
|
| 690 |
+
"epoch": 10.907877169559413,
|
| 691 |
+
"grad_norm": 0.1610051542520523,
|
| 692 |
+
"learning_rate": 1.011229314420804e-05,
|
| 693 |
+
"loss": 0.2265,
|
| 694 |
+
"step": 2050
|
| 695 |
+
},
|
| 696 |
+
{
|
| 697 |
+
"epoch": 11.0,
|
| 698 |
+
"eval_accuracy": 0.8387096774193549,
|
| 699 |
+
"eval_loss": 0.892425000667572,
|
| 700 |
+
"eval_macro_f1": 0.8382187728419929,
|
| 701 |
+
"eval_runtime": 1.5046,
|
| 702 |
+
"eval_samples_per_second": 535.696,
|
| 703 |
+
"eval_steps_per_second": 33.896,
|
| 704 |
+
"eval_weighted_f1": 0.8380808910159788,
|
| 705 |
+
"step": 2068
|
| 706 |
+
},
|
| 707 |
+
{
|
| 708 |
+
"epoch": 11.037383177570094,
|
| 709 |
+
"grad_norm": 0.15600396692752838,
|
| 710 |
+
"learning_rate": 9.9645390070922e-06,
|
| 711 |
+
"loss": 0.0409,
|
| 712 |
+
"step": 2075
|
| 713 |
+
},
|
| 714 |
+
{
|
| 715 |
+
"epoch": 11.170894526034713,
|
| 716 |
+
"grad_norm": 67.61930084228516,
|
| 717 |
+
"learning_rate": 9.81678486997636e-06,
|
| 718 |
+
"loss": 0.1319,
|
| 719 |
+
"step": 2100
|
| 720 |
+
},
|
| 721 |
+
{
|
| 722 |
+
"epoch": 11.304405874499333,
|
| 723 |
+
"grad_norm": 0.18008147180080414,
|
| 724 |
+
"learning_rate": 9.66903073286052e-06,
|
| 725 |
+
"loss": 0.0594,
|
| 726 |
+
"step": 2125
|
| 727 |
+
},
|
| 728 |
+
{
|
| 729 |
+
"epoch": 11.437917222963952,
|
| 730 |
+
"grad_norm": 0.13436464965343475,
|
| 731 |
+
"learning_rate": 9.521276595744681e-06,
|
| 732 |
+
"loss": 0.0949,
|
| 733 |
+
"step": 2150
|
| 734 |
+
},
|
| 735 |
+
{
|
| 736 |
+
"epoch": 11.571428571428571,
|
| 737 |
+
"grad_norm": 9.68471622467041,
|
| 738 |
+
"learning_rate": 9.373522458628842e-06,
|
| 739 |
+
"loss": 0.1318,
|
| 740 |
+
"step": 2175
|
| 741 |
+
},
|
| 742 |
+
{
|
| 743 |
+
"epoch": 11.70493991989319,
|
| 744 |
+
"grad_norm": 7.8745198249816895,
|
| 745 |
+
"learning_rate": 9.225768321513003e-06,
|
| 746 |
+
"loss": 0.1741,
|
| 747 |
+
"step": 2200
|
| 748 |
+
},
|
| 749 |
+
{
|
| 750 |
+
"epoch": 11.83845126835781,
|
| 751 |
+
"grad_norm": 0.11418383568525314,
|
| 752 |
+
"learning_rate": 9.078014184397164e-06,
|
| 753 |
+
"loss": 0.0177,
|
| 754 |
+
"step": 2225
|
| 755 |
+
},
|
| 756 |
+
{
|
| 757 |
+
"epoch": 11.97196261682243,
|
| 758 |
+
"grad_norm": 0.11664965748786926,
|
| 759 |
+
"learning_rate": 8.930260047281325e-06,
|
| 760 |
+
"loss": 0.2132,
|
| 761 |
+
"step": 2250
|
| 762 |
+
},
|
| 763 |
+
{
|
| 764 |
+
"epoch": 12.0,
|
| 765 |
+
"eval_accuracy": 0.8449131513647643,
|
| 766 |
+
"eval_loss": 0.9302653074264526,
|
| 767 |
+
"eval_macro_f1": 0.8454159860586411,
|
| 768 |
+
"eval_runtime": 1.5574,
|
| 769 |
+
"eval_samples_per_second": 517.518,
|
| 770 |
+
"eval_steps_per_second": 32.746,
|
| 771 |
+
"eval_weighted_f1": 0.8452329573587182,
|
| 772 |
+
"step": 2256
|
| 773 |
+
},
|
| 774 |
+
{
|
| 775 |
+
"epoch": 12.10146862483311,
|
| 776 |
+
"grad_norm": 0.1059003695845604,
|
| 777 |
+
"learning_rate": 8.782505910165485e-06,
|
| 778 |
+
"loss": 0.0804,
|
| 779 |
+
"step": 2275
|
| 780 |
+
},
|
| 781 |
+
{
|
| 782 |
+
"epoch": 12.23497997329773,
|
| 783 |
+
"grad_norm": 0.1035916805267334,
|
| 784 |
+
"learning_rate": 8.634751773049647e-06,
|
| 785 |
+
"loss": 0.0484,
|
| 786 |
+
"step": 2300
|
| 787 |
+
},
|
| 788 |
+
{
|
| 789 |
+
"epoch": 12.368491321762349,
|
| 790 |
+
"grad_norm": 0.11409342288970947,
|
| 791 |
+
"learning_rate": 8.486997635933807e-06,
|
| 792 |
+
"loss": 0.0732,
|
| 793 |
+
"step": 2325
|
| 794 |
+
},
|
| 795 |
+
{
|
| 796 |
+
"epoch": 12.50200267022697,
|
| 797 |
+
"grad_norm": 1.819298267364502,
|
| 798 |
+
"learning_rate": 8.339243498817968e-06,
|
| 799 |
+
"loss": 0.1016,
|
| 800 |
+
"step": 2350
|
| 801 |
+
},
|
| 802 |
+
{
|
| 803 |
+
"epoch": 12.63551401869159,
|
| 804 |
+
"grad_norm": 1.8857816457748413,
|
| 805 |
+
"learning_rate": 8.191489361702128e-06,
|
| 806 |
+
"loss": 0.1649,
|
| 807 |
+
"step": 2375
|
| 808 |
+
},
|
| 809 |
+
{
|
| 810 |
+
"epoch": 12.769025367156209,
|
| 811 |
+
"grad_norm": 0.09834053367376328,
|
| 812 |
+
"learning_rate": 8.04373522458629e-06,
|
| 813 |
+
"loss": 0.181,
|
| 814 |
+
"step": 2400
|
| 815 |
+
},
|
| 816 |
+
{
|
| 817 |
+
"epoch": 12.902536715620828,
|
| 818 |
+
"grad_norm": 0.09086725860834122,
|
| 819 |
+
"learning_rate": 7.89598108747045e-06,
|
| 820 |
+
"loss": 0.0256,
|
| 821 |
+
"step": 2425
|
| 822 |
+
},
|
| 823 |
+
{
|
| 824 |
+
"epoch": 13.0,
|
| 825 |
+
"eval_accuracy": 0.8461538461538461,
|
| 826 |
+
"eval_loss": 0.9327659606933594,
|
| 827 |
+
"eval_macro_f1": 0.8467000742994991,
|
| 828 |
+
"eval_runtime": 1.4831,
|
| 829 |
+
"eval_samples_per_second": 543.449,
|
| 830 |
+
"eval_steps_per_second": 34.387,
|
| 831 |
+
"eval_weighted_f1": 0.8464963612837884,
|
| 832 |
+
"step": 2444
|
| 833 |
+
}
|
| 834 |
+
],
|
| 835 |
+
"logging_steps": 25,
|
| 836 |
+
"max_steps": 3760,
|
| 837 |
+
"num_input_tokens_seen": 0,
|
| 838 |
+
"num_train_epochs": 20,
|
| 839 |
+
"save_steps": 500,
|
| 840 |
+
"stateful_callbacks": {
|
| 841 |
+
"EarlyStoppingCallback": {
|
| 842 |
+
"args": {
|
| 843 |
+
"early_stopping_patience": 3,
|
| 844 |
+
"early_stopping_threshold": 0.0
|
| 845 |
+
},
|
| 846 |
+
"attributes": {
|
| 847 |
+
"early_stopping_patience_counter": 3
|
| 848 |
+
}
|
| 849 |
+
},
|
| 850 |
+
"TrainerControl": {
|
| 851 |
+
"args": {
|
| 852 |
+
"should_epoch_stop": false,
|
| 853 |
+
"should_evaluate": false,
|
| 854 |
+
"should_log": false,
|
| 855 |
+
"should_save": true,
|
| 856 |
+
"should_training_stop": true
|
| 857 |
+
},
|
| 858 |
+
"attributes": {}
|
| 859 |
+
}
|
| 860 |
+
},
|
| 861 |
+
"total_flos": 0.0,
|
| 862 |
+
"train_batch_size": 8,
|
| 863 |
+
"trial_name": null,
|
| 864 |
+
"trial_params": null
|
| 865 |
+
}
|
ood-vifactcheck-phobert-topic-domain-04-s43-20260814T141531Z-bdd2c308/trainer/checkpoint-2444/vocab.txt
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-vifactcheck-phobert-topic-domain-05-s43-20260814T143211Z-bdd2c308/trainer/checkpoint-1225/added_tokens.json
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"<mask>": 64000
|
| 3 |
+
}
|
ood-vifactcheck-phobert-topic-domain-05-s43-20260814T143211Z-bdd2c308/trainer/checkpoint-1225/bpe.codes
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-vifactcheck-phobert-topic-domain-05-s43-20260814T143211Z-bdd2c308/trainer/checkpoint-1225/special_tokens_map.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token": "<s>",
|
| 3 |
+
"cls_token": "<s>",
|
| 4 |
+
"eos_token": "</s>",
|
| 5 |
+
"mask_token": "<mask>",
|
| 6 |
+
"pad_token": "<pad>",
|
| 7 |
+
"sep_token": "</s>",
|
| 8 |
+
"unk_token": "<unk>"
|
| 9 |
+
}
|
ood-vifactcheck-phobert-topic-domain-05-s43-20260814T143211Z-bdd2c308/trainer/checkpoint-1225/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 |
+
"64000": {
|
| 36 |
+
"content": "<mask>",
|
| 37 |
+
"lstrip": false,
|
| 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": 256,
|
| 51 |
+
"pad_token": "<pad>",
|
| 52 |
+
"sep_token": "</s>",
|
| 53 |
+
"tokenizer_class": "PhobertTokenizer",
|
| 54 |
+
"unk_token": "<unk>"
|
| 55 |
+
}
|
ood-vifactcheck-phobert-topic-domain-05-s43-20260814T143211Z-bdd2c308/trainer/checkpoint-1225/trainer_state.json
ADDED
|
@@ -0,0 +1,463 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": 700,
|
| 3 |
+
"best_metric": 0.8318811799298261,
|
| 4 |
+
"best_model_checkpoint": "/root/VNEncoder/results/checkpoints/ood-vifactcheck-phobert-topic-domain-05-s43-20260814T143211Z-bdd2c308/trainer/checkpoint-700",
|
| 5 |
+
"epoch": 7.0,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 1225,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"epoch": 0.1430615164520744,
|
| 14 |
+
"grad_norm": 9.187152862548828,
|
| 15 |
+
"learning_rate": 1.3714285714285717e-06,
|
| 16 |
+
"loss": 4.4471,
|
| 17 |
+
"step": 25
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.2861230329041488,
|
| 21 |
+
"grad_norm": 6.098942279815674,
|
| 22 |
+
"learning_rate": 2.8000000000000003e-06,
|
| 23 |
+
"loss": 4.418,
|
| 24 |
+
"step": 50
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.4291845493562232,
|
| 28 |
+
"grad_norm": 5.86117696762085,
|
| 29 |
+
"learning_rate": 4.228571428571429e-06,
|
| 30 |
+
"loss": 4.403,
|
| 31 |
+
"step": 75
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 0.5722460658082976,
|
| 35 |
+
"grad_norm": 6.91613245010376,
|
| 36 |
+
"learning_rate": 5.6571428571428576e-06,
|
| 37 |
+
"loss": 4.3734,
|
| 38 |
+
"step": 100
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.7153075822603719,
|
| 42 |
+
"grad_norm": 6.6782917976379395,
|
| 43 |
+
"learning_rate": 7.085714285714286e-06,
|
| 44 |
+
"loss": 4.3521,
|
| 45 |
+
"step": 125
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.8583690987124464,
|
| 49 |
+
"grad_norm": 9.129453659057617,
|
| 50 |
+
"learning_rate": 8.514285714285715e-06,
|
| 51 |
+
"loss": 4.2872,
|
| 52 |
+
"step": 150
|
| 53 |
+
},
|
| 54 |
+
{
|
| 55 |
+
"epoch": 1.0,
|
| 56 |
+
"grad_norm": 7.63546895980835,
|
| 57 |
+
"learning_rate": 9.942857142857145e-06,
|
| 58 |
+
"loss": 4.1545,
|
| 59 |
+
"step": 175
|
| 60 |
+
},
|
| 61 |
+
{
|
| 62 |
+
"epoch": 1.0,
|
| 63 |
+
"eval_accuracy": 0.49326145552560646,
|
| 64 |
+
"eval_loss": 1.0274596214294434,
|
| 65 |
+
"eval_macro_f1": 0.43124266727220806,
|
| 66 |
+
"eval_runtime": 1.5939,
|
| 67 |
+
"eval_samples_per_second": 465.511,
|
| 68 |
+
"eval_steps_per_second": 29.487,
|
| 69 |
+
"eval_weighted_f1": 0.43703064131624236,
|
| 70 |
+
"step": 175
|
| 71 |
+
},
|
| 72 |
+
{
|
| 73 |
+
"epoch": 1.1430615164520743,
|
| 74 |
+
"grad_norm": 22.98823356628418,
|
| 75 |
+
"learning_rate": 1.1371428571428573e-05,
|
| 76 |
+
"loss": 4.0057,
|
| 77 |
+
"step": 200
|
| 78 |
+
},
|
| 79 |
+
{
|
| 80 |
+
"epoch": 1.2861230329041489,
|
| 81 |
+
"grad_norm": 39.52901077270508,
|
| 82 |
+
"learning_rate": 1.2800000000000001e-05,
|
| 83 |
+
"loss": 3.6023,
|
| 84 |
+
"step": 225
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"epoch": 1.4291845493562232,
|
| 88 |
+
"grad_norm": 35.6060791015625,
|
| 89 |
+
"learning_rate": 1.422857142857143e-05,
|
| 90 |
+
"loss": 3.3551,
|
| 91 |
+
"step": 250
|
| 92 |
+
},
|
| 93 |
+
{
|
| 94 |
+
"epoch": 1.5722460658082977,
|
| 95 |
+
"grad_norm": 54.77183151245117,
|
| 96 |
+
"learning_rate": 1.5657142857142856e-05,
|
| 97 |
+
"loss": 2.8497,
|
| 98 |
+
"step": 275
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 1.715307582260372,
|
| 102 |
+
"grad_norm": 52.70783615112305,
|
| 103 |
+
"learning_rate": 1.708571428571429e-05,
|
| 104 |
+
"loss": 2.4412,
|
| 105 |
+
"step": 300
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 1.8583690987124464,
|
| 109 |
+
"grad_norm": 34.87353515625,
|
| 110 |
+
"learning_rate": 1.8514285714285716e-05,
|
| 111 |
+
"loss": 2.2117,
|
| 112 |
+
"step": 325
|
| 113 |
+
},
|
| 114 |
+
{
|
| 115 |
+
"epoch": 2.0,
|
| 116 |
+
"grad_norm": 42.81010055541992,
|
| 117 |
+
"learning_rate": 1.9942857142857142e-05,
|
| 118 |
+
"loss": 2.1342,
|
| 119 |
+
"step": 350
|
| 120 |
+
},
|
| 121 |
+
{
|
| 122 |
+
"epoch": 2.0,
|
| 123 |
+
"eval_accuracy": 0.7776280323450134,
|
| 124 |
+
"eval_loss": 0.5863093137741089,
|
| 125 |
+
"eval_macro_f1": 0.7763857523713211,
|
| 126 |
+
"eval_runtime": 1.5314,
|
| 127 |
+
"eval_samples_per_second": 484.529,
|
| 128 |
+
"eval_steps_per_second": 30.691,
|
| 129 |
+
"eval_weighted_f1": 0.7768329835985338,
|
| 130 |
+
"step": 350
|
| 131 |
+
},
|
| 132 |
+
{
|
| 133 |
+
"epoch": 2.1430615164520743,
|
| 134 |
+
"grad_norm": 92.29073333740234,
|
| 135 |
+
"learning_rate": 1.984761904761905e-05,
|
| 136 |
+
"loss": 1.8224,
|
| 137 |
+
"step": 375
|
| 138 |
+
},
|
| 139 |
+
{
|
| 140 |
+
"epoch": 2.2861230329041486,
|
| 141 |
+
"grad_norm": 45.62285614013672,
|
| 142 |
+
"learning_rate": 1.968888888888889e-05,
|
| 143 |
+
"loss": 1.7858,
|
| 144 |
+
"step": 400
|
| 145 |
+
},
|
| 146 |
+
{
|
| 147 |
+
"epoch": 2.429184549356223,
|
| 148 |
+
"grad_norm": 43.795448303222656,
|
| 149 |
+
"learning_rate": 1.953015873015873e-05,
|
| 150 |
+
"loss": 1.8205,
|
| 151 |
+
"step": 425
|
| 152 |
+
},
|
| 153 |
+
{
|
| 154 |
+
"epoch": 2.5722460658082977,
|
| 155 |
+
"grad_norm": 22.248594284057617,
|
| 156 |
+
"learning_rate": 1.937142857142857e-05,
|
| 157 |
+
"loss": 1.6616,
|
| 158 |
+
"step": 450
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 2.715307582260372,
|
| 162 |
+
"grad_norm": 39.29319763183594,
|
| 163 |
+
"learning_rate": 1.9212698412698414e-05,
|
| 164 |
+
"loss": 1.4153,
|
| 165 |
+
"step": 475
|
| 166 |
+
},
|
| 167 |
+
{
|
| 168 |
+
"epoch": 2.8583690987124464,
|
| 169 |
+
"grad_norm": 58.74412155151367,
|
| 170 |
+
"learning_rate": 1.9053968253968257e-05,
|
| 171 |
+
"loss": 1.4515,
|
| 172 |
+
"step": 500
|
| 173 |
+
},
|
| 174 |
+
{
|
| 175 |
+
"epoch": 3.0,
|
| 176 |
+
"grad_norm": 39.221473693847656,
|
| 177 |
+
"learning_rate": 1.8895238095238096e-05,
|
| 178 |
+
"loss": 1.3692,
|
| 179 |
+
"step": 525
|
| 180 |
+
},
|
| 181 |
+
{
|
| 182 |
+
"epoch": 3.0,
|
| 183 |
+
"eval_accuracy": 0.8207547169811321,
|
| 184 |
+
"eval_loss": 0.48869648575782776,
|
| 185 |
+
"eval_macro_f1": 0.8206853516657854,
|
| 186 |
+
"eval_runtime": 1.5123,
|
| 187 |
+
"eval_samples_per_second": 490.63,
|
| 188 |
+
"eval_steps_per_second": 31.078,
|
| 189 |
+
"eval_weighted_f1": 0.8207046685890464,
|
| 190 |
+
"step": 525
|
| 191 |
+
},
|
| 192 |
+
{
|
| 193 |
+
"epoch": 3.1430615164520743,
|
| 194 |
+
"grad_norm": 21.472152709960938,
|
| 195 |
+
"learning_rate": 1.8736507936507936e-05,
|
| 196 |
+
"loss": 0.9876,
|
| 197 |
+
"step": 550
|
| 198 |
+
},
|
| 199 |
+
{
|
| 200 |
+
"epoch": 3.2861230329041486,
|
| 201 |
+
"grad_norm": 58.34779357910156,
|
| 202 |
+
"learning_rate": 1.857777777777778e-05,
|
| 203 |
+
"loss": 1.0658,
|
| 204 |
+
"step": 575
|
| 205 |
+
},
|
| 206 |
+
{
|
| 207 |
+
"epoch": 3.429184549356223,
|
| 208 |
+
"grad_norm": 39.19213104248047,
|
| 209 |
+
"learning_rate": 1.8419047619047622e-05,
|
| 210 |
+
"loss": 1.0417,
|
| 211 |
+
"step": 600
|
| 212 |
+
},
|
| 213 |
+
{
|
| 214 |
+
"epoch": 3.5722460658082977,
|
| 215 |
+
"grad_norm": 27.044824600219727,
|
| 216 |
+
"learning_rate": 1.8260317460317462e-05,
|
| 217 |
+
"loss": 1.1198,
|
| 218 |
+
"step": 625
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 3.715307582260372,
|
| 222 |
+
"grad_norm": 50.7132682800293,
|
| 223 |
+
"learning_rate": 1.81015873015873e-05,
|
| 224 |
+
"loss": 1.041,
|
| 225 |
+
"step": 650
|
| 226 |
+
},
|
| 227 |
+
{
|
| 228 |
+
"epoch": 3.8583690987124464,
|
| 229 |
+
"grad_norm": 60.58425521850586,
|
| 230 |
+
"learning_rate": 1.7942857142857144e-05,
|
| 231 |
+
"loss": 1.1259,
|
| 232 |
+
"step": 675
|
| 233 |
+
},
|
| 234 |
+
{
|
| 235 |
+
"epoch": 4.0,
|
| 236 |
+
"grad_norm": 39.099082946777344,
|
| 237 |
+
"learning_rate": 1.7784126984126987e-05,
|
| 238 |
+
"loss": 1.0999,
|
| 239 |
+
"step": 700
|
| 240 |
+
},
|
| 241 |
+
{
|
| 242 |
+
"epoch": 4.0,
|
| 243 |
+
"eval_accuracy": 0.8328840970350404,
|
| 244 |
+
"eval_loss": 0.47056081891059875,
|
| 245 |
+
"eval_macro_f1": 0.8318811799298261,
|
| 246 |
+
"eval_runtime": 1.5558,
|
| 247 |
+
"eval_samples_per_second": 476.93,
|
| 248 |
+
"eval_steps_per_second": 30.21,
|
| 249 |
+
"eval_weighted_f1": 0.8323409074500697,
|
| 250 |
+
"step": 700
|
| 251 |
+
},
|
| 252 |
+
{
|
| 253 |
+
"epoch": 4.143061516452074,
|
| 254 |
+
"grad_norm": 64.90669250488281,
|
| 255 |
+
"learning_rate": 1.7625396825396827e-05,
|
| 256 |
+
"loss": 0.9448,
|
| 257 |
+
"step": 725
|
| 258 |
+
},
|
| 259 |
+
{
|
| 260 |
+
"epoch": 4.286123032904149,
|
| 261 |
+
"grad_norm": 16.821197509765625,
|
| 262 |
+
"learning_rate": 1.7466666666666667e-05,
|
| 263 |
+
"loss": 0.8834,
|
| 264 |
+
"step": 750
|
| 265 |
+
},
|
| 266 |
+
{
|
| 267 |
+
"epoch": 4.429184549356223,
|
| 268 |
+
"grad_norm": 11.424208641052246,
|
| 269 |
+
"learning_rate": 1.730793650793651e-05,
|
| 270 |
+
"loss": 0.7242,
|
| 271 |
+
"step": 775
|
| 272 |
+
},
|
| 273 |
+
{
|
| 274 |
+
"epoch": 4.572246065808297,
|
| 275 |
+
"grad_norm": 104.45195770263672,
|
| 276 |
+
"learning_rate": 1.7149206349206353e-05,
|
| 277 |
+
"loss": 0.7625,
|
| 278 |
+
"step": 800
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 4.715307582260372,
|
| 282 |
+
"grad_norm": 47.22216796875,
|
| 283 |
+
"learning_rate": 1.6990476190476192e-05,
|
| 284 |
+
"loss": 0.6285,
|
| 285 |
+
"step": 825
|
| 286 |
+
},
|
| 287 |
+
{
|
| 288 |
+
"epoch": 4.858369098712446,
|
| 289 |
+
"grad_norm": 0.7487726211547852,
|
| 290 |
+
"learning_rate": 1.6831746031746032e-05,
|
| 291 |
+
"loss": 0.459,
|
| 292 |
+
"step": 850
|
| 293 |
+
},
|
| 294 |
+
{
|
| 295 |
+
"epoch": 5.0,
|
| 296 |
+
"grad_norm": 79.56040954589844,
|
| 297 |
+
"learning_rate": 1.6673015873015875e-05,
|
| 298 |
+
"loss": 0.7389,
|
| 299 |
+
"step": 875
|
| 300 |
+
},
|
| 301 |
+
{
|
| 302 |
+
"epoch": 5.0,
|
| 303 |
+
"eval_accuracy": 0.8274932614555256,
|
| 304 |
+
"eval_loss": 0.6653628945350647,
|
| 305 |
+
"eval_macro_f1": 0.8272892791775878,
|
| 306 |
+
"eval_runtime": 1.5643,
|
| 307 |
+
"eval_samples_per_second": 474.334,
|
| 308 |
+
"eval_steps_per_second": 30.045,
|
| 309 |
+
"eval_weighted_f1": 0.8270613133933576,
|
| 310 |
+
"step": 875
|
| 311 |
+
},
|
| 312 |
+
{
|
| 313 |
+
"epoch": 5.143061516452074,
|
| 314 |
+
"grad_norm": 18.06866455078125,
|
| 315 |
+
"learning_rate": 1.6514285714285714e-05,
|
| 316 |
+
"loss": 0.5253,
|
| 317 |
+
"step": 900
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 5.286123032904149,
|
| 321 |
+
"grad_norm": 12.360490798950195,
|
| 322 |
+
"learning_rate": 1.6355555555555557e-05,
|
| 323 |
+
"loss": 0.5756,
|
| 324 |
+
"step": 925
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"epoch": 5.429184549356223,
|
| 328 |
+
"grad_norm": 72.94513702392578,
|
| 329 |
+
"learning_rate": 1.6196825396825397e-05,
|
| 330 |
+
"loss": 0.4826,
|
| 331 |
+
"step": 950
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 5.572246065808297,
|
| 335 |
+
"grad_norm": 243.04159545898438,
|
| 336 |
+
"learning_rate": 1.603809523809524e-05,
|
| 337 |
+
"loss": 0.4979,
|
| 338 |
+
"step": 975
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 5.715307582260372,
|
| 342 |
+
"grad_norm": 103.33642578125,
|
| 343 |
+
"learning_rate": 1.587936507936508e-05,
|
| 344 |
+
"loss": 0.6762,
|
| 345 |
+
"step": 1000
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 5.858369098712446,
|
| 349 |
+
"grad_norm": 47.4416389465332,
|
| 350 |
+
"learning_rate": 1.5720634920634923e-05,
|
| 351 |
+
"loss": 0.5239,
|
| 352 |
+
"step": 1025
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 6.0,
|
| 356 |
+
"grad_norm": 0.9615633487701416,
|
| 357 |
+
"learning_rate": 1.5561904761904762e-05,
|
| 358 |
+
"loss": 0.4068,
|
| 359 |
+
"step": 1050
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 6.0,
|
| 363 |
+
"eval_accuracy": 0.8247978436657682,
|
| 364 |
+
"eval_loss": 0.7345940470695496,
|
| 365 |
+
"eval_macro_f1": 0.8224694249284413,
|
| 366 |
+
"eval_runtime": 1.4985,
|
| 367 |
+
"eval_samples_per_second": 495.173,
|
| 368 |
+
"eval_steps_per_second": 31.365,
|
| 369 |
+
"eval_weighted_f1": 0.822666601437977,
|
| 370 |
+
"step": 1050
|
| 371 |
+
},
|
| 372 |
+
{
|
| 373 |
+
"epoch": 6.143061516452074,
|
| 374 |
+
"grad_norm": 14.029003143310547,
|
| 375 |
+
"learning_rate": 1.5403174603174605e-05,
|
| 376 |
+
"loss": 0.4339,
|
| 377 |
+
"step": 1075
|
| 378 |
+
},
|
| 379 |
+
{
|
| 380 |
+
"epoch": 6.286123032904149,
|
| 381 |
+
"grad_norm": 43.50464630126953,
|
| 382 |
+
"learning_rate": 1.5244444444444447e-05,
|
| 383 |
+
"loss": 0.3476,
|
| 384 |
+
"step": 1100
|
| 385 |
+
},
|
| 386 |
+
{
|
| 387 |
+
"epoch": 6.429184549356223,
|
| 388 |
+
"grad_norm": 40.38178253173828,
|
| 389 |
+
"learning_rate": 1.5085714285714288e-05,
|
| 390 |
+
"loss": 0.3248,
|
| 391 |
+
"step": 1125
|
| 392 |
+
},
|
| 393 |
+
{
|
| 394 |
+
"epoch": 6.572246065808297,
|
| 395 |
+
"grad_norm": 7.844557762145996,
|
| 396 |
+
"learning_rate": 1.4926984126984128e-05,
|
| 397 |
+
"loss": 0.4011,
|
| 398 |
+
"step": 1150
|
| 399 |
+
},
|
| 400 |
+
{
|
| 401 |
+
"epoch": 6.715307582260372,
|
| 402 |
+
"grad_norm": 18.742788314819336,
|
| 403 |
+
"learning_rate": 1.4768253968253969e-05,
|
| 404 |
+
"loss": 0.4739,
|
| 405 |
+
"step": 1175
|
| 406 |
+
},
|
| 407 |
+
{
|
| 408 |
+
"epoch": 6.858369098712446,
|
| 409 |
+
"grad_norm": 15.948999404907227,
|
| 410 |
+
"learning_rate": 1.4609523809523812e-05,
|
| 411 |
+
"loss": 0.5922,
|
| 412 |
+
"step": 1200
|
| 413 |
+
},
|
| 414 |
+
{
|
| 415 |
+
"epoch": 7.0,
|
| 416 |
+
"grad_norm": 9.971810340881348,
|
| 417 |
+
"learning_rate": 1.4450793650793653e-05,
|
| 418 |
+
"loss": 0.4831,
|
| 419 |
+
"step": 1225
|
| 420 |
+
},
|
| 421 |
+
{
|
| 422 |
+
"epoch": 7.0,
|
| 423 |
+
"eval_accuracy": 0.8288409703504043,
|
| 424 |
+
"eval_loss": 0.7552890181541443,
|
| 425 |
+
"eval_macro_f1": 0.8283480411819761,
|
| 426 |
+
"eval_runtime": 1.5405,
|
| 427 |
+
"eval_samples_per_second": 481.668,
|
| 428 |
+
"eval_steps_per_second": 30.51,
|
| 429 |
+
"eval_weighted_f1": 0.8287061751340411,
|
| 430 |
+
"step": 1225
|
| 431 |
+
}
|
| 432 |
+
],
|
| 433 |
+
"logging_steps": 25,
|
| 434 |
+
"max_steps": 3500,
|
| 435 |
+
"num_input_tokens_seen": 0,
|
| 436 |
+
"num_train_epochs": 20,
|
| 437 |
+
"save_steps": 500,
|
| 438 |
+
"stateful_callbacks": {
|
| 439 |
+
"EarlyStoppingCallback": {
|
| 440 |
+
"args": {
|
| 441 |
+
"early_stopping_patience": 3,
|
| 442 |
+
"early_stopping_threshold": 0.0
|
| 443 |
+
},
|
| 444 |
+
"attributes": {
|
| 445 |
+
"early_stopping_patience_counter": 3
|
| 446 |
+
}
|
| 447 |
+
},
|
| 448 |
+
"TrainerControl": {
|
| 449 |
+
"args": {
|
| 450 |
+
"should_epoch_stop": false,
|
| 451 |
+
"should_evaluate": false,
|
| 452 |
+
"should_log": false,
|
| 453 |
+
"should_save": true,
|
| 454 |
+
"should_training_stop": true
|
| 455 |
+
},
|
| 456 |
+
"attributes": {}
|
| 457 |
+
}
|
| 458 |
+
},
|
| 459 |
+
"total_flos": 0.0,
|
| 460 |
+
"train_batch_size": 8,
|
| 461 |
+
"trial_name": null,
|
| 462 |
+
"trial_params": null
|
| 463 |
+
}
|
ood-vifactcheck-phobert-topic-domain-05-s43-20260814T143211Z-bdd2c308/trainer/checkpoint-1225/vocab.txt
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-vifactcheck-phobert-topic-domain-05-s43-20260814T143211Z-bdd2c308/trainer/checkpoint-700/added_tokens.json
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"<mask>": 64000
|
| 3 |
+
}
|
ood-vifactcheck-phobert-topic-domain-05-s43-20260814T143211Z-bdd2c308/trainer/checkpoint-700/bpe.codes
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-vifactcheck-phobert-topic-domain-05-s43-20260814T143211Z-bdd2c308/trainer/checkpoint-700/special_tokens_map.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token": "<s>",
|
| 3 |
+
"cls_token": "<s>",
|
| 4 |
+
"eos_token": "</s>",
|
| 5 |
+
"mask_token": "<mask>",
|
| 6 |
+
"pad_token": "<pad>",
|
| 7 |
+
"sep_token": "</s>",
|
| 8 |
+
"unk_token": "<unk>"
|
| 9 |
+
}
|
ood-vifactcheck-phobert-topic-domain-05-s43-20260814T143211Z-bdd2c308/trainer/checkpoint-700/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 |
+
"64000": {
|
| 36 |
+
"content": "<mask>",
|
| 37 |
+
"lstrip": false,
|
| 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": 256,
|
| 51 |
+
"pad_token": "<pad>",
|
| 52 |
+
"sep_token": "</s>",
|
| 53 |
+
"tokenizer_class": "PhobertTokenizer",
|
| 54 |
+
"unk_token": "<unk>"
|
| 55 |
+
}
|
ood-vifactcheck-phobert-topic-domain-05-s43-20260814T143211Z-bdd2c308/trainer/checkpoint-700/trainer_state.json
ADDED
|
@@ -0,0 +1,283 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": 700,
|
| 3 |
+
"best_metric": 0.8318811799298261,
|
| 4 |
+
"best_model_checkpoint": "/root/VNEncoder/results/checkpoints/ood-vifactcheck-phobert-topic-domain-05-s43-20260814T143211Z-bdd2c308/trainer/checkpoint-700",
|
| 5 |
+
"epoch": 4.0,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 700,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"epoch": 0.1430615164520744,
|
| 14 |
+
"grad_norm": 9.187152862548828,
|
| 15 |
+
"learning_rate": 1.3714285714285717e-06,
|
| 16 |
+
"loss": 4.4471,
|
| 17 |
+
"step": 25
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.2861230329041488,
|
| 21 |
+
"grad_norm": 6.098942279815674,
|
| 22 |
+
"learning_rate": 2.8000000000000003e-06,
|
| 23 |
+
"loss": 4.418,
|
| 24 |
+
"step": 50
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.4291845493562232,
|
| 28 |
+
"grad_norm": 5.86117696762085,
|
| 29 |
+
"learning_rate": 4.228571428571429e-06,
|
| 30 |
+
"loss": 4.403,
|
| 31 |
+
"step": 75
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 0.5722460658082976,
|
| 35 |
+
"grad_norm": 6.91613245010376,
|
| 36 |
+
"learning_rate": 5.6571428571428576e-06,
|
| 37 |
+
"loss": 4.3734,
|
| 38 |
+
"step": 100
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.7153075822603719,
|
| 42 |
+
"grad_norm": 6.6782917976379395,
|
| 43 |
+
"learning_rate": 7.085714285714286e-06,
|
| 44 |
+
"loss": 4.3521,
|
| 45 |
+
"step": 125
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.8583690987124464,
|
| 49 |
+
"grad_norm": 9.129453659057617,
|
| 50 |
+
"learning_rate": 8.514285714285715e-06,
|
| 51 |
+
"loss": 4.2872,
|
| 52 |
+
"step": 150
|
| 53 |
+
},
|
| 54 |
+
{
|
| 55 |
+
"epoch": 1.0,
|
| 56 |
+
"grad_norm": 7.63546895980835,
|
| 57 |
+
"learning_rate": 9.942857142857145e-06,
|
| 58 |
+
"loss": 4.1545,
|
| 59 |
+
"step": 175
|
| 60 |
+
},
|
| 61 |
+
{
|
| 62 |
+
"epoch": 1.0,
|
| 63 |
+
"eval_accuracy": 0.49326145552560646,
|
| 64 |
+
"eval_loss": 1.0274596214294434,
|
| 65 |
+
"eval_macro_f1": 0.43124266727220806,
|
| 66 |
+
"eval_runtime": 1.5939,
|
| 67 |
+
"eval_samples_per_second": 465.511,
|
| 68 |
+
"eval_steps_per_second": 29.487,
|
| 69 |
+
"eval_weighted_f1": 0.43703064131624236,
|
| 70 |
+
"step": 175
|
| 71 |
+
},
|
| 72 |
+
{
|
| 73 |
+
"epoch": 1.1430615164520743,
|
| 74 |
+
"grad_norm": 22.98823356628418,
|
| 75 |
+
"learning_rate": 1.1371428571428573e-05,
|
| 76 |
+
"loss": 4.0057,
|
| 77 |
+
"step": 200
|
| 78 |
+
},
|
| 79 |
+
{
|
| 80 |
+
"epoch": 1.2861230329041489,
|
| 81 |
+
"grad_norm": 39.52901077270508,
|
| 82 |
+
"learning_rate": 1.2800000000000001e-05,
|
| 83 |
+
"loss": 3.6023,
|
| 84 |
+
"step": 225
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"epoch": 1.4291845493562232,
|
| 88 |
+
"grad_norm": 35.6060791015625,
|
| 89 |
+
"learning_rate": 1.422857142857143e-05,
|
| 90 |
+
"loss": 3.3551,
|
| 91 |
+
"step": 250
|
| 92 |
+
},
|
| 93 |
+
{
|
| 94 |
+
"epoch": 1.5722460658082977,
|
| 95 |
+
"grad_norm": 54.77183151245117,
|
| 96 |
+
"learning_rate": 1.5657142857142856e-05,
|
| 97 |
+
"loss": 2.8497,
|
| 98 |
+
"step": 275
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 1.715307582260372,
|
| 102 |
+
"grad_norm": 52.70783615112305,
|
| 103 |
+
"learning_rate": 1.708571428571429e-05,
|
| 104 |
+
"loss": 2.4412,
|
| 105 |
+
"step": 300
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 1.8583690987124464,
|
| 109 |
+
"grad_norm": 34.87353515625,
|
| 110 |
+
"learning_rate": 1.8514285714285716e-05,
|
| 111 |
+
"loss": 2.2117,
|
| 112 |
+
"step": 325
|
| 113 |
+
},
|
| 114 |
+
{
|
| 115 |
+
"epoch": 2.0,
|
| 116 |
+
"grad_norm": 42.81010055541992,
|
| 117 |
+
"learning_rate": 1.9942857142857142e-05,
|
| 118 |
+
"loss": 2.1342,
|
| 119 |
+
"step": 350
|
| 120 |
+
},
|
| 121 |
+
{
|
| 122 |
+
"epoch": 2.0,
|
| 123 |
+
"eval_accuracy": 0.7776280323450134,
|
| 124 |
+
"eval_loss": 0.5863093137741089,
|
| 125 |
+
"eval_macro_f1": 0.7763857523713211,
|
| 126 |
+
"eval_runtime": 1.5314,
|
| 127 |
+
"eval_samples_per_second": 484.529,
|
| 128 |
+
"eval_steps_per_second": 30.691,
|
| 129 |
+
"eval_weighted_f1": 0.7768329835985338,
|
| 130 |
+
"step": 350
|
| 131 |
+
},
|
| 132 |
+
{
|
| 133 |
+
"epoch": 2.1430615164520743,
|
| 134 |
+
"grad_norm": 92.29073333740234,
|
| 135 |
+
"learning_rate": 1.984761904761905e-05,
|
| 136 |
+
"loss": 1.8224,
|
| 137 |
+
"step": 375
|
| 138 |
+
},
|
| 139 |
+
{
|
| 140 |
+
"epoch": 2.2861230329041486,
|
| 141 |
+
"grad_norm": 45.62285614013672,
|
| 142 |
+
"learning_rate": 1.968888888888889e-05,
|
| 143 |
+
"loss": 1.7858,
|
| 144 |
+
"step": 400
|
| 145 |
+
},
|
| 146 |
+
{
|
| 147 |
+
"epoch": 2.429184549356223,
|
| 148 |
+
"grad_norm": 43.795448303222656,
|
| 149 |
+
"learning_rate": 1.953015873015873e-05,
|
| 150 |
+
"loss": 1.8205,
|
| 151 |
+
"step": 425
|
| 152 |
+
},
|
| 153 |
+
{
|
| 154 |
+
"epoch": 2.5722460658082977,
|
| 155 |
+
"grad_norm": 22.248594284057617,
|
| 156 |
+
"learning_rate": 1.937142857142857e-05,
|
| 157 |
+
"loss": 1.6616,
|
| 158 |
+
"step": 450
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 2.715307582260372,
|
| 162 |
+
"grad_norm": 39.29319763183594,
|
| 163 |
+
"learning_rate": 1.9212698412698414e-05,
|
| 164 |
+
"loss": 1.4153,
|
| 165 |
+
"step": 475
|
| 166 |
+
},
|
| 167 |
+
{
|
| 168 |
+
"epoch": 2.8583690987124464,
|
| 169 |
+
"grad_norm": 58.74412155151367,
|
| 170 |
+
"learning_rate": 1.9053968253968257e-05,
|
| 171 |
+
"loss": 1.4515,
|
| 172 |
+
"step": 500
|
| 173 |
+
},
|
| 174 |
+
{
|
| 175 |
+
"epoch": 3.0,
|
| 176 |
+
"grad_norm": 39.221473693847656,
|
| 177 |
+
"learning_rate": 1.8895238095238096e-05,
|
| 178 |
+
"loss": 1.3692,
|
| 179 |
+
"step": 525
|
| 180 |
+
},
|
| 181 |
+
{
|
| 182 |
+
"epoch": 3.0,
|
| 183 |
+
"eval_accuracy": 0.8207547169811321,
|
| 184 |
+
"eval_loss": 0.48869648575782776,
|
| 185 |
+
"eval_macro_f1": 0.8206853516657854,
|
| 186 |
+
"eval_runtime": 1.5123,
|
| 187 |
+
"eval_samples_per_second": 490.63,
|
| 188 |
+
"eval_steps_per_second": 31.078,
|
| 189 |
+
"eval_weighted_f1": 0.8207046685890464,
|
| 190 |
+
"step": 525
|
| 191 |
+
},
|
| 192 |
+
{
|
| 193 |
+
"epoch": 3.1430615164520743,
|
| 194 |
+
"grad_norm": 21.472152709960938,
|
| 195 |
+
"learning_rate": 1.8736507936507936e-05,
|
| 196 |
+
"loss": 0.9876,
|
| 197 |
+
"step": 550
|
| 198 |
+
},
|
| 199 |
+
{
|
| 200 |
+
"epoch": 3.2861230329041486,
|
| 201 |
+
"grad_norm": 58.34779357910156,
|
| 202 |
+
"learning_rate": 1.857777777777778e-05,
|
| 203 |
+
"loss": 1.0658,
|
| 204 |
+
"step": 575
|
| 205 |
+
},
|
| 206 |
+
{
|
| 207 |
+
"epoch": 3.429184549356223,
|
| 208 |
+
"grad_norm": 39.19213104248047,
|
| 209 |
+
"learning_rate": 1.8419047619047622e-05,
|
| 210 |
+
"loss": 1.0417,
|
| 211 |
+
"step": 600
|
| 212 |
+
},
|
| 213 |
+
{
|
| 214 |
+
"epoch": 3.5722460658082977,
|
| 215 |
+
"grad_norm": 27.044824600219727,
|
| 216 |
+
"learning_rate": 1.8260317460317462e-05,
|
| 217 |
+
"loss": 1.1198,
|
| 218 |
+
"step": 625
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 3.715307582260372,
|
| 222 |
+
"grad_norm": 50.7132682800293,
|
| 223 |
+
"learning_rate": 1.81015873015873e-05,
|
| 224 |
+
"loss": 1.041,
|
| 225 |
+
"step": 650
|
| 226 |
+
},
|
| 227 |
+
{
|
| 228 |
+
"epoch": 3.8583690987124464,
|
| 229 |
+
"grad_norm": 60.58425521850586,
|
| 230 |
+
"learning_rate": 1.7942857142857144e-05,
|
| 231 |
+
"loss": 1.1259,
|
| 232 |
+
"step": 675
|
| 233 |
+
},
|
| 234 |
+
{
|
| 235 |
+
"epoch": 4.0,
|
| 236 |
+
"grad_norm": 39.099082946777344,
|
| 237 |
+
"learning_rate": 1.7784126984126987e-05,
|
| 238 |
+
"loss": 1.0999,
|
| 239 |
+
"step": 700
|
| 240 |
+
},
|
| 241 |
+
{
|
| 242 |
+
"epoch": 4.0,
|
| 243 |
+
"eval_accuracy": 0.8328840970350404,
|
| 244 |
+
"eval_loss": 0.47056081891059875,
|
| 245 |
+
"eval_macro_f1": 0.8318811799298261,
|
| 246 |
+
"eval_runtime": 1.5558,
|
| 247 |
+
"eval_samples_per_second": 476.93,
|
| 248 |
+
"eval_steps_per_second": 30.21,
|
| 249 |
+
"eval_weighted_f1": 0.8323409074500697,
|
| 250 |
+
"step": 700
|
| 251 |
+
}
|
| 252 |
+
],
|
| 253 |
+
"logging_steps": 25,
|
| 254 |
+
"max_steps": 3500,
|
| 255 |
+
"num_input_tokens_seen": 0,
|
| 256 |
+
"num_train_epochs": 20,
|
| 257 |
+
"save_steps": 500,
|
| 258 |
+
"stateful_callbacks": {
|
| 259 |
+
"EarlyStoppingCallback": {
|
| 260 |
+
"args": {
|
| 261 |
+
"early_stopping_patience": 3,
|
| 262 |
+
"early_stopping_threshold": 0.0
|
| 263 |
+
},
|
| 264 |
+
"attributes": {
|
| 265 |
+
"early_stopping_patience_counter": 0
|
| 266 |
+
}
|
| 267 |
+
},
|
| 268 |
+
"TrainerControl": {
|
| 269 |
+
"args": {
|
| 270 |
+
"should_epoch_stop": false,
|
| 271 |
+
"should_evaluate": false,
|
| 272 |
+
"should_log": false,
|
| 273 |
+
"should_save": true,
|
| 274 |
+
"should_training_stop": false
|
| 275 |
+
},
|
| 276 |
+
"attributes": {}
|
| 277 |
+
}
|
| 278 |
+
},
|
| 279 |
+
"total_flos": 0.0,
|
| 280 |
+
"train_batch_size": 8,
|
| 281 |
+
"trial_name": null,
|
| 282 |
+
"trial_params": null
|
| 283 |
+
}
|
ood-vifactcheck-phobert-topic-domain-05-s43-20260814T143211Z-bdd2c308/trainer/checkpoint-700/vocab.txt
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|