Hugging Face's logo

arielcerdap
/
modernbert-disfluency-expE-large-oversample-rv

Token Classification
TensorBoard
Safetensors
English
modernbert
disfluency-detection
speech-pathology
Model card Files
xet
 Metrics Community
arielcerdap commited on
Commit
e3a1dbf
·
verified ·
1 Parent(s): 57d1fdd

Upload expE_large_oversample_rv — RV F1=0.4884 Macro=0.8253

Browse files
This view is limited to 50 files because it contains too many changes.   See raw diff
Files changed (50) hide show
  1. .gitattributes +1 -0
  2. README.md +53 -0
  3. checkpoint-1250/config.json +99 -0
  4. checkpoint-1250/model.safetensors +3 -0
  5. checkpoint-1250/optimizer.pt +3 -0
  6. checkpoint-1250/rng_state.pth +3 -0
  7. checkpoint-1250/scheduler.pt +3 -0
  8. checkpoint-1250/tokenizer.json +0 -0
  9. checkpoint-1250/tokenizer_config.json +16 -0
  10. checkpoint-1250/trainer_state.json +1078 -0
  11. checkpoint-1250/training_args.bin +3 -0
  12. checkpoint-1300/config.json +99 -0
  13. checkpoint-1300/model.safetensors +3 -0
  14. checkpoint-1300/optimizer.pt +3 -0
  15. checkpoint-1300/rng_state.pth +3 -0
  16. checkpoint-1300/scheduler.pt +3 -0
  17. checkpoint-1300/tokenizer.json +0 -0
  18. checkpoint-1300/tokenizer_config.json +16 -0
  19. checkpoint-1300/trainer_state.json +1118 -0
  20. checkpoint-1300/training_args.bin +3 -0
  21. checkpoint-1350/config.json +99 -0
  22. checkpoint-1350/model.safetensors +3 -0
  23. checkpoint-1350/optimizer.pt +3 -0
  24. checkpoint-1350/rng_state.pth +3 -0
  25. checkpoint-1350/scheduler.pt +3 -0
  26. checkpoint-1350/tokenizer.json +0 -0
  27. checkpoint-1350/tokenizer_config.json +16 -0
  28. checkpoint-1350/trainer_state.json +1158 -0
  29. checkpoint-1350/training_args.bin +3 -0
  30. checkpoint-1400/config.json +99 -0
  31. checkpoint-1400/model.safetensors +3 -0
  32. checkpoint-1400/optimizer.pt +3 -0
  33. checkpoint-1400/rng_state.pth +3 -0
  34. checkpoint-1400/scheduler.pt +3 -0
  35. checkpoint-1400/tokenizer.json +0 -0
  36. checkpoint-1400/tokenizer_config.json +16 -0
  37. checkpoint-1400/trainer_state.json +1198 -0
  38. checkpoint-1400/training_args.bin +3 -0
  39. checkpoint-900/config.json +99 -0
  40. checkpoint-900/model.safetensors +3 -0
  41. checkpoint-900/optimizer.pt +3 -0
  42. checkpoint-900/rng_state.pth +3 -0
  43. checkpoint-900/scheduler.pt +3 -0
  44. checkpoint-900/tokenizer.json +0 -0
  45. checkpoint-900/tokenizer_config.json +16 -0
  46. checkpoint-900/trainer_state.json +791 -0
  47. checkpoint-900/training_args.bin +3 -0
  48. config.json +99 -0
  49. confusion_matrix.png +3 -0
  50. model.safetensors +3 -0
.gitattributes CHANGED
@@ -33,3 +33,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
33
  *.zip filter=lfs diff=lfs merge=lfs -text
34
  *.zst filter=lfs diff=lfs merge=lfs -text
35
  *tfevents* filter=lfs diff=lfs merge=lfs -text
 
 
33
  *.zip filter=lfs diff=lfs merge=lfs -text
34
  *.zst filter=lfs diff=lfs merge=lfs -text
35
  *tfevents* filter=lfs diff=lfs merge=lfs -text
36
+ confusion_matrix.png filter=lfs diff=lfs merge=lfs -text
README.md ADDED
@@ -0,0 +1,53 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ language: en
3
+ tags:
4
+ - disfluency-detection
5
+ - token-classification
6
+ - modernbert
7
+ - speech-pathology
8
+ datasets:
9
+ - arielcerdap/disfluency-fluencybank
10
+ ---
11
+
12
+ # ModernBERT-large Disfluency — Exp E (Real + RV Oversample x2)
13
+
14
+ Fine-tuned from [answerdotai/ModernBERT-large](https://huggingface.co/answerdotai/ModernBERT-large).
15
+ Train set = real data with RV-containing segments oversampled 2x.
16
+ No synthetic data. Val/Test identical to all experiments.
17
+
18
+ ## Oversample stats
19
+ - Train original : 2737 segments
20
+ - RV segments : 348 (duplicated 1 extra time(s))
21
+ - Train final : 3085 segments
22
+ - RV tokens after: 1858 (4.21%)
23
+
24
+ ## Comparison Table
25
+ | Label | Paper | Exp A (base) | Exp C (large) | Exp D (large+mix) | Exp E (large+oversample) |
26
+ |---|---|---|---|---|---|
27
+ | FP | 1.000 | 0.9944 | 0.9944 | 0.9944 | 0.9915 |
28
+ | RP | 0.690 | 0.8022 | 0.8964 | 0.7253 | 0.8802 |
29
+ | RV | 0.400 | 0.3145 | 0.4974 | 0.3410 | 0.4884 |
30
+ | PW | 0.830 | 0.8879 | 0.9451 | 0.9348 | 0.9409 |
31
+ | Macro | 0.730 | 0.7497 | 0.8333 | 0.7489 | 0.8253 |
32
+ | Binary F1 | — | 0.8902 | 0.9250 | 0.8459 | 0.9344 |
33
+
34
+ ## Per-class Detail
35
+ | Label | P | R | F1 | Support |
36
+ |---|---|---|---|---|
37
+ | O | 0.9895 | 0.9881 | 0.9888 | 3704 |
38
+ | FP | 0.9832 | 1.0000 | 0.9915 | 176 |
39
+ | RP | 0.8541 | 0.9080 | 0.8802 | 174 |
40
+ | RV | 0.4884 | 0.4884 | 0.4884 | 86 |
41
+ | PW | 0.9598 | 0.9227 | 0.9409 | 233 |
42
+
43
+
44
+ ## Hyperparameters
45
+ - learning_rate: 5e-05
46
+ - batch_size effective: 32 (8 × 4 grad_accum)
47
+ - epochs: 15
48
+ - warmup_steps: 97
49
+ - weight_decay: 0.1
50
+ - classifier_dropout: 0.3
51
+ - focal_loss_gamma: 3.0 (adaptive)
52
+ - class_weights: O=1.0, FP=3.0, RP=6.0, RV=12.0, PW=5.0
53
+ - oversample_factor: 2x RV segments
checkpoint-1250/config.json ADDED
@@ -0,0 +1,99 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "architectures": [
3
+ "ModernBertForTokenClassification"
4
+ ],
5
+ "attention_bias": false,
6
+ "attention_dropout": 0.0,
7
+ "bos_token_id": 50281,
8
+ "classifier_activation": "gelu",
9
+ "classifier_bias": false,
10
+ "classifier_dropout": 0.3,
11
+ "classifier_pooling": "mean",
12
+ "cls_token_id": 50281,
13
+ "decoder_bias": true,
14
+ "deterministic_flash_attn": false,
15
+ "dtype": "float32",
16
+ "embedding_dropout": 0.0,
17
+ "eos_token_id": 50282,
18
+ "global_attn_every_n_layers": 3,
19
+ "gradient_checkpointing": false,
20
+ "hidden_activation": "gelu",
21
+ "hidden_size": 1024,
22
+ "id2label": {
23
+ "0": "O",
24
+ "1": "FP",
25
+ "2": "RP",
26
+ "3": "RV",
27
+ "4": "PW"
28
+ },
29
+ "initializer_cutoff_factor": 2.0,
30
+ "initializer_range": 0.02,
31
+ "intermediate_size": 2624,
32
+ "label2id": {
33
+ "FP": 1,
34
+ "O": 0,
35
+ "PW": 4,
36
+ "RP": 2,
37
+ "RV": 3
38
+ },
39
+ "layer_norm_eps": 1e-05,
40
+ "layer_types": [
41
+ "full_attention",
42
+ "sliding_attention",
43
+ "sliding_attention",
44
+ "full_attention",
45
+ "sliding_attention",
46
+ "sliding_attention",
47
+ "full_attention",
48
+ "sliding_attention",
49
+ "sliding_attention",
50
+ "full_attention",
51
+ "sliding_attention",
52
+ "sliding_attention",
53
+ "full_attention",
54
+ "sliding_attention",
55
+ "sliding_attention",
56
+ "full_attention",
57
+ "sliding_attention",
58
+ "sliding_attention",
59
+ "full_attention",
60
+ "sliding_attention",
61
+ "sliding_attention",
62
+ "full_attention",
63
+ "sliding_attention",
64
+ "sliding_attention",
65
+ "full_attention",
66
+ "sliding_attention",
67
+ "sliding_attention",
68
+ "full_attention"
69
+ ],
70
+ "local_attention": 128,
71
+ "max_position_embeddings": 8192,
72
+ "mlp_bias": false,
73
+ "mlp_dropout": 0.0,
74
+ "model_type": "modernbert",
75
+ "norm_bias": false,
76
+ "norm_eps": 1e-05,
77
+ "num_attention_heads": 16,
78
+ "num_hidden_layers": 28,
79
+ "pad_token_id": 50283,
80
+ "position_embedding_type": "absolute",
81
+ "repad_logits_with_grad": false,
82
+ "rope_parameters": {
83
+ "full_attention": {
84
+ "rope_theta": 160000.0,
85
+ "rope_type": "default"
86
+ },
87
+ "sliding_attention": {
88
+ "rope_theta": 10000.0,
89
+ "rope_type": "default"
90
+ }
91
+ },
92
+ "sep_token_id": 50282,
93
+ "sparse_pred_ignore_index": -100,
94
+ "sparse_prediction": false,
95
+ "tie_word_embeddings": true,
96
+ "transformers_version": "5.0.0",
97
+ "use_cache": false,
98
+ "vocab_size": 50368
99
+ }
checkpoint-1250/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:845477fff3a13c0defcb2861df6578bd2dbe0531192bb483cc574811216db455
3
+ size 1583363940
checkpoint-1250/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3f5d30d428915c1643b5306a321e4573b19de870ed7bcb2b3cc840177ce3296e
3
+ size 1112419851
checkpoint-1250/rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0e9a8f7c575b6bf96e1ad06fb4e2aab449490b3b4a08ab04caecb01c1b557c6a
3
+ size 14709
checkpoint-1250/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7a7816a7fa51468af523da3f8a4c3dfb8f596a7f008f74fb935c20fff6166629
3
+ size 1465
checkpoint-1250/tokenizer.json ADDED
The diff for this file is too large to render. See raw diff
 
checkpoint-1250/tokenizer_config.json ADDED
@@ -0,0 +1,16 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "backend": "tokenizers",
3
+ "clean_up_tokenization_spaces": true,
4
+ "cls_token": "[CLS]",
5
+ "is_local": false,
6
+ "mask_token": "[MASK]",
7
+ "model_input_names": [
8
+ "input_ids",
9
+ "attention_mask"
10
+ ],
11
+ "model_max_length": 8192,
12
+ "pad_token": "[PAD]",
13
+ "sep_token": "[SEP]",
14
+ "tokenizer_class": "TokenizersBackend",
15
+ "unk_token": "[UNK]"
16
+ }
checkpoint-1250/trainer_state.json ADDED
@@ -0,0 +1,1078 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_global_step": 900,
3
+ "best_metric": 0.8546255506607928,
4
+ "best_model_checkpoint": "/content/drive/MyDrive/expE_large_oversample_rv/checkpoint-900",
5
+ "epoch": 12.89119170984456,
6
+ "eval_steps": 50,
7
+ "global_step": 1250,
8
+ "is_hyper_param_search": false,
9
+ "is_local_process_zero": true,
10
+ "is_world_process_zero": true,
11
+ "log_history": [
12
+ {
13
+ "epoch": 0.010362694300518135,
14
+ "grad_norm": 19.391862869262695,
15
+ "learning_rate": 0.0,
16
+ "loss": 0.8294010162353516,
17
+ "step": 1
18
+ },
19
+ {
20
+ "epoch": 0.12435233160621761,
21
+ "grad_norm": 7.022524833679199,
22
+ "learning_rate": 5.670103092783505e-06,
23
+ "loss": 0.6434053074229847,
24
+ "step": 12
25
+ },
26
+ {
27
+ "epoch": 0.24870466321243523,
28
+ "grad_norm": 2.0728249549865723,
29
+ "learning_rate": 1.1855670103092783e-05,
30
+ "loss": 0.5182687441507975,
31
+ "step": 24
32
+ },
33
+ {
34
+ "epoch": 0.37305699481865284,
35
+ "grad_norm": 8.758841514587402,
36
+ "learning_rate": 1.8041237113402062e-05,
37
+ "loss": 0.39159242312113446,
38
+ "step": 36
39
+ },
40
+ {
41
+ "epoch": 0.49740932642487046,
42
+ "grad_norm": 10.565062522888184,
43
+ "learning_rate": 2.422680412371134e-05,
44
+ "loss": 0.31471407413482666,
45
+ "step": 48
46
+ },
47
+ {
48
+ "epoch": 0.5181347150259067,
49
+ "eval_accuracy": 0.7721899849559424,
50
+ "eval_f1": 0.4335260115606937,
51
+ "eval_loss": 0.05006009340286255,
52
+ "eval_precision": 0.32142857142857145,
53
+ "eval_recall": 0.665680473372781,
54
+ "eval_runtime": 5.6283,
55
+ "eval_samples_per_second": 60.586,
56
+ "eval_steps_per_second": 1.954,
57
+ "step": 50
58
+ },
59
+ {
60
+ "epoch": 0.6217616580310881,
61
+ "grad_norm": 5.744125843048096,
62
+ "learning_rate": 3.0412371134020617e-05,
63
+ "loss": 0.28232719500859577,
64
+ "step": 60
65
+ },
66
+ {
67
+ "epoch": 0.7461139896373057,
68
+ "grad_norm": 4.099153518676758,
69
+ "learning_rate": 3.6597938144329896e-05,
70
+ "loss": 0.22093822558720908,
71
+ "step": 72
72
+ },
73
+ {
74
+ "epoch": 0.8704663212435233,
75
+ "grad_norm": 12.179475784301758,
76
+ "learning_rate": 4.2783505154639175e-05,
77
+ "loss": 0.2065215309460958,
78
+ "step": 84
79
+ },
80
+ {
81
+ "epoch": 0.9948186528497409,
82
+ "grad_norm": 8.685556411743164,
83
+ "learning_rate": 4.8969072164948454e-05,
84
+ "loss": 0.16811863581339517,
85
+ "step": 96
86
+ },
87
+ {
88
+ "epoch": 1.0310880829015545,
89
+ "eval_accuracy": 0.7461852568235547,
90
+ "eval_f1": 0.4638047138047138,
91
+ "eval_loss": 0.03645005077123642,
92
+ "eval_precision": 0.3241176470588235,
93
+ "eval_recall": 0.8150887573964497,
94
+ "eval_runtime": 3.0466,
95
+ "eval_samples_per_second": 111.926,
96
+ "eval_steps_per_second": 3.611,
97
+ "step": 100
98
+ },
99
+ {
100
+ "epoch": 1.1139896373056994,
101
+ "grad_norm": 33.46553421020508,
102
+ "learning_rate": 4.999331054325924e-05,
103
+ "loss": 0.15307321151097616,
104
+ "step": 108
105
+ },
106
+ {
107
+ "epoch": 1.238341968911917,
108
+ "grad_norm": 5.198877811431885,
109
+ "learning_rate": 4.996762857385083e-05,
110
+ "loss": 0.1403733491897583,
111
+ "step": 120
112
+ },
113
+ {
114
+ "epoch": 1.3626943005181347,
115
+ "grad_norm": 1.8454481363296509,
116
+ "learning_rate": 4.9922706292838655e-05,
117
+ "loss": 0.13873055577278137,
118
+ "step": 132
119
+ },
120
+ {
121
+ "epoch": 1.4870466321243523,
122
+ "grad_norm": 2.0861477851867676,
123
+ "learning_rate": 4.9858578317794894e-05,
124
+ "loss": 0.11305248737335205,
125
+ "step": 144
126
+ },
127
+ {
128
+ "epoch": 1.549222797927461,
129
+ "eval_accuracy": 0.895766172361917,
130
+ "eval_f1": 0.6706515242080096,
131
+ "eval_loss": 0.029485274106264114,
132
+ "eval_precision": 0.5626880641925778,
133
+ "eval_recall": 0.8298816568047337,
134
+ "eval_runtime": 3.0006,
135
+ "eval_samples_per_second": 113.645,
136
+ "eval_steps_per_second": 3.666,
137
+ "step": 150
138
+ },
139
+ {
140
+ "epoch": 1.61139896373057,
141
+ "grad_norm": 6.362898349761963,
142
+ "learning_rate": 4.9775294066397274e-05,
143
+ "loss": 0.11805283029874165,
144
+ "step": 156
145
+ },
146
+ {
147
+ "epoch": 1.7357512953367875,
148
+ "grad_norm": 2.3624749183654785,
149
+ "learning_rate": 4.967291771834727e-05,
150
+ "loss": 0.11144719521204631,
151
+ "step": 168
152
+ },
153
+ {
154
+ "epoch": 1.8601036269430051,
155
+ "grad_norm": 0.9627630710601807,
156
+ "learning_rate": 4.955152816591259e-05,
157
+ "loss": 0.0819690078496933,
158
+ "step": 180
159
+ },
160
+ {
161
+ "epoch": 1.9844559585492227,
162
+ "grad_norm": 15.513121604919434,
163
+ "learning_rate": 4.9411218953131964e-05,
164
+ "loss": 0.10844972729682922,
165
+ "step": 192
166
+ },
167
+ {
168
+ "epoch": 2.062176165803109,
169
+ "eval_accuracy": 0.8074360627552117,
170
+ "eval_f1": 0.5410536307546274,
171
+ "eval_loss": 0.030047627165913582,
172
+ "eval_precision": 0.39832285115303984,
173
+ "eval_recall": 0.8431952662721893,
174
+ "eval_runtime": 3.037,
175
+ "eval_samples_per_second": 112.283,
176
+ "eval_steps_per_second": 3.622,
177
+ "step": 200
178
+ },
179
+ {
180
+ "epoch": 2.1036269430051813,
181
+ "grad_norm": 2.684053421020508,
182
+ "learning_rate": 4.9252098203729154e-05,
183
+ "loss": 0.08770081400871277,
184
+ "step": 204
185
+ },
186
+ {
187
+ "epoch": 2.227979274611399,
188
+ "grad_norm": 21.072328567504883,
189
+ "learning_rate": 4.907428853779167e-05,
190
+ "loss": 0.1440110206604004,
191
+ "step": 216
192
+ },
193
+ {
194
+ "epoch": 2.3523316062176165,
195
+ "grad_norm": 3.221358299255371,
196
+ "learning_rate": 4.887792697727854e-05,
197
+ "loss": 0.11832749843597412,
198
+ "step": 228
199
+ },
200
+ {
201
+ "epoch": 2.476683937823834,
202
+ "grad_norm": 5.4783101081848145,
203
+ "learning_rate": 4.866316484042968e-05,
204
+ "loss": 0.0703568160533905,
205
+ "step": 240
206
+ },
207
+ {
208
+ "epoch": 2.5803108808290154,
209
+ "eval_accuracy": 0.9129593810444874,
210
+ "eval_f1": 0.7118012422360248,
211
+ "eval_loss": 0.0311787910759449,
212
+ "eval_precision": 0.6134903640256959,
213
+ "eval_recall": 0.8476331360946746,
214
+ "eval_runtime": 3.0166,
215
+ "eval_samples_per_second": 113.041,
216
+ "eval_steps_per_second": 3.646,
217
+ "step": 250
218
+ },
219
+ {
220
+ "epoch": 2.6010362694300517,
221
+ "grad_norm": 6.384552478790283,
222
+ "learning_rate": 4.8430167625158595e-05,
223
+ "loss": 0.09287564953168233,
224
+ "step": 252
225
+ },
226
+ {
227
+ "epoch": 2.7253886010362693,
228
+ "grad_norm": 4.217259407043457,
229
+ "learning_rate": 4.817911488151791e-05,
230
+ "loss": 0.08391642570495605,
231
+ "step": 264
232
+ },
233
+ {
234
+ "epoch": 2.849740932642487,
235
+ "grad_norm": 3.1646366119384766,
236
+ "learning_rate": 4.791020007333629e-05,
237
+ "loss": 0.061600317557652794,
238
+ "step": 276
239
+ },
240
+ {
241
+ "epoch": 2.9740932642487046,
242
+ "grad_norm": 6.243884563446045,
243
+ "learning_rate": 4.762363042913322e-05,
244
+ "loss": 0.05470449725786845,
245
+ "step": 288
246
+ },
247
+ {
248
+ "epoch": 3.093264248704663,
249
+ "grad_norm": 2.5704944133758545,
250
+ "learning_rate": 4.7319626782426564e-05,
251
+ "loss": 0.05606028437614441,
252
+ "step": 300
253
+ },
254
+ {
255
+ "epoch": 3.093264248704663,
256
+ "eval_accuracy": 0.7977648828712659,
257
+ "eval_f1": 0.5429740791268759,
258
+ "eval_loss": 0.03227110207080841,
259
+ "eval_precision": 0.39198949441891007,
260
+ "eval_recall": 0.8831360946745562,
261
+ "eval_runtime": 3.0347,
262
+ "eval_samples_per_second": 112.367,
263
+ "eval_steps_per_second": 3.625,
264
+ "step": 300
265
+ },
266
+ {
267
+ "epoch": 3.2176165803108807,
268
+ "grad_norm": 2.420072317123413,
269
+ "learning_rate": 4.699842340155598e-05,
270
+ "loss": 0.053138395150502525,
271
+ "step": 312
272
+ },
273
+ {
274
+ "epoch": 3.3419689119170983,
275
+ "grad_norm": 67.39490509033203,
276
+ "learning_rate": 4.666026780915334e-05,
277
+ "loss": 0.10085086027781169,
278
+ "step": 324
279
+ },
280
+ {
281
+ "epoch": 3.466321243523316,
282
+ "grad_norm": 0.7816835045814514,
283
+ "learning_rate": 4.630542059139924e-05,
284
+ "loss": 0.0743176539738973,
285
+ "step": 336
286
+ },
287
+ {
288
+ "epoch": 3.5906735751295336,
289
+ "grad_norm": 0.7841729521751404,
290
+ "learning_rate": 4.593415519721257e-05,
291
+ "loss": 0.05843483408292135,
292
+ "step": 348
293
+ },
294
+ {
295
+ "epoch": 3.61139896373057,
296
+ "eval_accuracy": 0.9329464861379755,
297
+ "eval_f1": 0.7535070140280562,
298
+ "eval_loss": 0.030000876635313034,
299
+ "eval_precision": 0.6869671132764921,
300
+ "eval_recall": 0.834319526627219,
301
+ "eval_runtime": 3.0255,
302
+ "eval_samples_per_second": 112.709,
303
+ "eval_steps_per_second": 3.636,
304
+ "step": 350
305
+ },
306
+ {
307
+ "epoch": 3.715025906735751,
308
+ "grad_norm": 5.054598331451416,
309
+ "learning_rate": 4.5546757727528035e-05,
310
+ "loss": 0.03325176239013672,
311
+ "step": 360
312
+ },
313
+ {
314
+ "epoch": 3.839378238341969,
315
+ "grad_norm": 0.5077673196792603,
316
+ "learning_rate": 4.5143526714823766e-05,
317
+ "loss": 0.036297169824441276,
318
+ "step": 372
319
+ },
320
+ {
321
+ "epoch": 3.9637305699481864,
322
+ "grad_norm": 0.8186505436897278,
323
+ "learning_rate": 4.472477289306913e-05,
324
+ "loss": 0.026280519862969715,
325
+ "step": 384
326
+ },
327
+ {
328
+ "epoch": 4.082901554404145,
329
+ "grad_norm": 0.9546571373939514,
330
+ "learning_rate": 4.429081895826997e-05,
331
+ "loss": 0.038141210873921715,
332
+ "step": 396
333
+ },
334
+ {
335
+ "epoch": 4.124352331606218,
336
+ "eval_accuracy": 0.9116698903932946,
337
+ "eval_f1": 0.7129798903107862,
338
+ "eval_loss": 0.02928924560546875,
339
+ "eval_precision": 0.6062176165803109,
340
+ "eval_recall": 0.8653846153846154,
341
+ "eval_runtime": 3.0233,
342
+ "eval_samples_per_second": 112.79,
343
+ "eval_steps_per_second": 3.638,
344
+ "step": 400
345
+ },
346
+ {
347
+ "epoch": 4.2072538860103625,
348
+ "grad_norm": 0.38423895835876465,
349
+ "learning_rate": 4.38419993197957e-05,
350
+ "loss": 0.023756501575311024,
351
+ "step": 408
352
+ },
353
+ {
354
+ "epoch": 4.331606217616581,
355
+ "grad_norm": 0.261392205953598,
356
+ "learning_rate": 4.337865984268001e-05,
357
+ "loss": 0.02620025227467219,
358
+ "step": 420
359
+ },
360
+ {
361
+ "epoch": 4.455958549222798,
362
+ "grad_norm": 2.0392212867736816,
363
+ "learning_rate": 4.290115758109375e-05,
364
+ "loss": 0.02181059867143631,
365
+ "step": 432
366
+ },
367
+ {
368
+ "epoch": 4.580310880829016,
369
+ "grad_norm": 0.4685036242008209,
370
+ "learning_rate": 4.240986050319524e-05,
371
+ "loss": 0.023431827624638874,
372
+ "step": 444
373
+ },
374
+ {
375
+ "epoch": 4.642487046632124,
376
+ "eval_accuracy": 0.9593810444874274,
377
+ "eval_f1": 0.828297715549005,
378
+ "eval_loss": 0.04263457655906677,
379
+ "eval_precision": 0.8252569750367107,
380
+ "eval_recall": 0.8313609467455622,
381
+ "eval_runtime": 3.0402,
382
+ "eval_samples_per_second": 112.164,
383
+ "eval_steps_per_second": 3.618,
384
+ "step": 450
385
+ },
386
+ {
387
+ "epoch": 4.704663212435233,
388
+ "grad_norm": 0.535210371017456,
389
+ "learning_rate": 4.19051472075703e-05,
390
+ "loss": 0.021484971046447754,
391
+ "step": 456
392
+ },
393
+ {
394
+ "epoch": 4.829015544041451,
395
+ "grad_norm": 3.4881668090820312,
396
+ "learning_rate": 4.138740663148019e-05,
397
+ "loss": 0.03922188033660253,
398
+ "step": 468
399
+ },
400
+ {
401
+ "epoch": 4.953367875647668,
402
+ "grad_norm": 0.809377908706665,
403
+ "learning_rate": 4.085703775114266e-05,
404
+ "loss": 0.018220727642377216,
405
+ "step": 480
406
+ },
407
+ {
408
+ "epoch": 5.072538860103627,
409
+ "grad_norm": 3.266059160232544,
410
+ "learning_rate": 4.0314449274276663e-05,
411
+ "loss": 0.015343919396400452,
412
+ "step": 492
413
+ },
414
+ {
415
+ "epoch": 5.155440414507772,
416
+ "eval_accuracy": 0.9574468085106383,
417
+ "eval_f1": 0.8162672476397966,
418
+ "eval_loss": 0.04887477308511734,
419
+ "eval_precision": 0.8017118402282454,
420
+ "eval_recall": 0.8313609467455622,
421
+ "eval_runtime": 3.0737,
422
+ "eval_samples_per_second": 110.94,
423
+ "eval_steps_per_second": 3.579,
424
+ "step": 500
425
+ },
426
+ {
427
+ "epoch": 5.196891191709844,
428
+ "grad_norm": 0.14335286617279053,
429
+ "learning_rate": 3.976005932514807e-05,
430
+ "loss": 0.007969564447800318,
431
+ "step": 504
432
+ },
433
+ {
434
+ "epoch": 5.321243523316062,
435
+ "grad_norm": 0.24810202419757843,
436
+ "learning_rate": 3.919429512235874e-05,
437
+ "loss": 0.007458047941327095,
438
+ "step": 516
439
+ },
440
+ {
441
+ "epoch": 5.44559585492228,
442
+ "grad_norm": 1.7369506359100342,
443
+ "learning_rate": 3.86175926496275e-05,
444
+ "loss": 0.012967783957719803,
445
+ "step": 528
446
+ },
447
+ {
448
+ "epoch": 5.569948186528498,
449
+ "grad_norm": 0.40230634808540344,
450
+ "learning_rate": 3.8030396319816585e-05,
451
+ "loss": 0.008079041416446367,
452
+ "step": 540
453
+ },
454
+ {
455
+ "epoch": 5.6735751295336785,
456
+ "eval_accuracy": 0.9591661293788953,
457
+ "eval_f1": 0.8310168251645941,
458
+ "eval_loss": 0.04243873059749603,
459
+ "eval_precision": 0.8219971056439942,
460
+ "eval_recall": 0.8402366863905325,
461
+ "eval_runtime": 3.0267,
462
+ "eval_samples_per_second": 112.662,
463
+ "eval_steps_per_second": 3.634,
464
+ "step": 550
465
+ },
466
+ {
467
+ "epoch": 5.694300518134715,
468
+ "grad_norm": 0.46429362893104553,
469
+ "learning_rate": 3.743315863246256e-05,
470
+ "loss": 0.005399468044439952,
471
+ "step": 552
472
+ },
473
+ {
474
+ "epoch": 5.818652849740933,
475
+ "grad_norm": 1.60989248752594,
476
+ "learning_rate": 3.682633982507543e-05,
477
+ "loss": 0.006222298989693324,
478
+ "step": 564
479
+ },
480
+ {
481
+ "epoch": 5.94300518134715,
482
+ "grad_norm": 0.5653553009033203,
483
+ "learning_rate": 3.621040751847497e-05,
484
+ "loss": 0.015771896888812382,
485
+ "step": 576
486
+ },
487
+ {
488
+ "epoch": 6.062176165803109,
489
+ "grad_norm": 0.34870776534080505,
490
+ "learning_rate": 3.5585836356437264e-05,
491
+ "loss": 0.0066079869866371155,
492
+ "step": 588
493
+ },
494
+ {
495
+ "epoch": 6.186528497409326,
496
+ "grad_norm": 0.2842996120452881,
497
+ "learning_rate": 3.495310763992928e-05,
498
+ "loss": 0.0051950619866450625,
499
+ "step": 600
500
+ },
501
+ {
502
+ "epoch": 6.186528497409326,
503
+ "eval_accuracy": 0.9580915538362347,
504
+ "eval_f1": 0.8308351177730193,
505
+ "eval_loss": 0.04230675473809242,
506
+ "eval_precision": 0.8027586206896552,
507
+ "eval_recall": 0.8609467455621301,
508
+ "eval_runtime": 3.0559,
509
+ "eval_samples_per_second": 111.586,
510
+ "eval_steps_per_second": 3.6,
511
+ "step": 600
512
+ },
513
+ {
514
+ "epoch": 6.310880829015544,
515
+ "grad_norm": 0.20262780785560608,
516
+ "learning_rate": 3.431270895621349e-05,
517
+ "loss": 0.0038215465222795806,
518
+ "step": 612
519
+ },
520
+ {
521
+ "epoch": 6.435233160621761,
522
+ "grad_norm": 0.08969880640506744,
523
+ "learning_rate": 3.3665133803108035e-05,
524
+ "loss": 0.002952239786585172,
525
+ "step": 624
526
+ },
527
+ {
528
+ "epoch": 6.5595854922279795,
529
+ "grad_norm": 7.001764297485352,
530
+ "learning_rate": 3.301088120869224e-05,
531
+ "loss": 0.006995424007376035,
532
+ "step": 636
533
+ },
534
+ {
535
+ "epoch": 6.683937823834197,
536
+ "grad_norm": 1.759656548500061,
537
+ "learning_rate": 3.235045534675045e-05,
538
+ "loss": 0.0032693296670913696,
539
+ "step": 648
540
+ },
541
+ {
542
+ "epoch": 6.704663212435233,
543
+ "eval_accuracy": 0.9537932516655921,
544
+ "eval_f1": 0.8183098591549295,
545
+ "eval_loss": 0.04099448770284653,
546
+ "eval_precision": 0.7809139784946236,
547
+ "eval_recall": 0.8594674556213018,
548
+ "eval_runtime": 3.0102,
549
+ "eval_samples_per_second": 113.281,
550
+ "eval_steps_per_second": 3.654,
551
+ "step": 650
552
+ },
553
+ {
554
+ "epoch": 6.808290155440415,
555
+ "grad_norm": 0.113290935754776,
556
+ "learning_rate": 3.168436514825052e-05,
557
+ "loss": 0.003636385935048262,
558
+ "step": 660
559
+ },
560
+ {
561
+ "epoch": 6.932642487046632,
562
+ "grad_norm": 0.4593416452407837,
563
+ "learning_rate": 3.101312390915634e-05,
564
+ "loss": 0.006359308337171872,
565
+ "step": 672
566
+ },
567
+ {
568
+ "epoch": 7.051813471502591,
569
+ "grad_norm": 0.5328609943389893,
570
+ "learning_rate": 3.0337248894876746e-05,
571
+ "loss": 0.0031539862975478172,
572
+ "step": 684
573
+ },
574
+ {
575
+ "epoch": 7.176165803108808,
576
+ "grad_norm": 0.13898538053035736,
577
+ "learning_rate": 2.9657260941655417e-05,
578
+ "loss": 0.0014738718358178933,
579
+ "step": 696
580
+ },
581
+ {
582
+ "epoch": 7.217616580310881,
583
+ "eval_accuracy": 0.9636793466580701,
584
+ "eval_f1": 0.8498542274052479,
585
+ "eval_loss": 0.041410233825445175,
586
+ "eval_precision": 0.8376436781609196,
587
+ "eval_recall": 0.8624260355029586,
588
+ "eval_runtime": 3.0173,
589
+ "eval_samples_per_second": 113.015,
590
+ "eval_steps_per_second": 3.646,
591
+ "step": 700
592
+ },
593
+ {
594
+ "epoch": 7.300518134715026,
595
+ "grad_norm": 0.020028043538331985,
596
+ "learning_rate": 2.8973684055209106e-05,
597
+ "loss": 0.0009494482850035032,
598
+ "step": 708
599
+ },
600
+ {
601
+ "epoch": 7.424870466321243,
602
+ "grad_norm": 0.038508884608745575,
603
+ "learning_rate": 2.8287045006923473e-05,
604
+ "loss": 0.002579918752113978,
605
+ "step": 720
606
+ },
607
+ {
608
+ "epoch": 7.549222797927461,
609
+ "grad_norm": 0.46958020329475403,
610
+ "learning_rate": 2.7597872927917613e-05,
611
+ "loss": 0.0010905038410176833,
612
+ "step": 732
613
+ },
614
+ {
615
+ "epoch": 7.6735751295336785,
616
+ "grad_norm": 0.1547262817621231,
617
+ "learning_rate": 2.6906698901290167e-05,
618
+ "loss": 0.0028232391923666,
619
+ "step": 744
620
+ },
621
+ {
622
+ "epoch": 7.7357512953367875,
623
+ "eval_accuracy": 0.957016978293574,
624
+ "eval_f1": 0.829302987197724,
625
+ "eval_loss": 0.04088299721479416,
626
+ "eval_precision": 0.7986301369863014,
627
+ "eval_recall": 0.8624260355029586,
628
+ "eval_runtime": 3.0234,
629
+ "eval_samples_per_second": 112.787,
630
+ "eval_steps_per_second": 3.638,
631
+ "step": 750
632
+ },
633
+ {
634
+ "epoch": 7.7979274611398965,
635
+ "grad_norm": 0.0519477054476738,
636
+ "learning_rate": 2.621405555286121e-05,
637
+ "loss": 0.0010418800326685111,
638
+ "step": 756
639
+ },
640
+ {
641
+ "epoch": 7.922279792746114,
642
+ "grad_norm": 1.0287493467330933,
643
+ "learning_rate": 2.5520476640725342e-05,
644
+ "loss": 0.001736572477966547,
645
+ "step": 768
646
+ },
647
+ {
648
+ "epoch": 8.041450777202073,
649
+ "grad_norm": 0.013289994560182095,
650
+ "learning_rate": 2.4826496643932183e-05,
651
+ "loss": 0.000329588889144361,
652
+ "step": 780
653
+ },
654
+ {
655
+ "epoch": 8.16580310880829,
656
+ "grad_norm": 0.011798656545579433,
657
+ "learning_rate": 2.413265035061131e-05,
658
+ "loss": 0.002016165293753147,
659
+ "step": 792
660
+ },
661
+ {
662
+ "epoch": 8.248704663212436,
663
+ "eval_accuracy": 0.9630346013324736,
664
+ "eval_f1": 0.8407079646017699,
665
+ "eval_loss": 0.05126392841339111,
666
+ "eval_precision": 0.8382352941176471,
667
+ "eval_recall": 0.8431952662721893,
668
+ "eval_runtime": 3.0399,
669
+ "eval_samples_per_second": 112.176,
670
+ "eval_steps_per_second": 3.619,
671
+ "step": 800
672
+ },
673
+ {
674
+ "epoch": 8.290155440414507,
675
+ "grad_norm": 0.016478369012475014,
676
+ "learning_rate": 2.343947244585902e-05,
677
+ "loss": 0.001583070649454991,
678
+ "step": 804
679
+ },
680
+ {
681
+ "epoch": 8.414507772020725,
682
+ "grad_norm": 0.08483371138572693,
683
+ "learning_rate": 2.2747497099704533e-05,
684
+ "loss": 0.0006651872924218575,
685
+ "step": 816
686
+ },
687
+ {
688
+ "epoch": 8.538860103626943,
689
+ "grad_norm": 0.02956315688788891,
690
+ "learning_rate": 2.2057257555472998e-05,
691
+ "loss": 0.001361117387811343,
692
+ "step": 828
693
+ },
694
+ {
695
+ "epoch": 8.663212435233161,
696
+ "grad_norm": 0.1703813225030899,
697
+ "learning_rate": 2.136928571886275e-05,
698
+ "loss": 0.0004198816216861208,
699
+ "step": 840
700
+ },
701
+ {
702
+ "epoch": 8.766839378238341,
703
+ "eval_accuracy": 0.9636793466580701,
704
+ "eval_f1": 0.851528384279476,
705
+ "eval_loss": 0.04630826786160469,
706
+ "eval_precision": 0.83810888252149,
707
+ "eval_recall": 0.8653846153846154,
708
+ "eval_runtime": 3.0066,
709
+ "eval_samples_per_second": 113.418,
710
+ "eval_steps_per_second": 3.659,
711
+ "step": 850
712
+ },
713
+ {
714
+ "epoch": 8.787564766839377,
715
+ "grad_norm": 0.08187761902809143,
716
+ "learning_rate": 2.068411174805327e-05,
717
+ "loss": 0.00021340335176015893,
718
+ "step": 852
719
+ },
720
+ {
721
+ "epoch": 8.911917098445596,
722
+ "grad_norm": 0.1566244512796402,
723
+ "learning_rate": 2.0002263645159892e-05,
724
+ "loss": 0.0004732525364185373,
725
+ "step": 864
726
+ },
727
+ {
728
+ "epoch": 9.031088082901555,
729
+ "grad_norm": 0.007061833515763283,
730
+ "learning_rate": 1.9324266849349848e-05,
731
+ "loss": 9.06402322774132e-05,
732
+ "step": 876
733
+ },
734
+ {
735
+ "epoch": 9.155440414507773,
736
+ "grad_norm": 0.010846185497939587,
737
+ "learning_rate": 1.8650643831933502e-05,
738
+ "loss": 0.00011390061505759756,
739
+ "step": 888
740
+ },
741
+ {
742
+ "epoch": 9.279792746113989,
743
+ "grad_norm": 0.03851994872093201,
744
+ "learning_rate": 1.798191369374258e-05,
745
+ "loss": 0.00019203778356313705,
746
+ "step": 900
747
+ },
748
+ {
749
+ "epoch": 9.279792746113989,
750
+ "eval_accuracy": 0.9649688373092629,
751
+ "eval_f1": 0.8546255506607928,
752
+ "eval_loss": 0.05020074546337128,
753
+ "eval_precision": 0.8483965014577259,
754
+ "eval_recall": 0.8609467455621301,
755
+ "eval_runtime": 3.0339,
756
+ "eval_samples_per_second": 112.397,
757
+ "eval_steps_per_second": 3.626,
758
+ "step": 900
759
+ },
760
+ {
761
+ "epoch": 9.404145077720207,
762
+ "grad_norm": 0.006672121584415436,
763
+ "learning_rate": 1.731859176510569e-05,
764
+ "loss": 9.344543407981594e-05,
765
+ "step": 912
766
+ },
767
+ {
768
+ "epoch": 9.528497409326425,
769
+ "grad_norm": 0.028979798778891563,
770
+ "learning_rate": 1.666118920872949e-05,
771
+ "loss": 0.0009806355616698663,
772
+ "step": 924
773
+ },
774
+ {
775
+ "epoch": 9.652849740932643,
776
+ "grad_norm": 0.010359357111155987,
777
+ "learning_rate": 1.6010212625791422e-05,
778
+ "loss": 0.00018761774602656564,
779
+ "step": 936
780
+ },
781
+ {
782
+ "epoch": 9.77720207253886,
783
+ "grad_norm": 0.0016930641140788794,
784
+ "learning_rate": 1.536616366554768e-05,
785
+ "loss": 0.00010497189941816032,
786
+ "step": 948
787
+ },
788
+ {
789
+ "epoch": 9.797927461139896,
790
+ "eval_accuracy": 0.9641091768751343,
791
+ "eval_f1": 0.849112426035503,
792
+ "eval_loss": 0.05002701282501221,
793
+ "eval_precision": 0.849112426035503,
794
+ "eval_recall": 0.849112426035503,
795
+ "eval_runtime": 3.0151,
796
+ "eval_samples_per_second": 113.098,
797
+ "eval_steps_per_second": 3.648,
798
+ "step": 950
799
+ },
800
+ {
801
+ "epoch": 9.901554404145077,
802
+ "grad_norm": 0.0343826562166214,
803
+ "learning_rate": 1.472953863875706e-05,
804
+ "loss": 8.524257767324646e-05,
805
+ "step": 960
806
+ },
807
+ {
808
+ "epoch": 10.020725388601036,
809
+ "grad_norm": 0.006111979950219393,
810
+ "learning_rate": 1.4100828135218829e-05,
811
+ "loss": 0.00012398386024869978,
812
+ "step": 972
813
+ },
814
+ {
815
+ "epoch": 10.145077720207254,
816
+ "grad_norm": 0.010606050491333008,
817
+ "learning_rate": 1.3480516645719172e-05,
818
+ "loss": 3.8064138304131724e-05,
819
+ "step": 984
820
+ },
821
+ {
822
+ "epoch": 10.26943005181347,
823
+ "grad_norm": 0.0009117216104641557,
824
+ "learning_rate": 1.2869082188677572e-05,
825
+ "loss": 0.0001223099728425344,
826
+ "step": 996
827
+ },
828
+ {
829
+ "epoch": 10.310880829015543,
830
+ "eval_accuracy": 0.9645390070921985,
831
+ "eval_f1": 0.8514412416851442,
832
+ "eval_loss": 0.05051806941628456,
833
+ "eval_precision": 0.8508124076809453,
834
+ "eval_recall": 0.8520710059171598,
835
+ "eval_runtime": 3.0202,
836
+ "eval_samples_per_second": 112.906,
837
+ "eval_steps_per_second": 3.642,
838
+ "step": 1000
839
+ },
840
+ {
841
+ "epoch": 10.393782383419689,
842
+ "grad_norm": 0.0027385777793824673,
843
+ "learning_rate": 1.2266995941780934e-05,
844
+ "loss": 7.653955738836278e-05,
845
+ "step": 1008
846
+ },
847
+ {
848
+ "epoch": 10.518134715025907,
849
+ "grad_norm": 0.00011921550321858376,
850
+ "learning_rate": 1.1674721878889171e-05,
851
+ "loss": 0.0001742592042622467,
852
+ "step": 1020
853
+ },
854
+ {
855
+ "epoch": 10.642487046632125,
856
+ "grad_norm": 0.001234560739248991,
857
+ "learning_rate": 1.1092716412492226e-05,
858
+ "loss": 7.672411447856575e-05,
859
+ "step": 1032
860
+ },
861
+ {
862
+ "epoch": 10.766839378238341,
863
+ "grad_norm": 0.0005847270367667079,
864
+ "learning_rate": 1.0521428041993848e-05,
865
+ "loss": 9.932015624751026e-06,
866
+ "step": 1044
867
+ },
868
+ {
869
+ "epoch": 10.82901554404145,
870
+ "eval_accuracy": 0.9643240919836664,
871
+ "eval_f1": 0.8508124076809453,
872
+ "eval_loss": 0.05043612793087959,
873
+ "eval_precision": 0.8495575221238938,
874
+ "eval_recall": 0.8520710059171598,
875
+ "eval_runtime": 3.024,
876
+ "eval_samples_per_second": 112.764,
877
+ "eval_steps_per_second": 3.638,
878
+ "step": 1050
879
+ },
880
+ {
881
+ "epoch": 10.89119170984456,
882
+ "grad_norm": 0.0753927081823349,
883
+ "learning_rate": 9.961297008093357e-06,
884
+ "loss": 0.00011471325221161048,
885
+ "step": 1056
886
+ },
887
+ {
888
+ "epoch": 11.010362694300518,
889
+ "grad_norm": 0.003958659712225199,
890
+ "learning_rate": 9.412754953531663e-06,
891
+ "loss": 1.709427791259562e-05,
892
+ "step": 1068
893
+ },
894
+ {
895
+ "epoch": 11.134715025906736,
896
+ "grad_norm": 0.001931481878273189,
897
+ "learning_rate": 8.876224590462873e-06,
898
+ "loss": 1.0689032933441922e-05,
899
+ "step": 1080
900
+ },
901
+ {
902
+ "epoch": 11.259067357512953,
903
+ "grad_norm": 0.012324509210884571,
904
+ "learning_rate": 8.352119374707978e-06,
905
+ "loss": 5.9023882689264916e-05,
906
+ "step": 1092
907
+ },
908
+ {
909
+ "epoch": 11.341968911917098,
910
+ "eval_accuracy": 0.9643240919836664,
911
+ "eval_f1": 0.8510324483775811,
912
+ "eval_loss": 0.050491511821746826,
913
+ "eval_precision": 0.8485294117647059,
914
+ "eval_recall": 0.8535502958579881,
915
+ "eval_runtime": 3.0303,
916
+ "eval_samples_per_second": 112.529,
917
+ "eval_steps_per_second": 3.63,
918
+ "step": 1100
919
+ },
920
+ {
921
+ "epoch": 11.38341968911917,
922
+ "grad_norm": 0.002621831139549613,
923
+ "learning_rate": 7.840843187141544e-06,
924
+ "loss": 5.975224000091354e-05,
925
+ "step": 1104
926
+ },
927
+ {
928
+ "epoch": 11.507772020725389,
929
+ "grad_norm": 0.0006910113152116537,
930
+ "learning_rate": 7.3427900224569606e-06,
931
+ "loss": 7.24908944296961e-05,
932
+ "step": 1116
933
+ },
934
+ {
935
+ "epoch": 11.632124352331607,
936
+ "grad_norm": 0.01195142138749361,
937
+ "learning_rate": 6.858343685550048e-06,
938
+ "loss": 3.2996586620962866e-05,
939
+ "step": 1128
940
+ },
941
+ {
942
+ "epoch": 11.756476683937823,
943
+ "grad_norm": 0.00024170154938474298,
944
+ "learning_rate": 6.387877495755065e-06,
945
+ "loss": 1.4914168180742612e-05,
946
+ "step": 1140
947
+ },
948
+ {
949
+ "epoch": 11.860103626943005,
950
+ "eval_accuracy": 0.9645390070921985,
951
+ "eval_f1": 0.8514412416851442,
952
+ "eval_loss": 0.05080362409353256,
953
+ "eval_precision": 0.8508124076809453,
954
+ "eval_recall": 0.8520710059171598,
955
+ "eval_runtime": 3.0169,
956
+ "eval_samples_per_second": 113.03,
957
+ "eval_steps_per_second": 3.646,
958
+ "step": 1150
959
+ },
960
+ {
961
+ "epoch": 11.880829015544041,
962
+ "grad_norm": 0.00013972603483125567,
963
+ "learning_rate": 5.93175399916101e-06,
964
+ "loss": 0.00017348669158915678,
965
+ "step": 1152
966
+ },
967
+ {
968
+ "epoch": 12.0,
969
+ "grad_norm": 0.014590195380151272,
970
+ "learning_rate": 5.490324689229834e-06,
971
+ "loss": 6.250141692968707e-05,
972
+ "step": 1164
973
+ },
974
+ {
975
+ "epoch": 12.124352331606218,
976
+ "grad_norm": 0.002015709411352873,
977
+ "learning_rate": 5.063929735931985e-06,
978
+ "loss": 1.4617435226682574e-05,
979
+ "step": 1176
980
+ },
981
+ {
982
+ "epoch": 12.248704663212436,
983
+ "grad_norm": 0.0005398832727223635,
984
+ "learning_rate": 4.652897723607913e-06,
985
+ "loss": 1.6918191249715164e-05,
986
+ "step": 1188
987
+ },
988
+ {
989
+ "epoch": 12.373056994818652,
990
+ "grad_norm": 0.0006741413380950689,
991
+ "learning_rate": 4.257545397757623e-06,
992
+ "loss": 4.216020170133561e-05,
993
+ "step": 1200
994
+ },
995
+ {
996
+ "epoch": 12.373056994818652,
997
+ "eval_accuracy": 0.9645390070921985,
998
+ "eval_f1": 0.8514412416851442,
999
+ "eval_loss": 0.05097455903887749,
1000
+ "eval_precision": 0.8508124076809453,
1001
+ "eval_recall": 0.8520710059171598,
1002
+ "eval_runtime": 3.0165,
1003
+ "eval_samples_per_second": 113.043,
1004
+ "eval_steps_per_second": 3.647,
1005
+ "step": 1200
1006
+ },
1007
+ {
1008
+ "epoch": 12.49740932642487,
1009
+ "grad_norm": 0.0007238159305416048,
1010
+ "learning_rate": 3.878177420953291e-06,
1011
+ "loss": 5.365341106274476e-05,
1012
+ "step": 1212
1013
+ },
1014
+ {
1015
+ "epoch": 12.621761658031089,
1016
+ "grad_norm": 0.002653967123478651,
1017
+ "learning_rate": 3.5150861380631755e-06,
1018
+ "loss": 3.87875431139643e-05,
1019
+ "step": 1224
1020
+ },
1021
+ {
1022
+ "epoch": 12.746113989637305,
1023
+ "grad_norm": 0.0012897817650809884,
1024
+ "learning_rate": 3.1685513509676606e-06,
1025
+ "loss": 1.6210615285672247e-05,
1026
+ "step": 1236
1027
+ },
1028
+ {
1029
+ "epoch": 12.870466321243523,
1030
+ "grad_norm": 0.010877900756895542,
1031
+ "learning_rate": 2.838840102941015e-06,
1032
+ "loss": 4.428145863736669e-05,
1033
+ "step": 1248
1034
+ },
1035
+ {
1036
+ "epoch": 12.89119170984456,
1037
+ "eval_accuracy": 0.9645390070921985,
1038
+ "eval_f1": 0.8514412416851442,
1039
+ "eval_loss": 0.05106152221560478,
1040
+ "eval_precision": 0.8508124076809453,
1041
+ "eval_recall": 0.8520710059171598,
1042
+ "eval_runtime": 3.0194,
1043
+ "eval_samples_per_second": 112.937,
1044
+ "eval_steps_per_second": 3.643,
1045
+ "step": 1250
1046
+ }
1047
+ ],
1048
+ "logging_steps": 12,
1049
+ "max_steps": 1455,
1050
+ "num_input_tokens_seen": 0,
1051
+ "num_train_epochs": 15,
1052
+ "save_steps": 50,
1053
+ "stateful_callbacks": {
1054
+ "EarlyStoppingCallback": {
1055
+ "args": {
1056
+ "early_stopping_patience": 10,
1057
+ "early_stopping_threshold": 0.0
1058
+ },
1059
+ "attributes": {
1060
+ "early_stopping_patience_counter": 7
1061
+ }
1062
+ },
1063
+ "TrainerControl": {
1064
+ "args": {
1065
+ "should_epoch_stop": false,
1066
+ "should_evaluate": false,
1067
+ "should_log": false,
1068
+ "should_save": true,
1069
+ "should_training_stop": false
1070
+ },
1071
+ "attributes": {}
1072
+ }
1073
+ },
1074
+ "total_flos": 2.103074668601549e+16,
1075
+ "train_batch_size": 8,
1076
+ "trial_name": null,
1077
+ "trial_params": null
1078
+ }
checkpoint-1250/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:32e421217e1eea826c4025f111c9f4fa0b9d53a8bca71632cae9ad6aed6b26f4
3
+ size 5201
checkpoint-1300/config.json ADDED
@@ -0,0 +1,99 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "architectures": [
3
+ "ModernBertForTokenClassification"
4
+ ],
5
+ "attention_bias": false,
6
+ "attention_dropout": 0.0,
7
+ "bos_token_id": 50281,
8
+ "classifier_activation": "gelu",
9
+ "classifier_bias": false,
10
+ "classifier_dropout": 0.3,
11
+ "classifier_pooling": "mean",
12
+ "cls_token_id": 50281,
13
+ "decoder_bias": true,
14
+ "deterministic_flash_attn": false,
15
+ "dtype": "float32",
16
+ "embedding_dropout": 0.0,
17
+ "eos_token_id": 50282,
18
+ "global_attn_every_n_layers": 3,
19
+ "gradient_checkpointing": false,
20
+ "hidden_activation": "gelu",
21
+ "hidden_size": 1024,
22
+ "id2label": {
23
+ "0": "O",
24
+ "1": "FP",
25
+ "2": "RP",
26
+ "3": "RV",
27
+ "4": "PW"
28
+ },
29
+ "initializer_cutoff_factor": 2.0,
30
+ "initializer_range": 0.02,
31
+ "intermediate_size": 2624,
32
+ "label2id": {
33
+ "FP": 1,
34
+ "O": 0,
35
+ "PW": 4,
36
+ "RP": 2,
37
+ "RV": 3
38
+ },
39
+ "layer_norm_eps": 1e-05,
40
+ "layer_types": [
41
+ "full_attention",
42
+ "sliding_attention",
43
+ "sliding_attention",
44
+ "full_attention",
45
+ "sliding_attention",
46
+ "sliding_attention",
47
+ "full_attention",
48
+ "sliding_attention",
49
+ "sliding_attention",
50
+ "full_attention",
51
+ "sliding_attention",
52
+ "sliding_attention",
53
+ "full_attention",
54
+ "sliding_attention",
55
+ "sliding_attention",
56
+ "full_attention",
57
+ "sliding_attention",
58
+ "sliding_attention",
59
+ "full_attention",
60
+ "sliding_attention",
61
+ "sliding_attention",
62
+ "full_attention",
63
+ "sliding_attention",
64
+ "sliding_attention",
65
+ "full_attention",
66
+ "sliding_attention",
67
+ "sliding_attention",
68
+ "full_attention"
69
+ ],
70
+ "local_attention": 128,
71
+ "max_position_embeddings": 8192,
72
+ "mlp_bias": false,
73
+ "mlp_dropout": 0.0,
74
+ "model_type": "modernbert",
75
+ "norm_bias": false,
76
+ "norm_eps": 1e-05,
77
+ "num_attention_heads": 16,
78
+ "num_hidden_layers": 28,
79
+ "pad_token_id": 50283,
80
+ "position_embedding_type": "absolute",
81
+ "repad_logits_with_grad": false,
82
+ "rope_parameters": {
83
+ "full_attention": {
84
+ "rope_theta": 160000.0,
85
+ "rope_type": "default"
86
+ },
87
+ "sliding_attention": {
88
+ "rope_theta": 10000.0,
89
+ "rope_type": "default"
90
+ }
91
+ },
92
+ "sep_token_id": 50282,
93
+ "sparse_pred_ignore_index": -100,
94
+ "sparse_prediction": false,
95
+ "tie_word_embeddings": true,
96
+ "transformers_version": "5.0.0",
97
+ "use_cache": false,
98
+ "vocab_size": 50368
99
+ }
checkpoint-1300/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:6abd07e14472da63133a66a4e46d4ad54588546b3e07824e6c431a49d09f1832
3
+ size 1583363940
checkpoint-1300/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d8216115ef02e9e575f508c994af124ca8d4e6ca5f23d751fef603273270123c
3
+ size 1112419851
checkpoint-1300/rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:af4cd8335cc00a9c9a53cbb03ec5aa4e79b0165fc60849e892bc46a20865fea6
3
+ size 14709
checkpoint-1300/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:77cca423b90dadc1335febc60d6a4d0c1f35705f91d027853d3b2308cbbc1d37
3
+ size 1465
checkpoint-1300/tokenizer.json ADDED
The diff for this file is too large to render. See raw diff
 
checkpoint-1300/tokenizer_config.json ADDED
@@ -0,0 +1,16 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "backend": "tokenizers",
3
+ "clean_up_tokenization_spaces": true,
4
+ "cls_token": "[CLS]",
5
+ "is_local": false,
6
+ "mask_token": "[MASK]",
7
+ "model_input_names": [
8
+ "input_ids",
9
+ "attention_mask"
10
+ ],
11
+ "model_max_length": 8192,
12
+ "pad_token": "[PAD]",
13
+ "sep_token": "[SEP]",
14
+ "tokenizer_class": "TokenizersBackend",
15
+ "unk_token": "[UNK]"
16
+ }
checkpoint-1300/trainer_state.json ADDED
@@ -0,0 +1,1118 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_global_step": 900,
3
+ "best_metric": 0.8546255506607928,
4
+ "best_model_checkpoint": "/content/drive/MyDrive/expE_large_oversample_rv/checkpoint-900",
5
+ "epoch": 13.404145077720207,
6
+ "eval_steps": 50,
7
+ "global_step": 1300,
8
+ "is_hyper_param_search": false,
9
+ "is_local_process_zero": true,
10
+ "is_world_process_zero": true,
11
+ "log_history": [
12
+ {
13
+ "epoch": 0.010362694300518135,
14
+ "grad_norm": 19.391862869262695,
15
+ "learning_rate": 0.0,
16
+ "loss": 0.8294010162353516,
17
+ "step": 1
18
+ },
19
+ {
20
+ "epoch": 0.12435233160621761,
21
+ "grad_norm": 7.022524833679199,
22
+ "learning_rate": 5.670103092783505e-06,
23
+ "loss": 0.6434053074229847,
24
+ "step": 12
25
+ },
26
+ {
27
+ "epoch": 0.24870466321243523,
28
+ "grad_norm": 2.0728249549865723,
29
+ "learning_rate": 1.1855670103092783e-05,
30
+ "loss": 0.5182687441507975,
31
+ "step": 24
32
+ },
33
+ {
34
+ "epoch": 0.37305699481865284,
35
+ "grad_norm": 8.758841514587402,
36
+ "learning_rate": 1.8041237113402062e-05,
37
+ "loss": 0.39159242312113446,
38
+ "step": 36
39
+ },
40
+ {
41
+ "epoch": 0.49740932642487046,
42
+ "grad_norm": 10.565062522888184,
43
+ "learning_rate": 2.422680412371134e-05,
44
+ "loss": 0.31471407413482666,
45
+ "step": 48
46
+ },
47
+ {
48
+ "epoch": 0.5181347150259067,
49
+ "eval_accuracy": 0.7721899849559424,
50
+ "eval_f1": 0.4335260115606937,
51
+ "eval_loss": 0.05006009340286255,
52
+ "eval_precision": 0.32142857142857145,
53
+ "eval_recall": 0.665680473372781,
54
+ "eval_runtime": 5.6283,
55
+ "eval_samples_per_second": 60.586,
56
+ "eval_steps_per_second": 1.954,
57
+ "step": 50
58
+ },
59
+ {
60
+ "epoch": 0.6217616580310881,
61
+ "grad_norm": 5.744125843048096,
62
+ "learning_rate": 3.0412371134020617e-05,
63
+ "loss": 0.28232719500859577,
64
+ "step": 60
65
+ },
66
+ {
67
+ "epoch": 0.7461139896373057,
68
+ "grad_norm": 4.099153518676758,
69
+ "learning_rate": 3.6597938144329896e-05,
70
+ "loss": 0.22093822558720908,
71
+ "step": 72
72
+ },
73
+ {
74
+ "epoch": 0.8704663212435233,
75
+ "grad_norm": 12.179475784301758,
76
+ "learning_rate": 4.2783505154639175e-05,
77
+ "loss": 0.2065215309460958,
78
+ "step": 84
79
+ },
80
+ {
81
+ "epoch": 0.9948186528497409,
82
+ "grad_norm": 8.685556411743164,
83
+ "learning_rate": 4.8969072164948454e-05,
84
+ "loss": 0.16811863581339517,
85
+ "step": 96
86
+ },
87
+ {
88
+ "epoch": 1.0310880829015545,
89
+ "eval_accuracy": 0.7461852568235547,
90
+ "eval_f1": 0.4638047138047138,
91
+ "eval_loss": 0.03645005077123642,
92
+ "eval_precision": 0.3241176470588235,
93
+ "eval_recall": 0.8150887573964497,
94
+ "eval_runtime": 3.0466,
95
+ "eval_samples_per_second": 111.926,
96
+ "eval_steps_per_second": 3.611,
97
+ "step": 100
98
+ },
99
+ {
100
+ "epoch": 1.1139896373056994,
101
+ "grad_norm": 33.46553421020508,
102
+ "learning_rate": 4.999331054325924e-05,
103
+ "loss": 0.15307321151097616,
104
+ "step": 108
105
+ },
106
+ {
107
+ "epoch": 1.238341968911917,
108
+ "grad_norm": 5.198877811431885,
109
+ "learning_rate": 4.996762857385083e-05,
110
+ "loss": 0.1403733491897583,
111
+ "step": 120
112
+ },
113
+ {
114
+ "epoch": 1.3626943005181347,
115
+ "grad_norm": 1.8454481363296509,
116
+ "learning_rate": 4.9922706292838655e-05,
117
+ "loss": 0.13873055577278137,
118
+ "step": 132
119
+ },
120
+ {
121
+ "epoch": 1.4870466321243523,
122
+ "grad_norm": 2.0861477851867676,
123
+ "learning_rate": 4.9858578317794894e-05,
124
+ "loss": 0.11305248737335205,
125
+ "step": 144
126
+ },
127
+ {
128
+ "epoch": 1.549222797927461,
129
+ "eval_accuracy": 0.895766172361917,
130
+ "eval_f1": 0.6706515242080096,
131
+ "eval_loss": 0.029485274106264114,
132
+ "eval_precision": 0.5626880641925778,
133
+ "eval_recall": 0.8298816568047337,
134
+ "eval_runtime": 3.0006,
135
+ "eval_samples_per_second": 113.645,
136
+ "eval_steps_per_second": 3.666,
137
+ "step": 150
138
+ },
139
+ {
140
+ "epoch": 1.61139896373057,
141
+ "grad_norm": 6.362898349761963,
142
+ "learning_rate": 4.9775294066397274e-05,
143
+ "loss": 0.11805283029874165,
144
+ "step": 156
145
+ },
146
+ {
147
+ "epoch": 1.7357512953367875,
148
+ "grad_norm": 2.3624749183654785,
149
+ "learning_rate": 4.967291771834727e-05,
150
+ "loss": 0.11144719521204631,
151
+ "step": 168
152
+ },
153
+ {
154
+ "epoch": 1.8601036269430051,
155
+ "grad_norm": 0.9627630710601807,
156
+ "learning_rate": 4.955152816591259e-05,
157
+ "loss": 0.0819690078496933,
158
+ "step": 180
159
+ },
160
+ {
161
+ "epoch": 1.9844559585492227,
162
+ "grad_norm": 15.513121604919434,
163
+ "learning_rate": 4.9411218953131964e-05,
164
+ "loss": 0.10844972729682922,
165
+ "step": 192
166
+ },
167
+ {
168
+ "epoch": 2.062176165803109,
169
+ "eval_accuracy": 0.8074360627552117,
170
+ "eval_f1": 0.5410536307546274,
171
+ "eval_loss": 0.030047627165913582,
172
+ "eval_precision": 0.39832285115303984,
173
+ "eval_recall": 0.8431952662721893,
174
+ "eval_runtime": 3.037,
175
+ "eval_samples_per_second": 112.283,
176
+ "eval_steps_per_second": 3.622,
177
+ "step": 200
178
+ },
179
+ {
180
+ "epoch": 2.1036269430051813,
181
+ "grad_norm": 2.684053421020508,
182
+ "learning_rate": 4.9252098203729154e-05,
183
+ "loss": 0.08770081400871277,
184
+ "step": 204
185
+ },
186
+ {
187
+ "epoch": 2.227979274611399,
188
+ "grad_norm": 21.072328567504883,
189
+ "learning_rate": 4.907428853779167e-05,
190
+ "loss": 0.1440110206604004,
191
+ "step": 216
192
+ },
193
+ {
194
+ "epoch": 2.3523316062176165,
195
+ "grad_norm": 3.221358299255371,
196
+ "learning_rate": 4.887792697727854e-05,
197
+ "loss": 0.11832749843597412,
198
+ "step": 228
199
+ },
200
+ {
201
+ "epoch": 2.476683937823834,
202
+ "grad_norm": 5.4783101081848145,
203
+ "learning_rate": 4.866316484042968e-05,
204
+ "loss": 0.0703568160533905,
205
+ "step": 240
206
+ },
207
+ {
208
+ "epoch": 2.5803108808290154,
209
+ "eval_accuracy": 0.9129593810444874,
210
+ "eval_f1": 0.7118012422360248,
211
+ "eval_loss": 0.0311787910759449,
212
+ "eval_precision": 0.6134903640256959,
213
+ "eval_recall": 0.8476331360946746,
214
+ "eval_runtime": 3.0166,
215
+ "eval_samples_per_second": 113.041,
216
+ "eval_steps_per_second": 3.646,
217
+ "step": 250
218
+ },
219
+ {
220
+ "epoch": 2.6010362694300517,
221
+ "grad_norm": 6.384552478790283,
222
+ "learning_rate": 4.8430167625158595e-05,
223
+ "loss": 0.09287564953168233,
224
+ "step": 252
225
+ },
226
+ {
227
+ "epoch": 2.7253886010362693,
228
+ "grad_norm": 4.217259407043457,
229
+ "learning_rate": 4.817911488151791e-05,
230
+ "loss": 0.08391642570495605,
231
+ "step": 264
232
+ },
233
+ {
234
+ "epoch": 2.849740932642487,
235
+ "grad_norm": 3.1646366119384766,
236
+ "learning_rate": 4.791020007333629e-05,
237
+ "loss": 0.061600317557652794,
238
+ "step": 276
239
+ },
240
+ {
241
+ "epoch": 2.9740932642487046,
242
+ "grad_norm": 6.243884563446045,
243
+ "learning_rate": 4.762363042913322e-05,
244
+ "loss": 0.05470449725786845,
245
+ "step": 288
246
+ },
247
+ {
248
+ "epoch": 3.093264248704663,
249
+ "grad_norm": 2.5704944133758545,
250
+ "learning_rate": 4.7319626782426564e-05,
251
+ "loss": 0.05606028437614441,
252
+ "step": 300
253
+ },
254
+ {
255
+ "epoch": 3.093264248704663,
256
+ "eval_accuracy": 0.7977648828712659,
257
+ "eval_f1": 0.5429740791268759,
258
+ "eval_loss": 0.03227110207080841,
259
+ "eval_precision": 0.39198949441891007,
260
+ "eval_recall": 0.8831360946745562,
261
+ "eval_runtime": 3.0347,
262
+ "eval_samples_per_second": 112.367,
263
+ "eval_steps_per_second": 3.625,
264
+ "step": 300
265
+ },
266
+ {
267
+ "epoch": 3.2176165803108807,
268
+ "grad_norm": 2.420072317123413,
269
+ "learning_rate": 4.699842340155598e-05,
270
+ "loss": 0.053138395150502525,
271
+ "step": 312
272
+ },
273
+ {
274
+ "epoch": 3.3419689119170983,
275
+ "grad_norm": 67.39490509033203,
276
+ "learning_rate": 4.666026780915334e-05,
277
+ "loss": 0.10085086027781169,
278
+ "step": 324
279
+ },
280
+ {
281
+ "epoch": 3.466321243523316,
282
+ "grad_norm": 0.7816835045814514,
283
+ "learning_rate": 4.630542059139924e-05,
284
+ "loss": 0.0743176539738973,
285
+ "step": 336
286
+ },
287
+ {
288
+ "epoch": 3.5906735751295336,
289
+ "grad_norm": 0.7841729521751404,
290
+ "learning_rate": 4.593415519721257e-05,
291
+ "loss": 0.05843483408292135,
292
+ "step": 348
293
+ },
294
+ {
295
+ "epoch": 3.61139896373057,
296
+ "eval_accuracy": 0.9329464861379755,
297
+ "eval_f1": 0.7535070140280562,
298
+ "eval_loss": 0.030000876635313034,
299
+ "eval_precision": 0.6869671132764921,
300
+ "eval_recall": 0.834319526627219,
301
+ "eval_runtime": 3.0255,
302
+ "eval_samples_per_second": 112.709,
303
+ "eval_steps_per_second": 3.636,
304
+ "step": 350
305
+ },
306
+ {
307
+ "epoch": 3.715025906735751,
308
+ "grad_norm": 5.054598331451416,
309
+ "learning_rate": 4.5546757727528035e-05,
310
+ "loss": 0.03325176239013672,
311
+ "step": 360
312
+ },
313
+ {
314
+ "epoch": 3.839378238341969,
315
+ "grad_norm": 0.5077673196792603,
316
+ "learning_rate": 4.5143526714823766e-05,
317
+ "loss": 0.036297169824441276,
318
+ "step": 372
319
+ },
320
+ {
321
+ "epoch": 3.9637305699481864,
322
+ "grad_norm": 0.8186505436897278,
323
+ "learning_rate": 4.472477289306913e-05,
324
+ "loss": 0.026280519862969715,
325
+ "step": 384
326
+ },
327
+ {
328
+ "epoch": 4.082901554404145,
329
+ "grad_norm": 0.9546571373939514,
330
+ "learning_rate": 4.429081895826997e-05,
331
+ "loss": 0.038141210873921715,
332
+ "step": 396
333
+ },
334
+ {
335
+ "epoch": 4.124352331606218,
336
+ "eval_accuracy": 0.9116698903932946,
337
+ "eval_f1": 0.7129798903107862,
338
+ "eval_loss": 0.02928924560546875,
339
+ "eval_precision": 0.6062176165803109,
340
+ "eval_recall": 0.8653846153846154,
341
+ "eval_runtime": 3.0233,
342
+ "eval_samples_per_second": 112.79,
343
+ "eval_steps_per_second": 3.638,
344
+ "step": 400
345
+ },
346
+ {
347
+ "epoch": 4.2072538860103625,
348
+ "grad_norm": 0.38423895835876465,
349
+ "learning_rate": 4.38419993197957e-05,
350
+ "loss": 0.023756501575311024,
351
+ "step": 408
352
+ },
353
+ {
354
+ "epoch": 4.331606217616581,
355
+ "grad_norm": 0.261392205953598,
356
+ "learning_rate": 4.337865984268001e-05,
357
+ "loss": 0.02620025227467219,
358
+ "step": 420
359
+ },
360
+ {
361
+ "epoch": 4.455958549222798,
362
+ "grad_norm": 2.0392212867736816,
363
+ "learning_rate": 4.290115758109375e-05,
364
+ "loss": 0.02181059867143631,
365
+ "step": 432
366
+ },
367
+ {
368
+ "epoch": 4.580310880829016,
369
+ "grad_norm": 0.4685036242008209,
370
+ "learning_rate": 4.240986050319524e-05,
371
+ "loss": 0.023431827624638874,
372
+ "step": 444
373
+ },
374
+ {
375
+ "epoch": 4.642487046632124,
376
+ "eval_accuracy": 0.9593810444874274,
377
+ "eval_f1": 0.828297715549005,
378
+ "eval_loss": 0.04263457655906677,
379
+ "eval_precision": 0.8252569750367107,
380
+ "eval_recall": 0.8313609467455622,
381
+ "eval_runtime": 3.0402,
382
+ "eval_samples_per_second": 112.164,
383
+ "eval_steps_per_second": 3.618,
384
+ "step": 450
385
+ },
386
+ {
387
+ "epoch": 4.704663212435233,
388
+ "grad_norm": 0.535210371017456,
389
+ "learning_rate": 4.19051472075703e-05,
390
+ "loss": 0.021484971046447754,
391
+ "step": 456
392
+ },
393
+ {
394
+ "epoch": 4.829015544041451,
395
+ "grad_norm": 3.4881668090820312,
396
+ "learning_rate": 4.138740663148019e-05,
397
+ "loss": 0.03922188033660253,
398
+ "step": 468
399
+ },
400
+ {
401
+ "epoch": 4.953367875647668,
402
+ "grad_norm": 0.809377908706665,
403
+ "learning_rate": 4.085703775114266e-05,
404
+ "loss": 0.018220727642377216,
405
+ "step": 480
406
+ },
407
+ {
408
+ "epoch": 5.072538860103627,
409
+ "grad_norm": 3.266059160232544,
410
+ "learning_rate": 4.0314449274276663e-05,
411
+ "loss": 0.015343919396400452,
412
+ "step": 492
413
+ },
414
+ {
415
+ "epoch": 5.155440414507772,
416
+ "eval_accuracy": 0.9574468085106383,
417
+ "eval_f1": 0.8162672476397966,
418
+ "eval_loss": 0.04887477308511734,
419
+ "eval_precision": 0.8017118402282454,
420
+ "eval_recall": 0.8313609467455622,
421
+ "eval_runtime": 3.0737,
422
+ "eval_samples_per_second": 110.94,
423
+ "eval_steps_per_second": 3.579,
424
+ "step": 500
425
+ },
426
+ {
427
+ "epoch": 5.196891191709844,
428
+ "grad_norm": 0.14335286617279053,
429
+ "learning_rate": 3.976005932514807e-05,
430
+ "loss": 0.007969564447800318,
431
+ "step": 504
432
+ },
433
+ {
434
+ "epoch": 5.321243523316062,
435
+ "grad_norm": 0.24810202419757843,
436
+ "learning_rate": 3.919429512235874e-05,
437
+ "loss": 0.007458047941327095,
438
+ "step": 516
439
+ },
440
+ {
441
+ "epoch": 5.44559585492228,
442
+ "grad_norm": 1.7369506359100342,
443
+ "learning_rate": 3.86175926496275e-05,
444
+ "loss": 0.012967783957719803,
445
+ "step": 528
446
+ },
447
+ {
448
+ "epoch": 5.569948186528498,
449
+ "grad_norm": 0.40230634808540344,
450
+ "learning_rate": 3.8030396319816585e-05,
451
+ "loss": 0.008079041416446367,
452
+ "step": 540
453
+ },
454
+ {
455
+ "epoch": 5.6735751295336785,
456
+ "eval_accuracy": 0.9591661293788953,
457
+ "eval_f1": 0.8310168251645941,
458
+ "eval_loss": 0.04243873059749603,
459
+ "eval_precision": 0.8219971056439942,
460
+ "eval_recall": 0.8402366863905325,
461
+ "eval_runtime": 3.0267,
462
+ "eval_samples_per_second": 112.662,
463
+ "eval_steps_per_second": 3.634,
464
+ "step": 550
465
+ },
466
+ {
467
+ "epoch": 5.694300518134715,
468
+ "grad_norm": 0.46429362893104553,
469
+ "learning_rate": 3.743315863246256e-05,
470
+ "loss": 0.005399468044439952,
471
+ "step": 552
472
+ },
473
+ {
474
+ "epoch": 5.818652849740933,
475
+ "grad_norm": 1.60989248752594,
476
+ "learning_rate": 3.682633982507543e-05,
477
+ "loss": 0.006222298989693324,
478
+ "step": 564
479
+ },
480
+ {
481
+ "epoch": 5.94300518134715,
482
+ "grad_norm": 0.5653553009033203,
483
+ "learning_rate": 3.621040751847497e-05,
484
+ "loss": 0.015771896888812382,
485
+ "step": 576
486
+ },
487
+ {
488
+ "epoch": 6.062176165803109,
489
+ "grad_norm": 0.34870776534080505,
490
+ "learning_rate": 3.5585836356437264e-05,
491
+ "loss": 0.0066079869866371155,
492
+ "step": 588
493
+ },
494
+ {
495
+ "epoch": 6.186528497409326,
496
+ "grad_norm": 0.2842996120452881,
497
+ "learning_rate": 3.495310763992928e-05,
498
+ "loss": 0.0051950619866450625,
499
+ "step": 600
500
+ },
501
+ {
502
+ "epoch": 6.186528497409326,
503
+ "eval_accuracy": 0.9580915538362347,
504
+ "eval_f1": 0.8308351177730193,
505
+ "eval_loss": 0.04230675473809242,
506
+ "eval_precision": 0.8027586206896552,
507
+ "eval_recall": 0.8609467455621301,
508
+ "eval_runtime": 3.0559,
509
+ "eval_samples_per_second": 111.586,
510
+ "eval_steps_per_second": 3.6,
511
+ "step": 600
512
+ },
513
+ {
514
+ "epoch": 6.310880829015544,
515
+ "grad_norm": 0.20262780785560608,
516
+ "learning_rate": 3.431270895621349e-05,
517
+ "loss": 0.0038215465222795806,
518
+ "step": 612
519
+ },
520
+ {
521
+ "epoch": 6.435233160621761,
522
+ "grad_norm": 0.08969880640506744,
523
+ "learning_rate": 3.3665133803108035e-05,
524
+ "loss": 0.002952239786585172,
525
+ "step": 624
526
+ },
527
+ {
528
+ "epoch": 6.5595854922279795,
529
+ "grad_norm": 7.001764297485352,
530
+ "learning_rate": 3.301088120869224e-05,
531
+ "loss": 0.006995424007376035,
532
+ "step": 636
533
+ },
534
+ {
535
+ "epoch": 6.683937823834197,
536
+ "grad_norm": 1.759656548500061,
537
+ "learning_rate": 3.235045534675045e-05,
538
+ "loss": 0.0032693296670913696,
539
+ "step": 648
540
+ },
541
+ {
542
+ "epoch": 6.704663212435233,
543
+ "eval_accuracy": 0.9537932516655921,
544
+ "eval_f1": 0.8183098591549295,
545
+ "eval_loss": 0.04099448770284653,
546
+ "eval_precision": 0.7809139784946236,
547
+ "eval_recall": 0.8594674556213018,
548
+ "eval_runtime": 3.0102,
549
+ "eval_samples_per_second": 113.281,
550
+ "eval_steps_per_second": 3.654,
551
+ "step": 650
552
+ },
553
+ {
554
+ "epoch": 6.808290155440415,
555
+ "grad_norm": 0.113290935754776,
556
+ "learning_rate": 3.168436514825052e-05,
557
+ "loss": 0.003636385935048262,
558
+ "step": 660
559
+ },
560
+ {
561
+ "epoch": 6.932642487046632,
562
+ "grad_norm": 0.4593416452407837,
563
+ "learning_rate": 3.101312390915634e-05,
564
+ "loss": 0.006359308337171872,
565
+ "step": 672
566
+ },
567
+ {
568
+ "epoch": 7.051813471502591,
569
+ "grad_norm": 0.5328609943389893,
570
+ "learning_rate": 3.0337248894876746e-05,
571
+ "loss": 0.0031539862975478172,
572
+ "step": 684
573
+ },
574
+ {
575
+ "epoch": 7.176165803108808,
576
+ "grad_norm": 0.13898538053035736,
577
+ "learning_rate": 2.9657260941655417e-05,
578
+ "loss": 0.0014738718358178933,
579
+ "step": 696
580
+ },
581
+ {
582
+ "epoch": 7.217616580310881,
583
+ "eval_accuracy": 0.9636793466580701,
584
+ "eval_f1": 0.8498542274052479,
585
+ "eval_loss": 0.041410233825445175,
586
+ "eval_precision": 0.8376436781609196,
587
+ "eval_recall": 0.8624260355029586,
588
+ "eval_runtime": 3.0173,
589
+ "eval_samples_per_second": 113.015,
590
+ "eval_steps_per_second": 3.646,
591
+ "step": 700
592
+ },
593
+ {
594
+ "epoch": 7.300518134715026,
595
+ "grad_norm": 0.020028043538331985,
596
+ "learning_rate": 2.8973684055209106e-05,
597
+ "loss": 0.0009494482850035032,
598
+ "step": 708
599
+ },
600
+ {
601
+ "epoch": 7.424870466321243,
602
+ "grad_norm": 0.038508884608745575,
603
+ "learning_rate": 2.8287045006923473e-05,
604
+ "loss": 0.002579918752113978,
605
+ "step": 720
606
+ },
607
+ {
608
+ "epoch": 7.549222797927461,
609
+ "grad_norm": 0.46958020329475403,
610
+ "learning_rate": 2.7597872927917613e-05,
611
+ "loss": 0.0010905038410176833,
612
+ "step": 732
613
+ },
614
+ {
615
+ "epoch": 7.6735751295336785,
616
+ "grad_norm": 0.1547262817621231,
617
+ "learning_rate": 2.6906698901290167e-05,
618
+ "loss": 0.0028232391923666,
619
+ "step": 744
620
+ },
621
+ {
622
+ "epoch": 7.7357512953367875,
623
+ "eval_accuracy": 0.957016978293574,
624
+ "eval_f1": 0.829302987197724,
625
+ "eval_loss": 0.04088299721479416,
626
+ "eval_precision": 0.7986301369863014,
627
+ "eval_recall": 0.8624260355029586,
628
+ "eval_runtime": 3.0234,
629
+ "eval_samples_per_second": 112.787,
630
+ "eval_steps_per_second": 3.638,
631
+ "step": 750
632
+ },
633
+ {
634
+ "epoch": 7.7979274611398965,
635
+ "grad_norm": 0.0519477054476738,
636
+ "learning_rate": 2.621405555286121e-05,
637
+ "loss": 0.0010418800326685111,
638
+ "step": 756
639
+ },
640
+ {
641
+ "epoch": 7.922279792746114,
642
+ "grad_norm": 1.0287493467330933,
643
+ "learning_rate": 2.5520476640725342e-05,
644
+ "loss": 0.001736572477966547,
645
+ "step": 768
646
+ },
647
+ {
648
+ "epoch": 8.041450777202073,
649
+ "grad_norm": 0.013289994560182095,
650
+ "learning_rate": 2.4826496643932183e-05,
651
+ "loss": 0.000329588889144361,
652
+ "step": 780
653
+ },
654
+ {
655
+ "epoch": 8.16580310880829,
656
+ "grad_norm": 0.011798656545579433,
657
+ "learning_rate": 2.413265035061131e-05,
658
+ "loss": 0.002016165293753147,
659
+ "step": 792
660
+ },
661
+ {
662
+ "epoch": 8.248704663212436,
663
+ "eval_accuracy": 0.9630346013324736,
664
+ "eval_f1": 0.8407079646017699,
665
+ "eval_loss": 0.05126392841339111,
666
+ "eval_precision": 0.8382352941176471,
667
+ "eval_recall": 0.8431952662721893,
668
+ "eval_runtime": 3.0399,
669
+ "eval_samples_per_second": 112.176,
670
+ "eval_steps_per_second": 3.619,
671
+ "step": 800
672
+ },
673
+ {
674
+ "epoch": 8.290155440414507,
675
+ "grad_norm": 0.016478369012475014,
676
+ "learning_rate": 2.343947244585902e-05,
677
+ "loss": 0.001583070649454991,
678
+ "step": 804
679
+ },
680
+ {
681
+ "epoch": 8.414507772020725,
682
+ "grad_norm": 0.08483371138572693,
683
+ "learning_rate": 2.2747497099704533e-05,
684
+ "loss": 0.0006651872924218575,
685
+ "step": 816
686
+ },
687
+ {
688
+ "epoch": 8.538860103626943,
689
+ "grad_norm": 0.02956315688788891,
690
+ "learning_rate": 2.2057257555472998e-05,
691
+ "loss": 0.001361117387811343,
692
+ "step": 828
693
+ },
694
+ {
695
+ "epoch": 8.663212435233161,
696
+ "grad_norm": 0.1703813225030899,
697
+ "learning_rate": 2.136928571886275e-05,
698
+ "loss": 0.0004198816216861208,
699
+ "step": 840
700
+ },
701
+ {
702
+ "epoch": 8.766839378238341,
703
+ "eval_accuracy": 0.9636793466580701,
704
+ "eval_f1": 0.851528384279476,
705
+ "eval_loss": 0.04630826786160469,
706
+ "eval_precision": 0.83810888252149,
707
+ "eval_recall": 0.8653846153846154,
708
+ "eval_runtime": 3.0066,
709
+ "eval_samples_per_second": 113.418,
710
+ "eval_steps_per_second": 3.659,
711
+ "step": 850
712
+ },
713
+ {
714
+ "epoch": 8.787564766839377,
715
+ "grad_norm": 0.08187761902809143,
716
+ "learning_rate": 2.068411174805327e-05,
717
+ "loss": 0.00021340335176015893,
718
+ "step": 852
719
+ },
720
+ {
721
+ "epoch": 8.911917098445596,
722
+ "grad_norm": 0.1566244512796402,
723
+ "learning_rate": 2.0002263645159892e-05,
724
+ "loss": 0.0004732525364185373,
725
+ "step": 864
726
+ },
727
+ {
728
+ "epoch": 9.031088082901555,
729
+ "grad_norm": 0.007061833515763283,
730
+ "learning_rate": 1.9324266849349848e-05,
731
+ "loss": 9.06402322774132e-05,
732
+ "step": 876
733
+ },
734
+ {
735
+ "epoch": 9.155440414507773,
736
+ "grad_norm": 0.010846185497939587,
737
+ "learning_rate": 1.8650643831933502e-05,
738
+ "loss": 0.00011390061505759756,
739
+ "step": 888
740
+ },
741
+ {
742
+ "epoch": 9.279792746113989,
743
+ "grad_norm": 0.03851994872093201,
744
+ "learning_rate": 1.798191369374258e-05,
745
+ "loss": 0.00019203778356313705,
746
+ "step": 900
747
+ },
748
+ {
749
+ "epoch": 9.279792746113989,
750
+ "eval_accuracy": 0.9649688373092629,
751
+ "eval_f1": 0.8546255506607928,
752
+ "eval_loss": 0.05020074546337128,
753
+ "eval_precision": 0.8483965014577259,
754
+ "eval_recall": 0.8609467455621301,
755
+ "eval_runtime": 3.0339,
756
+ "eval_samples_per_second": 112.397,
757
+ "eval_steps_per_second": 3.626,
758
+ "step": 900
759
+ },
760
+ {
761
+ "epoch": 9.404145077720207,
762
+ "grad_norm": 0.006672121584415436,
763
+ "learning_rate": 1.731859176510569e-05,
764
+ "loss": 9.344543407981594e-05,
765
+ "step": 912
766
+ },
767
+ {
768
+ "epoch": 9.528497409326425,
769
+ "grad_norm": 0.028979798778891563,
770
+ "learning_rate": 1.666118920872949e-05,
771
+ "loss": 0.0009806355616698663,
772
+ "step": 924
773
+ },
774
+ {
775
+ "epoch": 9.652849740932643,
776
+ "grad_norm": 0.010359357111155987,
777
+ "learning_rate": 1.6010212625791422e-05,
778
+ "loss": 0.00018761774602656564,
779
+ "step": 936
780
+ },
781
+ {
782
+ "epoch": 9.77720207253886,
783
+ "grad_norm": 0.0016930641140788794,
784
+ "learning_rate": 1.536616366554768e-05,
785
+ "loss": 0.00010497189941816032,
786
+ "step": 948
787
+ },
788
+ {
789
+ "epoch": 9.797927461139896,
790
+ "eval_accuracy": 0.9641091768751343,
791
+ "eval_f1": 0.849112426035503,
792
+ "eval_loss": 0.05002701282501221,
793
+ "eval_precision": 0.849112426035503,
794
+ "eval_recall": 0.849112426035503,
795
+ "eval_runtime": 3.0151,
796
+ "eval_samples_per_second": 113.098,
797
+ "eval_steps_per_second": 3.648,
798
+ "step": 950
799
+ },
800
+ {
801
+ "epoch": 9.901554404145077,
802
+ "grad_norm": 0.0343826562166214,
803
+ "learning_rate": 1.472953863875706e-05,
804
+ "loss": 8.524257767324646e-05,
805
+ "step": 960
806
+ },
807
+ {
808
+ "epoch": 10.020725388601036,
809
+ "grad_norm": 0.006111979950219393,
810
+ "learning_rate": 1.4100828135218829e-05,
811
+ "loss": 0.00012398386024869978,
812
+ "step": 972
813
+ },
814
+ {
815
+ "epoch": 10.145077720207254,
816
+ "grad_norm": 0.010606050491333008,
817
+ "learning_rate": 1.3480516645719172e-05,
818
+ "loss": 3.8064138304131724e-05,
819
+ "step": 984
820
+ },
821
+ {
822
+ "epoch": 10.26943005181347,
823
+ "grad_norm": 0.0009117216104641557,
824
+ "learning_rate": 1.2869082188677572e-05,
825
+ "loss": 0.0001223099728425344,
826
+ "step": 996
827
+ },
828
+ {
829
+ "epoch": 10.310880829015543,
830
+ "eval_accuracy": 0.9645390070921985,
831
+ "eval_f1": 0.8514412416851442,
832
+ "eval_loss": 0.05051806941628456,
833
+ "eval_precision": 0.8508124076809453,
834
+ "eval_recall": 0.8520710059171598,
835
+ "eval_runtime": 3.0202,
836
+ "eval_samples_per_second": 112.906,
837
+ "eval_steps_per_second": 3.642,
838
+ "step": 1000
839
+ },
840
+ {
841
+ "epoch": 10.393782383419689,
842
+ "grad_norm": 0.0027385777793824673,
843
+ "learning_rate": 1.2266995941780934e-05,
844
+ "loss": 7.653955738836278e-05,
845
+ "step": 1008
846
+ },
847
+ {
848
+ "epoch": 10.518134715025907,
849
+ "grad_norm": 0.00011921550321858376,
850
+ "learning_rate": 1.1674721878889171e-05,
851
+ "loss": 0.0001742592042622467,
852
+ "step": 1020
853
+ },
854
+ {
855
+ "epoch": 10.642487046632125,
856
+ "grad_norm": 0.001234560739248991,
857
+ "learning_rate": 1.1092716412492226e-05,
858
+ "loss": 7.672411447856575e-05,
859
+ "step": 1032
860
+ },
861
+ {
862
+ "epoch": 10.766839378238341,
863
+ "grad_norm": 0.0005847270367667079,
864
+ "learning_rate": 1.0521428041993848e-05,
865
+ "loss": 9.932015624751026e-06,
866
+ "step": 1044
867
+ },
868
+ {
869
+ "epoch": 10.82901554404145,
870
+ "eval_accuracy": 0.9643240919836664,
871
+ "eval_f1": 0.8508124076809453,
872
+ "eval_loss": 0.05043612793087959,
873
+ "eval_precision": 0.8495575221238938,
874
+ "eval_recall": 0.8520710059171598,
875
+ "eval_runtime": 3.024,
876
+ "eval_samples_per_second": 112.764,
877
+ "eval_steps_per_second": 3.638,
878
+ "step": 1050
879
+ },
880
+ {
881
+ "epoch": 10.89119170984456,
882
+ "grad_norm": 0.0753927081823349,
883
+ "learning_rate": 9.961297008093357e-06,
884
+ "loss": 0.00011471325221161048,
885
+ "step": 1056
886
+ },
887
+ {
888
+ "epoch": 11.010362694300518,
889
+ "grad_norm": 0.003958659712225199,
890
+ "learning_rate": 9.412754953531663e-06,
891
+ "loss": 1.709427791259562e-05,
892
+ "step": 1068
893
+ },
894
+ {
895
+ "epoch": 11.134715025906736,
896
+ "grad_norm": 0.001931481878273189,
897
+ "learning_rate": 8.876224590462873e-06,
898
+ "loss": 1.0689032933441922e-05,
899
+ "step": 1080
900
+ },
901
+ {
902
+ "epoch": 11.259067357512953,
903
+ "grad_norm": 0.012324509210884571,
904
+ "learning_rate": 8.352119374707978e-06,
905
+ "loss": 5.9023882689264916e-05,
906
+ "step": 1092
907
+ },
908
+ {
909
+ "epoch": 11.341968911917098,
910
+ "eval_accuracy": 0.9643240919836664,
911
+ "eval_f1": 0.8510324483775811,
912
+ "eval_loss": 0.050491511821746826,
913
+ "eval_precision": 0.8485294117647059,
914
+ "eval_recall": 0.8535502958579881,
915
+ "eval_runtime": 3.0303,
916
+ "eval_samples_per_second": 112.529,
917
+ "eval_steps_per_second": 3.63,
918
+ "step": 1100
919
+ },
920
+ {
921
+ "epoch": 11.38341968911917,
922
+ "grad_norm": 0.002621831139549613,
923
+ "learning_rate": 7.840843187141544e-06,
924
+ "loss": 5.975224000091354e-05,
925
+ "step": 1104
926
+ },
927
+ {
928
+ "epoch": 11.507772020725389,
929
+ "grad_norm": 0.0006910113152116537,
930
+ "learning_rate": 7.3427900224569606e-06,
931
+ "loss": 7.24908944296961e-05,
932
+ "step": 1116
933
+ },
934
+ {
935
+ "epoch": 11.632124352331607,
936
+ "grad_norm": 0.01195142138749361,
937
+ "learning_rate": 6.858343685550048e-06,
938
+ "loss": 3.2996586620962866e-05,
939
+ "step": 1128
940
+ },
941
+ {
942
+ "epoch": 11.756476683937823,
943
+ "grad_norm": 0.00024170154938474298,
944
+ "learning_rate": 6.387877495755065e-06,
945
+ "loss": 1.4914168180742612e-05,
946
+ "step": 1140
947
+ },
948
+ {
949
+ "epoch": 11.860103626943005,
950
+ "eval_accuracy": 0.9645390070921985,
951
+ "eval_f1": 0.8514412416851442,
952
+ "eval_loss": 0.05080362409353256,
953
+ "eval_precision": 0.8508124076809453,
954
+ "eval_recall": 0.8520710059171598,
955
+ "eval_runtime": 3.0169,
956
+ "eval_samples_per_second": 113.03,
957
+ "eval_steps_per_second": 3.646,
958
+ "step": 1150
959
+ },
960
+ {
961
+ "epoch": 11.880829015544041,
962
+ "grad_norm": 0.00013972603483125567,
963
+ "learning_rate": 5.93175399916101e-06,
964
+ "loss": 0.00017348669158915678,
965
+ "step": 1152
966
+ },
967
+ {
968
+ "epoch": 12.0,
969
+ "grad_norm": 0.014590195380151272,
970
+ "learning_rate": 5.490324689229834e-06,
971
+ "loss": 6.250141692968707e-05,
972
+ "step": 1164
973
+ },
974
+ {
975
+ "epoch": 12.124352331606218,
976
+ "grad_norm": 0.002015709411352873,
977
+ "learning_rate": 5.063929735931985e-06,
978
+ "loss": 1.4617435226682574e-05,
979
+ "step": 1176
980
+ },
981
+ {
982
+ "epoch": 12.248704663212436,
983
+ "grad_norm": 0.0005398832727223635,
984
+ "learning_rate": 4.652897723607913e-06,
985
+ "loss": 1.6918191249715164e-05,
986
+ "step": 1188
987
+ },
988
+ {
989
+ "epoch": 12.373056994818652,
990
+ "grad_norm": 0.0006741413380950689,
991
+ "learning_rate": 4.257545397757623e-06,
992
+ "loss": 4.216020170133561e-05,
993
+ "step": 1200
994
+ },
995
+ {
996
+ "epoch": 12.373056994818652,
997
+ "eval_accuracy": 0.9645390070921985,
998
+ "eval_f1": 0.8514412416851442,
999
+ "eval_loss": 0.05097455903887749,
1000
+ "eval_precision": 0.8508124076809453,
1001
+ "eval_recall": 0.8520710059171598,
1002
+ "eval_runtime": 3.0165,
1003
+ "eval_samples_per_second": 113.043,
1004
+ "eval_steps_per_second": 3.647,
1005
+ "step": 1200
1006
+ },
1007
+ {
1008
+ "epoch": 12.49740932642487,
1009
+ "grad_norm": 0.0007238159305416048,
1010
+ "learning_rate": 3.878177420953291e-06,
1011
+ "loss": 5.365341106274476e-05,
1012
+ "step": 1212
1013
+ },
1014
+ {
1015
+ "epoch": 12.621761658031089,
1016
+ "grad_norm": 0.002653967123478651,
1017
+ "learning_rate": 3.5150861380631755e-06,
1018
+ "loss": 3.87875431139643e-05,
1019
+ "step": 1224
1020
+ },
1021
+ {
1022
+ "epoch": 12.746113989637305,
1023
+ "grad_norm": 0.0012897817650809884,
1024
+ "learning_rate": 3.1685513509676606e-06,
1025
+ "loss": 1.6210615285672247e-05,
1026
+ "step": 1236
1027
+ },
1028
+ {
1029
+ "epoch": 12.870466321243523,
1030
+ "grad_norm": 0.010877900756895542,
1031
+ "learning_rate": 2.838840102941015e-06,
1032
+ "loss": 4.428145863736669e-05,
1033
+ "step": 1248
1034
+ },
1035
+ {
1036
+ "epoch": 12.89119170984456,
1037
+ "eval_accuracy": 0.9645390070921985,
1038
+ "eval_f1": 0.8514412416851442,
1039
+ "eval_loss": 0.05106152221560478,
1040
+ "eval_precision": 0.8508124076809453,
1041
+ "eval_recall": 0.8520710059171598,
1042
+ "eval_runtime": 3.0194,
1043
+ "eval_samples_per_second": 112.937,
1044
+ "eval_steps_per_second": 3.643,
1045
+ "step": 1250
1046
+ },
1047
+ {
1048
+ "epoch": 12.994818652849741,
1049
+ "grad_norm": 0.004842772614210844,
1050
+ "learning_rate": 2.52620647286512e-06,
1051
+ "loss": 4.890701772334675e-05,
1052
+ "step": 1260
1053
+ },
1054
+ {
1055
+ "epoch": 13.1139896373057,
1056
+ "grad_norm": 0.005717657506465912,
1057
+ "learning_rate": 2.2308913794336583e-06,
1058
+ "loss": 8.704484207555652e-05,
1059
+ "step": 1272
1060
+ },
1061
+ {
1062
+ "epoch": 13.238341968911918,
1063
+ "grad_norm": 0.0031942392233759165,
1064
+ "learning_rate": 1.953122395497711e-06,
1065
+ "loss": 3.180075388324136e-05,
1066
+ "step": 1284
1067
+ },
1068
+ {
1069
+ "epoch": 13.362694300518134,
1070
+ "grad_norm": 0.006513912230730057,
1071
+ "learning_rate": 1.6931135726957315e-06,
1072
+ "loss": 1.0696584164785842e-05,
1073
+ "step": 1296
1074
+ },
1075
+ {
1076
+ "epoch": 13.404145077720207,
1077
+ "eval_accuracy": 0.9645390070921985,
1078
+ "eval_f1": 0.8514412416851442,
1079
+ "eval_loss": 0.05115008354187012,
1080
+ "eval_precision": 0.8508124076809453,
1081
+ "eval_recall": 0.8520710059171598,
1082
+ "eval_runtime": 3.0366,
1083
+ "eval_samples_per_second": 112.298,
1084
+ "eval_steps_per_second": 3.623,
1085
+ "step": 1300
1086
+ }
1087
+ ],
1088
+ "logging_steps": 12,
1089
+ "max_steps": 1455,
1090
+ "num_input_tokens_seen": 0,
1091
+ "num_train_epochs": 15,
1092
+ "save_steps": 50,
1093
+ "stateful_callbacks": {
1094
+ "EarlyStoppingCallback": {
1095
+ "args": {
1096
+ "early_stopping_patience": 10,
1097
+ "early_stopping_threshold": 0.0
1098
+ },
1099
+ "attributes": {
1100
+ "early_stopping_patience_counter": 8
1101
+ }
1102
+ },
1103
+ "TrainerControl": {
1104
+ "args": {
1105
+ "should_epoch_stop": false,
1106
+ "should_evaluate": false,
1107
+ "should_log": false,
1108
+ "should_save": true,
1109
+ "should_training_stop": false
1110
+ },
1111
+ "attributes": {}
1112
+ }
1113
+ },
1114
+ "total_flos": 2.186675218009651e+16,
1115
+ "train_batch_size": 8,
1116
+ "trial_name": null,
1117
+ "trial_params": null
1118
+ }
checkpoint-1300/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:32e421217e1eea826c4025f111c9f4fa0b9d53a8bca71632cae9ad6aed6b26f4
3
+ size 5201
checkpoint-1350/config.json ADDED
@@ -0,0 +1,99 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "architectures": [
3
+ "ModernBertForTokenClassification"
4
+ ],
5
+ "attention_bias": false,
6
+ "attention_dropout": 0.0,
7
+ "bos_token_id": 50281,
8
+ "classifier_activation": "gelu",
9
+ "classifier_bias": false,
10
+ "classifier_dropout": 0.3,
11
+ "classifier_pooling": "mean",
12
+ "cls_token_id": 50281,
13
+ "decoder_bias": true,
14
+ "deterministic_flash_attn": false,
15
+ "dtype": "float32",
16
+ "embedding_dropout": 0.0,
17
+ "eos_token_id": 50282,
18
+ "global_attn_every_n_layers": 3,
19
+ "gradient_checkpointing": false,
20
+ "hidden_activation": "gelu",
21
+ "hidden_size": 1024,
22
+ "id2label": {
23
+ "0": "O",
24
+ "1": "FP",
25
+ "2": "RP",
26
+ "3": "RV",
27
+ "4": "PW"
28
+ },
29
+ "initializer_cutoff_factor": 2.0,
30
+ "initializer_range": 0.02,
31
+ "intermediate_size": 2624,
32
+ "label2id": {
33
+ "FP": 1,
34
+ "O": 0,
35
+ "PW": 4,
36
+ "RP": 2,
37
+ "RV": 3
38
+ },
39
+ "layer_norm_eps": 1e-05,
40
+ "layer_types": [
41
+ "full_attention",
42
+ "sliding_attention",
43
+ "sliding_attention",
44
+ "full_attention",
45
+ "sliding_attention",
46
+ "sliding_attention",
47
+ "full_attention",
48
+ "sliding_attention",
49
+ "sliding_attention",
50
+ "full_attention",
51
+ "sliding_attention",
52
+ "sliding_attention",
53
+ "full_attention",
54
+ "sliding_attention",
55
+ "sliding_attention",
56
+ "full_attention",
57
+ "sliding_attention",
58
+ "sliding_attention",
59
+ "full_attention",
60
+ "sliding_attention",
61
+ "sliding_attention",
62
+ "full_attention",
63
+ "sliding_attention",
64
+ "sliding_attention",
65
+ "full_attention",
66
+ "sliding_attention",
67
+ "sliding_attention",
68
+ "full_attention"
69
+ ],
70
+ "local_attention": 128,
71
+ "max_position_embeddings": 8192,
72
+ "mlp_bias": false,
73
+ "mlp_dropout": 0.0,
74
+ "model_type": "modernbert",
75
+ "norm_bias": false,
76
+ "norm_eps": 1e-05,
77
+ "num_attention_heads": 16,
78
+ "num_hidden_layers": 28,
79
+ "pad_token_id": 50283,
80
+ "position_embedding_type": "absolute",
81
+ "repad_logits_with_grad": false,
82
+ "rope_parameters": {
83
+ "full_attention": {
84
+ "rope_theta": 160000.0,
85
+ "rope_type": "default"
86
+ },
87
+ "sliding_attention": {
88
+ "rope_theta": 10000.0,
89
+ "rope_type": "default"
90
+ }
91
+ },
92
+ "sep_token_id": 50282,
93
+ "sparse_pred_ignore_index": -100,
94
+ "sparse_prediction": false,
95
+ "tie_word_embeddings": true,
96
+ "transformers_version": "5.0.0",
97
+ "use_cache": false,
98
+ "vocab_size": 50368
99
+ }
checkpoint-1350/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e1d3c8d0e6df65f2c972fd0c87f6daa0a37b4babed4f2cc25c658eedd7edc347
3
+ size 1583363940
checkpoint-1350/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:71212c04e8ca6f276995294ba396d27c97a9b56f5a63a9a4a32d536893349e7e
3
+ size 1112419851
checkpoint-1350/rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5de33cb8e076033269f2ec33b8b28007426020c9486c01e0012591b6d3a72dbe
3
+ size 14709
checkpoint-1350/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ce882db0a748e79447221d3afedae8c50336e129b597c06ead5f3f97886558bc
3
+ size 1465
checkpoint-1350/tokenizer.json ADDED
The diff for this file is too large to render. See raw diff
 
checkpoint-1350/tokenizer_config.json ADDED
@@ -0,0 +1,16 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "backend": "tokenizers",
3
+ "clean_up_tokenization_spaces": true,
4
+ "cls_token": "[CLS]",
5
+ "is_local": false,
6
+ "mask_token": "[MASK]",
7
+ "model_input_names": [
8
+ "input_ids",
9
+ "attention_mask"
10
+ ],
11
+ "model_max_length": 8192,
12
+ "pad_token": "[PAD]",
13
+ "sep_token": "[SEP]",
14
+ "tokenizer_class": "TokenizersBackend",
15
+ "unk_token": "[UNK]"
16
+ }
checkpoint-1350/trainer_state.json ADDED
@@ -0,0 +1,1158 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_global_step": 900,
3
+ "best_metric": 0.8546255506607928,
4
+ "best_model_checkpoint": "/content/drive/MyDrive/expE_large_oversample_rv/checkpoint-900",
5
+ "epoch": 13.922279792746114,
6
+ "eval_steps": 50,
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": 0.010362694300518135,
14
+ "grad_norm": 19.391862869262695,
15
+ "learning_rate": 0.0,
16
+ "loss": 0.8294010162353516,
17
+ "step": 1
18
+ },
19
+ {
20
+ "epoch": 0.12435233160621761,
21
+ "grad_norm": 7.022524833679199,
22
+ "learning_rate": 5.670103092783505e-06,
23
+ "loss": 0.6434053074229847,
24
+ "step": 12
25
+ },
26
+ {
27
+ "epoch": 0.24870466321243523,
28
+ "grad_norm": 2.0728249549865723,
29
+ "learning_rate": 1.1855670103092783e-05,
30
+ "loss": 0.5182687441507975,
31
+ "step": 24
32
+ },
33
+ {
34
+ "epoch": 0.37305699481865284,
35
+ "grad_norm": 8.758841514587402,
36
+ "learning_rate": 1.8041237113402062e-05,
37
+ "loss": 0.39159242312113446,
38
+ "step": 36
39
+ },
40
+ {
41
+ "epoch": 0.49740932642487046,
42
+ "grad_norm": 10.565062522888184,
43
+ "learning_rate": 2.422680412371134e-05,
44
+ "loss": 0.31471407413482666,
45
+ "step": 48
46
+ },
47
+ {
48
+ "epoch": 0.5181347150259067,
49
+ "eval_accuracy": 0.7721899849559424,
50
+ "eval_f1": 0.4335260115606937,
51
+ "eval_loss": 0.05006009340286255,
52
+ "eval_precision": 0.32142857142857145,
53
+ "eval_recall": 0.665680473372781,
54
+ "eval_runtime": 5.6283,
55
+ "eval_samples_per_second": 60.586,
56
+ "eval_steps_per_second": 1.954,
57
+ "step": 50
58
+ },
59
+ {
60
+ "epoch": 0.6217616580310881,
61
+ "grad_norm": 5.744125843048096,
62
+ "learning_rate": 3.0412371134020617e-05,
63
+ "loss": 0.28232719500859577,
64
+ "step": 60
65
+ },
66
+ {
67
+ "epoch": 0.7461139896373057,
68
+ "grad_norm": 4.099153518676758,
69
+ "learning_rate": 3.6597938144329896e-05,
70
+ "loss": 0.22093822558720908,
71
+ "step": 72
72
+ },
73
+ {
74
+ "epoch": 0.8704663212435233,
75
+ "grad_norm": 12.179475784301758,
76
+ "learning_rate": 4.2783505154639175e-05,
77
+ "loss": 0.2065215309460958,
78
+ "step": 84
79
+ },
80
+ {
81
+ "epoch": 0.9948186528497409,
82
+ "grad_norm": 8.685556411743164,
83
+ "learning_rate": 4.8969072164948454e-05,
84
+ "loss": 0.16811863581339517,
85
+ "step": 96
86
+ },
87
+ {
88
+ "epoch": 1.0310880829015545,
89
+ "eval_accuracy": 0.7461852568235547,
90
+ "eval_f1": 0.4638047138047138,
91
+ "eval_loss": 0.03645005077123642,
92
+ "eval_precision": 0.3241176470588235,
93
+ "eval_recall": 0.8150887573964497,
94
+ "eval_runtime": 3.0466,
95
+ "eval_samples_per_second": 111.926,
96
+ "eval_steps_per_second": 3.611,
97
+ "step": 100
98
+ },
99
+ {
100
+ "epoch": 1.1139896373056994,
101
+ "grad_norm": 33.46553421020508,
102
+ "learning_rate": 4.999331054325924e-05,
103
+ "loss": 0.15307321151097616,
104
+ "step": 108
105
+ },
106
+ {
107
+ "epoch": 1.238341968911917,
108
+ "grad_norm": 5.198877811431885,
109
+ "learning_rate": 4.996762857385083e-05,
110
+ "loss": 0.1403733491897583,
111
+ "step": 120
112
+ },
113
+ {
114
+ "epoch": 1.3626943005181347,
115
+ "grad_norm": 1.8454481363296509,
116
+ "learning_rate": 4.9922706292838655e-05,
117
+ "loss": 0.13873055577278137,
118
+ "step": 132
119
+ },
120
+ {
121
+ "epoch": 1.4870466321243523,
122
+ "grad_norm": 2.0861477851867676,
123
+ "learning_rate": 4.9858578317794894e-05,
124
+ "loss": 0.11305248737335205,
125
+ "step": 144
126
+ },
127
+ {
128
+ "epoch": 1.549222797927461,
129
+ "eval_accuracy": 0.895766172361917,
130
+ "eval_f1": 0.6706515242080096,
131
+ "eval_loss": 0.029485274106264114,
132
+ "eval_precision": 0.5626880641925778,
133
+ "eval_recall": 0.8298816568047337,
134
+ "eval_runtime": 3.0006,
135
+ "eval_samples_per_second": 113.645,
136
+ "eval_steps_per_second": 3.666,
137
+ "step": 150
138
+ },
139
+ {
140
+ "epoch": 1.61139896373057,
141
+ "grad_norm": 6.362898349761963,
142
+ "learning_rate": 4.9775294066397274e-05,
143
+ "loss": 0.11805283029874165,
144
+ "step": 156
145
+ },
146
+ {
147
+ "epoch": 1.7357512953367875,
148
+ "grad_norm": 2.3624749183654785,
149
+ "learning_rate": 4.967291771834727e-05,
150
+ "loss": 0.11144719521204631,
151
+ "step": 168
152
+ },
153
+ {
154
+ "epoch": 1.8601036269430051,
155
+ "grad_norm": 0.9627630710601807,
156
+ "learning_rate": 4.955152816591259e-05,
157
+ "loss": 0.0819690078496933,
158
+ "step": 180
159
+ },
160
+ {
161
+ "epoch": 1.9844559585492227,
162
+ "grad_norm": 15.513121604919434,
163
+ "learning_rate": 4.9411218953131964e-05,
164
+ "loss": 0.10844972729682922,
165
+ "step": 192
166
+ },
167
+ {
168
+ "epoch": 2.062176165803109,
169
+ "eval_accuracy": 0.8074360627552117,
170
+ "eval_f1": 0.5410536307546274,
171
+ "eval_loss": 0.030047627165913582,
172
+ "eval_precision": 0.39832285115303984,
173
+ "eval_recall": 0.8431952662721893,
174
+ "eval_runtime": 3.037,
175
+ "eval_samples_per_second": 112.283,
176
+ "eval_steps_per_second": 3.622,
177
+ "step": 200
178
+ },
179
+ {
180
+ "epoch": 2.1036269430051813,
181
+ "grad_norm": 2.684053421020508,
182
+ "learning_rate": 4.9252098203729154e-05,
183
+ "loss": 0.08770081400871277,
184
+ "step": 204
185
+ },
186
+ {
187
+ "epoch": 2.227979274611399,
188
+ "grad_norm": 21.072328567504883,
189
+ "learning_rate": 4.907428853779167e-05,
190
+ "loss": 0.1440110206604004,
191
+ "step": 216
192
+ },
193
+ {
194
+ "epoch": 2.3523316062176165,
195
+ "grad_norm": 3.221358299255371,
196
+ "learning_rate": 4.887792697727854e-05,
197
+ "loss": 0.11832749843597412,
198
+ "step": 228
199
+ },
200
+ {
201
+ "epoch": 2.476683937823834,
202
+ "grad_norm": 5.4783101081848145,
203
+ "learning_rate": 4.866316484042968e-05,
204
+ "loss": 0.0703568160533905,
205
+ "step": 240
206
+ },
207
+ {
208
+ "epoch": 2.5803108808290154,
209
+ "eval_accuracy": 0.9129593810444874,
210
+ "eval_f1": 0.7118012422360248,
211
+ "eval_loss": 0.0311787910759449,
212
+ "eval_precision": 0.6134903640256959,
213
+ "eval_recall": 0.8476331360946746,
214
+ "eval_runtime": 3.0166,
215
+ "eval_samples_per_second": 113.041,
216
+ "eval_steps_per_second": 3.646,
217
+ "step": 250
218
+ },
219
+ {
220
+ "epoch": 2.6010362694300517,
221
+ "grad_norm": 6.384552478790283,
222
+ "learning_rate": 4.8430167625158595e-05,
223
+ "loss": 0.09287564953168233,
224
+ "step": 252
225
+ },
226
+ {
227
+ "epoch": 2.7253886010362693,
228
+ "grad_norm": 4.217259407043457,
229
+ "learning_rate": 4.817911488151791e-05,
230
+ "loss": 0.08391642570495605,
231
+ "step": 264
232
+ },
233
+ {
234
+ "epoch": 2.849740932642487,
235
+ "grad_norm": 3.1646366119384766,
236
+ "learning_rate": 4.791020007333629e-05,
237
+ "loss": 0.061600317557652794,
238
+ "step": 276
239
+ },
240
+ {
241
+ "epoch": 2.9740932642487046,
242
+ "grad_norm": 6.243884563446045,
243
+ "learning_rate": 4.762363042913322e-05,
244
+ "loss": 0.05470449725786845,
245
+ "step": 288
246
+ },
247
+ {
248
+ "epoch": 3.093264248704663,
249
+ "grad_norm": 2.5704944133758545,
250
+ "learning_rate": 4.7319626782426564e-05,
251
+ "loss": 0.05606028437614441,
252
+ "step": 300
253
+ },
254
+ {
255
+ "epoch": 3.093264248704663,
256
+ "eval_accuracy": 0.7977648828712659,
257
+ "eval_f1": 0.5429740791268759,
258
+ "eval_loss": 0.03227110207080841,
259
+ "eval_precision": 0.39198949441891007,
260
+ "eval_recall": 0.8831360946745562,
261
+ "eval_runtime": 3.0347,
262
+ "eval_samples_per_second": 112.367,
263
+ "eval_steps_per_second": 3.625,
264
+ "step": 300
265
+ },
266
+ {
267
+ "epoch": 3.2176165803108807,
268
+ "grad_norm": 2.420072317123413,
269
+ "learning_rate": 4.699842340155598e-05,
270
+ "loss": 0.053138395150502525,
271
+ "step": 312
272
+ },
273
+ {
274
+ "epoch": 3.3419689119170983,
275
+ "grad_norm": 67.39490509033203,
276
+ "learning_rate": 4.666026780915334e-05,
277
+ "loss": 0.10085086027781169,
278
+ "step": 324
279
+ },
280
+ {
281
+ "epoch": 3.466321243523316,
282
+ "grad_norm": 0.7816835045814514,
283
+ "learning_rate": 4.630542059139924e-05,
284
+ "loss": 0.0743176539738973,
285
+ "step": 336
286
+ },
287
+ {
288
+ "epoch": 3.5906735751295336,
289
+ "grad_norm": 0.7841729521751404,
290
+ "learning_rate": 4.593415519721257e-05,
291
+ "loss": 0.05843483408292135,
292
+ "step": 348
293
+ },
294
+ {
295
+ "epoch": 3.61139896373057,
296
+ "eval_accuracy": 0.9329464861379755,
297
+ "eval_f1": 0.7535070140280562,
298
+ "eval_loss": 0.030000876635313034,
299
+ "eval_precision": 0.6869671132764921,
300
+ "eval_recall": 0.834319526627219,
301
+ "eval_runtime": 3.0255,
302
+ "eval_samples_per_second": 112.709,
303
+ "eval_steps_per_second": 3.636,
304
+ "step": 350
305
+ },
306
+ {
307
+ "epoch": 3.715025906735751,
308
+ "grad_norm": 5.054598331451416,
309
+ "learning_rate": 4.5546757727528035e-05,
310
+ "loss": 0.03325176239013672,
311
+ "step": 360
312
+ },
313
+ {
314
+ "epoch": 3.839378238341969,
315
+ "grad_norm": 0.5077673196792603,
316
+ "learning_rate": 4.5143526714823766e-05,
317
+ "loss": 0.036297169824441276,
318
+ "step": 372
319
+ },
320
+ {
321
+ "epoch": 3.9637305699481864,
322
+ "grad_norm": 0.8186505436897278,
323
+ "learning_rate": 4.472477289306913e-05,
324
+ "loss": 0.026280519862969715,
325
+ "step": 384
326
+ },
327
+ {
328
+ "epoch": 4.082901554404145,
329
+ "grad_norm": 0.9546571373939514,
330
+ "learning_rate": 4.429081895826997e-05,
331
+ "loss": 0.038141210873921715,
332
+ "step": 396
333
+ },
334
+ {
335
+ "epoch": 4.124352331606218,
336
+ "eval_accuracy": 0.9116698903932946,
337
+ "eval_f1": 0.7129798903107862,
338
+ "eval_loss": 0.02928924560546875,
339
+ "eval_precision": 0.6062176165803109,
340
+ "eval_recall": 0.8653846153846154,
341
+ "eval_runtime": 3.0233,
342
+ "eval_samples_per_second": 112.79,
343
+ "eval_steps_per_second": 3.638,
344
+ "step": 400
345
+ },
346
+ {
347
+ "epoch": 4.2072538860103625,
348
+ "grad_norm": 0.38423895835876465,
349
+ "learning_rate": 4.38419993197957e-05,
350
+ "loss": 0.023756501575311024,
351
+ "step": 408
352
+ },
353
+ {
354
+ "epoch": 4.331606217616581,
355
+ "grad_norm": 0.261392205953598,
356
+ "learning_rate": 4.337865984268001e-05,
357
+ "loss": 0.02620025227467219,
358
+ "step": 420
359
+ },
360
+ {
361
+ "epoch": 4.455958549222798,
362
+ "grad_norm": 2.0392212867736816,
363
+ "learning_rate": 4.290115758109375e-05,
364
+ "loss": 0.02181059867143631,
365
+ "step": 432
366
+ },
367
+ {
368
+ "epoch": 4.580310880829016,
369
+ "grad_norm": 0.4685036242008209,
370
+ "learning_rate": 4.240986050319524e-05,
371
+ "loss": 0.023431827624638874,
372
+ "step": 444
373
+ },
374
+ {
375
+ "epoch": 4.642487046632124,
376
+ "eval_accuracy": 0.9593810444874274,
377
+ "eval_f1": 0.828297715549005,
378
+ "eval_loss": 0.04263457655906677,
379
+ "eval_precision": 0.8252569750367107,
380
+ "eval_recall": 0.8313609467455622,
381
+ "eval_runtime": 3.0402,
382
+ "eval_samples_per_second": 112.164,
383
+ "eval_steps_per_second": 3.618,
384
+ "step": 450
385
+ },
386
+ {
387
+ "epoch": 4.704663212435233,
388
+ "grad_norm": 0.535210371017456,
389
+ "learning_rate": 4.19051472075703e-05,
390
+ "loss": 0.021484971046447754,
391
+ "step": 456
392
+ },
393
+ {
394
+ "epoch": 4.829015544041451,
395
+ "grad_norm": 3.4881668090820312,
396
+ "learning_rate": 4.138740663148019e-05,
397
+ "loss": 0.03922188033660253,
398
+ "step": 468
399
+ },
400
+ {
401
+ "epoch": 4.953367875647668,
402
+ "grad_norm": 0.809377908706665,
403
+ "learning_rate": 4.085703775114266e-05,
404
+ "loss": 0.018220727642377216,
405
+ "step": 480
406
+ },
407
+ {
408
+ "epoch": 5.072538860103627,
409
+ "grad_norm": 3.266059160232544,
410
+ "learning_rate": 4.0314449274276663e-05,
411
+ "loss": 0.015343919396400452,
412
+ "step": 492
413
+ },
414
+ {
415
+ "epoch": 5.155440414507772,
416
+ "eval_accuracy": 0.9574468085106383,
417
+ "eval_f1": 0.8162672476397966,
418
+ "eval_loss": 0.04887477308511734,
419
+ "eval_precision": 0.8017118402282454,
420
+ "eval_recall": 0.8313609467455622,
421
+ "eval_runtime": 3.0737,
422
+ "eval_samples_per_second": 110.94,
423
+ "eval_steps_per_second": 3.579,
424
+ "step": 500
425
+ },
426
+ {
427
+ "epoch": 5.196891191709844,
428
+ "grad_norm": 0.14335286617279053,
429
+ "learning_rate": 3.976005932514807e-05,
430
+ "loss": 0.007969564447800318,
431
+ "step": 504
432
+ },
433
+ {
434
+ "epoch": 5.321243523316062,
435
+ "grad_norm": 0.24810202419757843,
436
+ "learning_rate": 3.919429512235874e-05,
437
+ "loss": 0.007458047941327095,
438
+ "step": 516
439
+ },
440
+ {
441
+ "epoch": 5.44559585492228,
442
+ "grad_norm": 1.7369506359100342,
443
+ "learning_rate": 3.86175926496275e-05,
444
+ "loss": 0.012967783957719803,
445
+ "step": 528
446
+ },
447
+ {
448
+ "epoch": 5.569948186528498,
449
+ "grad_norm": 0.40230634808540344,
450
+ "learning_rate": 3.8030396319816585e-05,
451
+ "loss": 0.008079041416446367,
452
+ "step": 540
453
+ },
454
+ {
455
+ "epoch": 5.6735751295336785,
456
+ "eval_accuracy": 0.9591661293788953,
457
+ "eval_f1": 0.8310168251645941,
458
+ "eval_loss": 0.04243873059749603,
459
+ "eval_precision": 0.8219971056439942,
460
+ "eval_recall": 0.8402366863905325,
461
+ "eval_runtime": 3.0267,
462
+ "eval_samples_per_second": 112.662,
463
+ "eval_steps_per_second": 3.634,
464
+ "step": 550
465
+ },
466
+ {
467
+ "epoch": 5.694300518134715,
468
+ "grad_norm": 0.46429362893104553,
469
+ "learning_rate": 3.743315863246256e-05,
470
+ "loss": 0.005399468044439952,
471
+ "step": 552
472
+ },
473
+ {
474
+ "epoch": 5.818652849740933,
475
+ "grad_norm": 1.60989248752594,
476
+ "learning_rate": 3.682633982507543e-05,
477
+ "loss": 0.006222298989693324,
478
+ "step": 564
479
+ },
480
+ {
481
+ "epoch": 5.94300518134715,
482
+ "grad_norm": 0.5653553009033203,
483
+ "learning_rate": 3.621040751847497e-05,
484
+ "loss": 0.015771896888812382,
485
+ "step": 576
486
+ },
487
+ {
488
+ "epoch": 6.062176165803109,
489
+ "grad_norm": 0.34870776534080505,
490
+ "learning_rate": 3.5585836356437264e-05,
491
+ "loss": 0.0066079869866371155,
492
+ "step": 588
493
+ },
494
+ {
495
+ "epoch": 6.186528497409326,
496
+ "grad_norm": 0.2842996120452881,
497
+ "learning_rate": 3.495310763992928e-05,
498
+ "loss": 0.0051950619866450625,
499
+ "step": 600
500
+ },
501
+ {
502
+ "epoch": 6.186528497409326,
503
+ "eval_accuracy": 0.9580915538362347,
504
+ "eval_f1": 0.8308351177730193,
505
+ "eval_loss": 0.04230675473809242,
506
+ "eval_precision": 0.8027586206896552,
507
+ "eval_recall": 0.8609467455621301,
508
+ "eval_runtime": 3.0559,
509
+ "eval_samples_per_second": 111.586,
510
+ "eval_steps_per_second": 3.6,
511
+ "step": 600
512
+ },
513
+ {
514
+ "epoch": 6.310880829015544,
515
+ "grad_norm": 0.20262780785560608,
516
+ "learning_rate": 3.431270895621349e-05,
517
+ "loss": 0.0038215465222795806,
518
+ "step": 612
519
+ },
520
+ {
521
+ "epoch": 6.435233160621761,
522
+ "grad_norm": 0.08969880640506744,
523
+ "learning_rate": 3.3665133803108035e-05,
524
+ "loss": 0.002952239786585172,
525
+ "step": 624
526
+ },
527
+ {
528
+ "epoch": 6.5595854922279795,
529
+ "grad_norm": 7.001764297485352,
530
+ "learning_rate": 3.301088120869224e-05,
531
+ "loss": 0.006995424007376035,
532
+ "step": 636
533
+ },
534
+ {
535
+ "epoch": 6.683937823834197,
536
+ "grad_norm": 1.759656548500061,
537
+ "learning_rate": 3.235045534675045e-05,
538
+ "loss": 0.0032693296670913696,
539
+ "step": 648
540
+ },
541
+ {
542
+ "epoch": 6.704663212435233,
543
+ "eval_accuracy": 0.9537932516655921,
544
+ "eval_f1": 0.8183098591549295,
545
+ "eval_loss": 0.04099448770284653,
546
+ "eval_precision": 0.7809139784946236,
547
+ "eval_recall": 0.8594674556213018,
548
+ "eval_runtime": 3.0102,
549
+ "eval_samples_per_second": 113.281,
550
+ "eval_steps_per_second": 3.654,
551
+ "step": 650
552
+ },
553
+ {
554
+ "epoch": 6.808290155440415,
555
+ "grad_norm": 0.113290935754776,
556
+ "learning_rate": 3.168436514825052e-05,
557
+ "loss": 0.003636385935048262,
558
+ "step": 660
559
+ },
560
+ {
561
+ "epoch": 6.932642487046632,
562
+ "grad_norm": 0.4593416452407837,
563
+ "learning_rate": 3.101312390915634e-05,
564
+ "loss": 0.006359308337171872,
565
+ "step": 672
566
+ },
567
+ {
568
+ "epoch": 7.051813471502591,
569
+ "grad_norm": 0.5328609943389893,
570
+ "learning_rate": 3.0337248894876746e-05,
571
+ "loss": 0.0031539862975478172,
572
+ "step": 684
573
+ },
574
+ {
575
+ "epoch": 7.176165803108808,
576
+ "grad_norm": 0.13898538053035736,
577
+ "learning_rate": 2.9657260941655417e-05,
578
+ "loss": 0.0014738718358178933,
579
+ "step": 696
580
+ },
581
+ {
582
+ "epoch": 7.217616580310881,
583
+ "eval_accuracy": 0.9636793466580701,
584
+ "eval_f1": 0.8498542274052479,
585
+ "eval_loss": 0.041410233825445175,
586
+ "eval_precision": 0.8376436781609196,
587
+ "eval_recall": 0.8624260355029586,
588
+ "eval_runtime": 3.0173,
589
+ "eval_samples_per_second": 113.015,
590
+ "eval_steps_per_second": 3.646,
591
+ "step": 700
592
+ },
593
+ {
594
+ "epoch": 7.300518134715026,
595
+ "grad_norm": 0.020028043538331985,
596
+ "learning_rate": 2.8973684055209106e-05,
597
+ "loss": 0.0009494482850035032,
598
+ "step": 708
599
+ },
600
+ {
601
+ "epoch": 7.424870466321243,
602
+ "grad_norm": 0.038508884608745575,
603
+ "learning_rate": 2.8287045006923473e-05,
604
+ "loss": 0.002579918752113978,
605
+ "step": 720
606
+ },
607
+ {
608
+ "epoch": 7.549222797927461,
609
+ "grad_norm": 0.46958020329475403,
610
+ "learning_rate": 2.7597872927917613e-05,
611
+ "loss": 0.0010905038410176833,
612
+ "step": 732
613
+ },
614
+ {
615
+ "epoch": 7.6735751295336785,
616
+ "grad_norm": 0.1547262817621231,
617
+ "learning_rate": 2.6906698901290167e-05,
618
+ "loss": 0.0028232391923666,
619
+ "step": 744
620
+ },
621
+ {
622
+ "epoch": 7.7357512953367875,
623
+ "eval_accuracy": 0.957016978293574,
624
+ "eval_f1": 0.829302987197724,
625
+ "eval_loss": 0.04088299721479416,
626
+ "eval_precision": 0.7986301369863014,
627
+ "eval_recall": 0.8624260355029586,
628
+ "eval_runtime": 3.0234,
629
+ "eval_samples_per_second": 112.787,
630
+ "eval_steps_per_second": 3.638,
631
+ "step": 750
632
+ },
633
+ {
634
+ "epoch": 7.7979274611398965,
635
+ "grad_norm": 0.0519477054476738,
636
+ "learning_rate": 2.621405555286121e-05,
637
+ "loss": 0.0010418800326685111,
638
+ "step": 756
639
+ },
640
+ {
641
+ "epoch": 7.922279792746114,
642
+ "grad_norm": 1.0287493467330933,
643
+ "learning_rate": 2.5520476640725342e-05,
644
+ "loss": 0.001736572477966547,
645
+ "step": 768
646
+ },
647
+ {
648
+ "epoch": 8.041450777202073,
649
+ "grad_norm": 0.013289994560182095,
650
+ "learning_rate": 2.4826496643932183e-05,
651
+ "loss": 0.000329588889144361,
652
+ "step": 780
653
+ },
654
+ {
655
+ "epoch": 8.16580310880829,
656
+ "grad_norm": 0.011798656545579433,
657
+ "learning_rate": 2.413265035061131e-05,
658
+ "loss": 0.002016165293753147,
659
+ "step": 792
660
+ },
661
+ {
662
+ "epoch": 8.248704663212436,
663
+ "eval_accuracy": 0.9630346013324736,
664
+ "eval_f1": 0.8407079646017699,
665
+ "eval_loss": 0.05126392841339111,
666
+ "eval_precision": 0.8382352941176471,
667
+ "eval_recall": 0.8431952662721893,
668
+ "eval_runtime": 3.0399,
669
+ "eval_samples_per_second": 112.176,
670
+ "eval_steps_per_second": 3.619,
671
+ "step": 800
672
+ },
673
+ {
674
+ "epoch": 8.290155440414507,
675
+ "grad_norm": 0.016478369012475014,
676
+ "learning_rate": 2.343947244585902e-05,
677
+ "loss": 0.001583070649454991,
678
+ "step": 804
679
+ },
680
+ {
681
+ "epoch": 8.414507772020725,
682
+ "grad_norm": 0.08483371138572693,
683
+ "learning_rate": 2.2747497099704533e-05,
684
+ "loss": 0.0006651872924218575,
685
+ "step": 816
686
+ },
687
+ {
688
+ "epoch": 8.538860103626943,
689
+ "grad_norm": 0.02956315688788891,
690
+ "learning_rate": 2.2057257555472998e-05,
691
+ "loss": 0.001361117387811343,
692
+ "step": 828
693
+ },
694
+ {
695
+ "epoch": 8.663212435233161,
696
+ "grad_norm": 0.1703813225030899,
697
+ "learning_rate": 2.136928571886275e-05,
698
+ "loss": 0.0004198816216861208,
699
+ "step": 840
700
+ },
701
+ {
702
+ "epoch": 8.766839378238341,
703
+ "eval_accuracy": 0.9636793466580701,
704
+ "eval_f1": 0.851528384279476,
705
+ "eval_loss": 0.04630826786160469,
706
+ "eval_precision": 0.83810888252149,
707
+ "eval_recall": 0.8653846153846154,
708
+ "eval_runtime": 3.0066,
709
+ "eval_samples_per_second": 113.418,
710
+ "eval_steps_per_second": 3.659,
711
+ "step": 850
712
+ },
713
+ {
714
+ "epoch": 8.787564766839377,
715
+ "grad_norm": 0.08187761902809143,
716
+ "learning_rate": 2.068411174805327e-05,
717
+ "loss": 0.00021340335176015893,
718
+ "step": 852
719
+ },
720
+ {
721
+ "epoch": 8.911917098445596,
722
+ "grad_norm": 0.1566244512796402,
723
+ "learning_rate": 2.0002263645159892e-05,
724
+ "loss": 0.0004732525364185373,
725
+ "step": 864
726
+ },
727
+ {
728
+ "epoch": 9.031088082901555,
729
+ "grad_norm": 0.007061833515763283,
730
+ "learning_rate": 1.9324266849349848e-05,
731
+ "loss": 9.06402322774132e-05,
732
+ "step": 876
733
+ },
734
+ {
735
+ "epoch": 9.155440414507773,
736
+ "grad_norm": 0.010846185497939587,
737
+ "learning_rate": 1.8650643831933502e-05,
738
+ "loss": 0.00011390061505759756,
739
+ "step": 888
740
+ },
741
+ {
742
+ "epoch": 9.279792746113989,
743
+ "grad_norm": 0.03851994872093201,
744
+ "learning_rate": 1.798191369374258e-05,
745
+ "loss": 0.00019203778356313705,
746
+ "step": 900
747
+ },
748
+ {
749
+ "epoch": 9.279792746113989,
750
+ "eval_accuracy": 0.9649688373092629,
751
+ "eval_f1": 0.8546255506607928,
752
+ "eval_loss": 0.05020074546337128,
753
+ "eval_precision": 0.8483965014577259,
754
+ "eval_recall": 0.8609467455621301,
755
+ "eval_runtime": 3.0339,
756
+ "eval_samples_per_second": 112.397,
757
+ "eval_steps_per_second": 3.626,
758
+ "step": 900
759
+ },
760
+ {
761
+ "epoch": 9.404145077720207,
762
+ "grad_norm": 0.006672121584415436,
763
+ "learning_rate": 1.731859176510569e-05,
764
+ "loss": 9.344543407981594e-05,
765
+ "step": 912
766
+ },
767
+ {
768
+ "epoch": 9.528497409326425,
769
+ "grad_norm": 0.028979798778891563,
770
+ "learning_rate": 1.666118920872949e-05,
771
+ "loss": 0.0009806355616698663,
772
+ "step": 924
773
+ },
774
+ {
775
+ "epoch": 9.652849740932643,
776
+ "grad_norm": 0.010359357111155987,
777
+ "learning_rate": 1.6010212625791422e-05,
778
+ "loss": 0.00018761774602656564,
779
+ "step": 936
780
+ },
781
+ {
782
+ "epoch": 9.77720207253886,
783
+ "grad_norm": 0.0016930641140788794,
784
+ "learning_rate": 1.536616366554768e-05,
785
+ "loss": 0.00010497189941816032,
786
+ "step": 948
787
+ },
788
+ {
789
+ "epoch": 9.797927461139896,
790
+ "eval_accuracy": 0.9641091768751343,
791
+ "eval_f1": 0.849112426035503,
792
+ "eval_loss": 0.05002701282501221,
793
+ "eval_precision": 0.849112426035503,
794
+ "eval_recall": 0.849112426035503,
795
+ "eval_runtime": 3.0151,
796
+ "eval_samples_per_second": 113.098,
797
+ "eval_steps_per_second": 3.648,
798
+ "step": 950
799
+ },
800
+ {
801
+ "epoch": 9.901554404145077,
802
+ "grad_norm": 0.0343826562166214,
803
+ "learning_rate": 1.472953863875706e-05,
804
+ "loss": 8.524257767324646e-05,
805
+ "step": 960
806
+ },
807
+ {
808
+ "epoch": 10.020725388601036,
809
+ "grad_norm": 0.006111979950219393,
810
+ "learning_rate": 1.4100828135218829e-05,
811
+ "loss": 0.00012398386024869978,
812
+ "step": 972
813
+ },
814
+ {
815
+ "epoch": 10.145077720207254,
816
+ "grad_norm": 0.010606050491333008,
817
+ "learning_rate": 1.3480516645719172e-05,
818
+ "loss": 3.8064138304131724e-05,
819
+ "step": 984
820
+ },
821
+ {
822
+ "epoch": 10.26943005181347,
823
+ "grad_norm": 0.0009117216104641557,
824
+ "learning_rate": 1.2869082188677572e-05,
825
+ "loss": 0.0001223099728425344,
826
+ "step": 996
827
+ },
828
+ {
829
+ "epoch": 10.310880829015543,
830
+ "eval_accuracy": 0.9645390070921985,
831
+ "eval_f1": 0.8514412416851442,
832
+ "eval_loss": 0.05051806941628456,
833
+ "eval_precision": 0.8508124076809453,
834
+ "eval_recall": 0.8520710059171598,
835
+ "eval_runtime": 3.0202,
836
+ "eval_samples_per_second": 112.906,
837
+ "eval_steps_per_second": 3.642,
838
+ "step": 1000
839
+ },
840
+ {
841
+ "epoch": 10.393782383419689,
842
+ "grad_norm": 0.0027385777793824673,
843
+ "learning_rate": 1.2266995941780934e-05,
844
+ "loss": 7.653955738836278e-05,
845
+ "step": 1008
846
+ },
847
+ {
848
+ "epoch": 10.518134715025907,
849
+ "grad_norm": 0.00011921550321858376,
850
+ "learning_rate": 1.1674721878889171e-05,
851
+ "loss": 0.0001742592042622467,
852
+ "step": 1020
853
+ },
854
+ {
855
+ "epoch": 10.642487046632125,
856
+ "grad_norm": 0.001234560739248991,
857
+ "learning_rate": 1.1092716412492226e-05,
858
+ "loss": 7.672411447856575e-05,
859
+ "step": 1032
860
+ },
861
+ {
862
+ "epoch": 10.766839378238341,
863
+ "grad_norm": 0.0005847270367667079,
864
+ "learning_rate": 1.0521428041993848e-05,
865
+ "loss": 9.932015624751026e-06,
866
+ "step": 1044
867
+ },
868
+ {
869
+ "epoch": 10.82901554404145,
870
+ "eval_accuracy": 0.9643240919836664,
871
+ "eval_f1": 0.8508124076809453,
872
+ "eval_loss": 0.05043612793087959,
873
+ "eval_precision": 0.8495575221238938,
874
+ "eval_recall": 0.8520710059171598,
875
+ "eval_runtime": 3.024,
876
+ "eval_samples_per_second": 112.764,
877
+ "eval_steps_per_second": 3.638,
878
+ "step": 1050
879
+ },
880
+ {
881
+ "epoch": 10.89119170984456,
882
+ "grad_norm": 0.0753927081823349,
883
+ "learning_rate": 9.961297008093357e-06,
884
+ "loss": 0.00011471325221161048,
885
+ "step": 1056
886
+ },
887
+ {
888
+ "epoch": 11.010362694300518,
889
+ "grad_norm": 0.003958659712225199,
890
+ "learning_rate": 9.412754953531663e-06,
891
+ "loss": 1.709427791259562e-05,
892
+ "step": 1068
893
+ },
894
+ {
895
+ "epoch": 11.134715025906736,
896
+ "grad_norm": 0.001931481878273189,
897
+ "learning_rate": 8.876224590462873e-06,
898
+ "loss": 1.0689032933441922e-05,
899
+ "step": 1080
900
+ },
901
+ {
902
+ "epoch": 11.259067357512953,
903
+ "grad_norm": 0.012324509210884571,
904
+ "learning_rate": 8.352119374707978e-06,
905
+ "loss": 5.9023882689264916e-05,
906
+ "step": 1092
907
+ },
908
+ {
909
+ "epoch": 11.341968911917098,
910
+ "eval_accuracy": 0.9643240919836664,
911
+ "eval_f1": 0.8510324483775811,
912
+ "eval_loss": 0.050491511821746826,
913
+ "eval_precision": 0.8485294117647059,
914
+ "eval_recall": 0.8535502958579881,
915
+ "eval_runtime": 3.0303,
916
+ "eval_samples_per_second": 112.529,
917
+ "eval_steps_per_second": 3.63,
918
+ "step": 1100
919
+ },
920
+ {
921
+ "epoch": 11.38341968911917,
922
+ "grad_norm": 0.002621831139549613,
923
+ "learning_rate": 7.840843187141544e-06,
924
+ "loss": 5.975224000091354e-05,
925
+ "step": 1104
926
+ },
927
+ {
928
+ "epoch": 11.507772020725389,
929
+ "grad_norm": 0.0006910113152116537,
930
+ "learning_rate": 7.3427900224569606e-06,
931
+ "loss": 7.24908944296961e-05,
932
+ "step": 1116
933
+ },
934
+ {
935
+ "epoch": 11.632124352331607,
936
+ "grad_norm": 0.01195142138749361,
937
+ "learning_rate": 6.858343685550048e-06,
938
+ "loss": 3.2996586620962866e-05,
939
+ "step": 1128
940
+ },
941
+ {
942
+ "epoch": 11.756476683937823,
943
+ "grad_norm": 0.00024170154938474298,
944
+ "learning_rate": 6.387877495755065e-06,
945
+ "loss": 1.4914168180742612e-05,
946
+ "step": 1140
947
+ },
948
+ {
949
+ "epoch": 11.860103626943005,
950
+ "eval_accuracy": 0.9645390070921985,
951
+ "eval_f1": 0.8514412416851442,
952
+ "eval_loss": 0.05080362409353256,
953
+ "eval_precision": 0.8508124076809453,
954
+ "eval_recall": 0.8520710059171598,
955
+ "eval_runtime": 3.0169,
956
+ "eval_samples_per_second": 113.03,
957
+ "eval_steps_per_second": 3.646,
958
+ "step": 1150
959
+ },
960
+ {
961
+ "epoch": 11.880829015544041,
962
+ "grad_norm": 0.00013972603483125567,
963
+ "learning_rate": 5.93175399916101e-06,
964
+ "loss": 0.00017348669158915678,
965
+ "step": 1152
966
+ },
967
+ {
968
+ "epoch": 12.0,
969
+ "grad_norm": 0.014590195380151272,
970
+ "learning_rate": 5.490324689229834e-06,
971
+ "loss": 6.250141692968707e-05,
972
+ "step": 1164
973
+ },
974
+ {
975
+ "epoch": 12.124352331606218,
976
+ "grad_norm": 0.002015709411352873,
977
+ "learning_rate": 5.063929735931985e-06,
978
+ "loss": 1.4617435226682574e-05,
979
+ "step": 1176
980
+ },
981
+ {
982
+ "epoch": 12.248704663212436,
983
+ "grad_norm": 0.0005398832727223635,
984
+ "learning_rate": 4.652897723607913e-06,
985
+ "loss": 1.6918191249715164e-05,
986
+ "step": 1188
987
+ },
988
+ {
989
+ "epoch": 12.373056994818652,
990
+ "grad_norm": 0.0006741413380950689,
991
+ "learning_rate": 4.257545397757623e-06,
992
+ "loss": 4.216020170133561e-05,
993
+ "step": 1200
994
+ },
995
+ {
996
+ "epoch": 12.373056994818652,
997
+ "eval_accuracy": 0.9645390070921985,
998
+ "eval_f1": 0.8514412416851442,
999
+ "eval_loss": 0.05097455903887749,
1000
+ "eval_precision": 0.8508124076809453,
1001
+ "eval_recall": 0.8520710059171598,
1002
+ "eval_runtime": 3.0165,
1003
+ "eval_samples_per_second": 113.043,
1004
+ "eval_steps_per_second": 3.647,
1005
+ "step": 1200
1006
+ },
1007
+ {
1008
+ "epoch": 12.49740932642487,
1009
+ "grad_norm": 0.0007238159305416048,
1010
+ "learning_rate": 3.878177420953291e-06,
1011
+ "loss": 5.365341106274476e-05,
1012
+ "step": 1212
1013
+ },
1014
+ {
1015
+ "epoch": 12.621761658031089,
1016
+ "grad_norm": 0.002653967123478651,
1017
+ "learning_rate": 3.5150861380631755e-06,
1018
+ "loss": 3.87875431139643e-05,
1019
+ "step": 1224
1020
+ },
1021
+ {
1022
+ "epoch": 12.746113989637305,
1023
+ "grad_norm": 0.0012897817650809884,
1024
+ "learning_rate": 3.1685513509676606e-06,
1025
+ "loss": 1.6210615285672247e-05,
1026
+ "step": 1236
1027
+ },
1028
+ {
1029
+ "epoch": 12.870466321243523,
1030
+ "grad_norm": 0.010877900756895542,
1031
+ "learning_rate": 2.838840102941015e-06,
1032
+ "loss": 4.428145863736669e-05,
1033
+ "step": 1248
1034
+ },
1035
+ {
1036
+ "epoch": 12.89119170984456,
1037
+ "eval_accuracy": 0.9645390070921985,
1038
+ "eval_f1": 0.8514412416851442,
1039
+ "eval_loss": 0.05106152221560478,
1040
+ "eval_precision": 0.8508124076809453,
1041
+ "eval_recall": 0.8520710059171598,
1042
+ "eval_runtime": 3.0194,
1043
+ "eval_samples_per_second": 112.937,
1044
+ "eval_steps_per_second": 3.643,
1045
+ "step": 1250
1046
+ },
1047
+ {
1048
+ "epoch": 12.994818652849741,
1049
+ "grad_norm": 0.004842772614210844,
1050
+ "learning_rate": 2.52620647286512e-06,
1051
+ "loss": 4.890701772334675e-05,
1052
+ "step": 1260
1053
+ },
1054
+ {
1055
+ "epoch": 13.1139896373057,
1056
+ "grad_norm": 0.005717657506465912,
1057
+ "learning_rate": 2.2308913794336583e-06,
1058
+ "loss": 8.704484207555652e-05,
1059
+ "step": 1272
1060
+ },
1061
+ {
1062
+ "epoch": 13.238341968911918,
1063
+ "grad_norm": 0.0031942392233759165,
1064
+ "learning_rate": 1.953122395497711e-06,
1065
+ "loss": 3.180075388324136e-05,
1066
+ "step": 1284
1067
+ },
1068
+ {
1069
+ "epoch": 13.362694300518134,
1070
+ "grad_norm": 0.006513912230730057,
1071
+ "learning_rate": 1.6931135726957315e-06,
1072
+ "loss": 1.0696584164785842e-05,
1073
+ "step": 1296
1074
+ },
1075
+ {
1076
+ "epoch": 13.404145077720207,
1077
+ "eval_accuracy": 0.9645390070921985,
1078
+ "eval_f1": 0.8514412416851442,
1079
+ "eval_loss": 0.05115008354187012,
1080
+ "eval_precision": 0.8508124076809453,
1081
+ "eval_recall": 0.8520710059171598,
1082
+ "eval_runtime": 3.0366,
1083
+ "eval_samples_per_second": 112.298,
1084
+ "eval_steps_per_second": 3.623,
1085
+ "step": 1300
1086
+ },
1087
+ {
1088
+ "epoch": 13.487046632124352,
1089
+ "grad_norm": 0.03252461925148964,
1090
+ "learning_rate": 1.4510652765031829e-06,
1091
+ "loss": 4.1350130535041295e-05,
1092
+ "step": 1308
1093
+ },
1094
+ {
1095
+ "epoch": 13.61139896373057,
1096
+ "grad_norm": 0.0007789316587150097,
1097
+ "learning_rate": 1.2271640318288397e-06,
1098
+ "loss": 9.945468870379651e-06,
1099
+ "step": 1320
1100
+ },
1101
+ {
1102
+ "epoch": 13.735751295336787,
1103
+ "grad_norm": 0.005432860925793648,
1104
+ "learning_rate": 1.0215823792767615e-06,
1105
+ "loss": 2.4644163204357028e-05,
1106
+ "step": 1332
1107
+ },
1108
+ {
1109
+ "epoch": 13.860103626943005,
1110
+ "grad_norm": 0.07517717033624649,
1111
+ "learning_rate": 8.344787421847217e-07,
1112
+ "loss": 0.00010969570333448549,
1113
+ "step": 1344
1114
+ },
1115
+ {
1116
+ "epoch": 13.922279792746114,
1117
+ "eval_accuracy": 0.9645390070921985,
1118
+ "eval_f1": 0.8514412416851442,
1119
+ "eval_loss": 0.05114062502980232,
1120
+ "eval_precision": 0.8508124076809453,
1121
+ "eval_recall": 0.8520710059171598,
1122
+ "eval_runtime": 3.0311,
1123
+ "eval_samples_per_second": 112.5,
1124
+ "eval_steps_per_second": 3.629,
1125
+ "step": 1350
1126
+ }
1127
+ ],
1128
+ "logging_steps": 12,
1129
+ "max_steps": 1455,
1130
+ "num_input_tokens_seen": 0,
1131
+ "num_train_epochs": 15,
1132
+ "save_steps": 50,
1133
+ "stateful_callbacks": {
1134
+ "EarlyStoppingCallback": {
1135
+ "args": {
1136
+ "early_stopping_patience": 10,
1137
+ "early_stopping_threshold": 0.0
1138
+ },
1139
+ "attributes": {
1140
+ "early_stopping_patience_counter": 9
1141
+ }
1142
+ },
1143
+ "TrainerControl": {
1144
+ "args": {
1145
+ "should_epoch_stop": false,
1146
+ "should_evaluate": false,
1147
+ "should_log": false,
1148
+ "should_save": true,
1149
+ "should_training_stop": false
1150
+ },
1151
+ "attributes": {}
1152
+ }
1153
+ },
1154
+ "total_flos": 2.271280454602291e+16,
1155
+ "train_batch_size": 8,
1156
+ "trial_name": null,
1157
+ "trial_params": null
1158
+ }
checkpoint-1350/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:32e421217e1eea826c4025f111c9f4fa0b9d53a8bca71632cae9ad6aed6b26f4
3
+ size 5201
checkpoint-1400/config.json ADDED
@@ -0,0 +1,99 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "architectures": [
3
+ "ModernBertForTokenClassification"
4
+ ],
5
+ "attention_bias": false,
6
+ "attention_dropout": 0.0,
7
+ "bos_token_id": 50281,
8
+ "classifier_activation": "gelu",
9
+ "classifier_bias": false,
10
+ "classifier_dropout": 0.3,
11
+ "classifier_pooling": "mean",
12
+ "cls_token_id": 50281,
13
+ "decoder_bias": true,
14
+ "deterministic_flash_attn": false,
15
+ "dtype": "float32",
16
+ "embedding_dropout": 0.0,
17
+ "eos_token_id": 50282,
18
+ "global_attn_every_n_layers": 3,
19
+ "gradient_checkpointing": false,
20
+ "hidden_activation": "gelu",
21
+ "hidden_size": 1024,
22
+ "id2label": {
23
+ "0": "O",
24
+ "1": "FP",
25
+ "2": "RP",
26
+ "3": "RV",
27
+ "4": "PW"
28
+ },
29
+ "initializer_cutoff_factor": 2.0,
30
+ "initializer_range": 0.02,
31
+ "intermediate_size": 2624,
32
+ "label2id": {
33
+ "FP": 1,
34
+ "O": 0,
35
+ "PW": 4,
36
+ "RP": 2,
37
+ "RV": 3
38
+ },
39
+ "layer_norm_eps": 1e-05,
40
+ "layer_types": [
41
+ "full_attention",
42
+ "sliding_attention",
43
+ "sliding_attention",
44
+ "full_attention",
45
+ "sliding_attention",
46
+ "sliding_attention",
47
+ "full_attention",
48
+ "sliding_attention",
49
+ "sliding_attention",
50
+ "full_attention",
51
+ "sliding_attention",
52
+ "sliding_attention",
53
+ "full_attention",
54
+ "sliding_attention",
55
+ "sliding_attention",
56
+ "full_attention",
57
+ "sliding_attention",
58
+ "sliding_attention",
59
+ "full_attention",
60
+ "sliding_attention",
61
+ "sliding_attention",
62
+ "full_attention",
63
+ "sliding_attention",
64
+ "sliding_attention",
65
+ "full_attention",
66
+ "sliding_attention",
67
+ "sliding_attention",
68
+ "full_attention"
69
+ ],
70
+ "local_attention": 128,
71
+ "max_position_embeddings": 8192,
72
+ "mlp_bias": false,
73
+ "mlp_dropout": 0.0,
74
+ "model_type": "modernbert",
75
+ "norm_bias": false,
76
+ "norm_eps": 1e-05,
77
+ "num_attention_heads": 16,
78
+ "num_hidden_layers": 28,
79
+ "pad_token_id": 50283,
80
+ "position_embedding_type": "absolute",
81
+ "repad_logits_with_grad": false,
82
+ "rope_parameters": {
83
+ "full_attention": {
84
+ "rope_theta": 160000.0,
85
+ "rope_type": "default"
86
+ },
87
+ "sliding_attention": {
88
+ "rope_theta": 10000.0,
89
+ "rope_type": "default"
90
+ }
91
+ },
92
+ "sep_token_id": 50282,
93
+ "sparse_pred_ignore_index": -100,
94
+ "sparse_prediction": false,
95
+ "tie_word_embeddings": true,
96
+ "transformers_version": "5.0.0",
97
+ "use_cache": false,
98
+ "vocab_size": 50368
99
+ }
checkpoint-1400/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7764237d1bdf8e9b32a79878020489bf00ac5f191a4980182317b52cd5236139
3
+ size 1583363940
checkpoint-1400/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9ee48109b855f4e60fa69febff63b78030e762af709f51b1a73d2b7e3f6f1722
3
+ size 1112419851
checkpoint-1400/rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9be962792ea7a3e9a9d059d1b9d81e1056932d080b9edf7387372512f7206890
3
+ size 14709
checkpoint-1400/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:983a4b44c32aa6763eaef66a4c7d7a7692b9f2967b3e887f3b6cdab4d01e2537
3
+ size 1465
checkpoint-1400/tokenizer.json ADDED
The diff for this file is too large to render. See raw diff
 
checkpoint-1400/tokenizer_config.json ADDED
@@ -0,0 +1,16 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "backend": "tokenizers",
3
+ "clean_up_tokenization_spaces": true,
4
+ "cls_token": "[CLS]",
5
+ "is_local": false,
6
+ "mask_token": "[MASK]",
7
+ "model_input_names": [
8
+ "input_ids",
9
+ "attention_mask"
10
+ ],
11
+ "model_max_length": 8192,
12
+ "pad_token": "[PAD]",
13
+ "sep_token": "[SEP]",
14
+ "tokenizer_class": "TokenizersBackend",
15
+ "unk_token": "[UNK]"
16
+ }
checkpoint-1400/trainer_state.json ADDED
@@ -0,0 +1,1198 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_global_step": 900,
3
+ "best_metric": 0.8546255506607928,
4
+ "best_model_checkpoint": "/content/drive/MyDrive/expE_large_oversample_rv/checkpoint-900",
5
+ "epoch": 14.435233160621761,
6
+ "eval_steps": 50,
7
+ "global_step": 1400,
8
+ "is_hyper_param_search": false,
9
+ "is_local_process_zero": true,
10
+ "is_world_process_zero": true,
11
+ "log_history": [
12
+ {
13
+ "epoch": 0.010362694300518135,
14
+ "grad_norm": 19.391862869262695,
15
+ "learning_rate": 0.0,
16
+ "loss": 0.8294010162353516,
17
+ "step": 1
18
+ },
19
+ {
20
+ "epoch": 0.12435233160621761,
21
+ "grad_norm": 7.022524833679199,
22
+ "learning_rate": 5.670103092783505e-06,
23
+ "loss": 0.6434053074229847,
24
+ "step": 12
25
+ },
26
+ {
27
+ "epoch": 0.24870466321243523,
28
+ "grad_norm": 2.0728249549865723,
29
+ "learning_rate": 1.1855670103092783e-05,
30
+ "loss": 0.5182687441507975,
31
+ "step": 24
32
+ },
33
+ {
34
+ "epoch": 0.37305699481865284,
35
+ "grad_norm": 8.758841514587402,
36
+ "learning_rate": 1.8041237113402062e-05,
37
+ "loss": 0.39159242312113446,
38
+ "step": 36
39
+ },
40
+ {
41
+ "epoch": 0.49740932642487046,
42
+ "grad_norm": 10.565062522888184,
43
+ "learning_rate": 2.422680412371134e-05,
44
+ "loss": 0.31471407413482666,
45
+ "step": 48
46
+ },
47
+ {
48
+ "epoch": 0.5181347150259067,
49
+ "eval_accuracy": 0.7721899849559424,
50
+ "eval_f1": 0.4335260115606937,
51
+ "eval_loss": 0.05006009340286255,
52
+ "eval_precision": 0.32142857142857145,
53
+ "eval_recall": 0.665680473372781,
54
+ "eval_runtime": 5.6283,
55
+ "eval_samples_per_second": 60.586,
56
+ "eval_steps_per_second": 1.954,
57
+ "step": 50
58
+ },
59
+ {
60
+ "epoch": 0.6217616580310881,
61
+ "grad_norm": 5.744125843048096,
62
+ "learning_rate": 3.0412371134020617e-05,
63
+ "loss": 0.28232719500859577,
64
+ "step": 60
65
+ },
66
+ {
67
+ "epoch": 0.7461139896373057,
68
+ "grad_norm": 4.099153518676758,
69
+ "learning_rate": 3.6597938144329896e-05,
70
+ "loss": 0.22093822558720908,
71
+ "step": 72
72
+ },
73
+ {
74
+ "epoch": 0.8704663212435233,
75
+ "grad_norm": 12.179475784301758,
76
+ "learning_rate": 4.2783505154639175e-05,
77
+ "loss": 0.2065215309460958,
78
+ "step": 84
79
+ },
80
+ {
81
+ "epoch": 0.9948186528497409,
82
+ "grad_norm": 8.685556411743164,
83
+ "learning_rate": 4.8969072164948454e-05,
84
+ "loss": 0.16811863581339517,
85
+ "step": 96
86
+ },
87
+ {
88
+ "epoch": 1.0310880829015545,
89
+ "eval_accuracy": 0.7461852568235547,
90
+ "eval_f1": 0.4638047138047138,
91
+ "eval_loss": 0.03645005077123642,
92
+ "eval_precision": 0.3241176470588235,
93
+ "eval_recall": 0.8150887573964497,
94
+ "eval_runtime": 3.0466,
95
+ "eval_samples_per_second": 111.926,
96
+ "eval_steps_per_second": 3.611,
97
+ "step": 100
98
+ },
99
+ {
100
+ "epoch": 1.1139896373056994,
101
+ "grad_norm": 33.46553421020508,
102
+ "learning_rate": 4.999331054325924e-05,
103
+ "loss": 0.15307321151097616,
104
+ "step": 108
105
+ },
106
+ {
107
+ "epoch": 1.238341968911917,
108
+ "grad_norm": 5.198877811431885,
109
+ "learning_rate": 4.996762857385083e-05,
110
+ "loss": 0.1403733491897583,
111
+ "step": 120
112
+ },
113
+ {
114
+ "epoch": 1.3626943005181347,
115
+ "grad_norm": 1.8454481363296509,
116
+ "learning_rate": 4.9922706292838655e-05,
117
+ "loss": 0.13873055577278137,
118
+ "step": 132
119
+ },
120
+ {
121
+ "epoch": 1.4870466321243523,
122
+ "grad_norm": 2.0861477851867676,
123
+ "learning_rate": 4.9858578317794894e-05,
124
+ "loss": 0.11305248737335205,
125
+ "step": 144
126
+ },
127
+ {
128
+ "epoch": 1.549222797927461,
129
+ "eval_accuracy": 0.895766172361917,
130
+ "eval_f1": 0.6706515242080096,
131
+ "eval_loss": 0.029485274106264114,
132
+ "eval_precision": 0.5626880641925778,
133
+ "eval_recall": 0.8298816568047337,
134
+ "eval_runtime": 3.0006,
135
+ "eval_samples_per_second": 113.645,
136
+ "eval_steps_per_second": 3.666,
137
+ "step": 150
138
+ },
139
+ {
140
+ "epoch": 1.61139896373057,
141
+ "grad_norm": 6.362898349761963,
142
+ "learning_rate": 4.9775294066397274e-05,
143
+ "loss": 0.11805283029874165,
144
+ "step": 156
145
+ },
146
+ {
147
+ "epoch": 1.7357512953367875,
148
+ "grad_norm": 2.3624749183654785,
149
+ "learning_rate": 4.967291771834727e-05,
150
+ "loss": 0.11144719521204631,
151
+ "step": 168
152
+ },
153
+ {
154
+ "epoch": 1.8601036269430051,
155
+ "grad_norm": 0.9627630710601807,
156
+ "learning_rate": 4.955152816591259e-05,
157
+ "loss": 0.0819690078496933,
158
+ "step": 180
159
+ },
160
+ {
161
+ "epoch": 1.9844559585492227,
162
+ "grad_norm": 15.513121604919434,
163
+ "learning_rate": 4.9411218953131964e-05,
164
+ "loss": 0.10844972729682922,
165
+ "step": 192
166
+ },
167
+ {
168
+ "epoch": 2.062176165803109,
169
+ "eval_accuracy": 0.8074360627552117,
170
+ "eval_f1": 0.5410536307546274,
171
+ "eval_loss": 0.030047627165913582,
172
+ "eval_precision": 0.39832285115303984,
173
+ "eval_recall": 0.8431952662721893,
174
+ "eval_runtime": 3.037,
175
+ "eval_samples_per_second": 112.283,
176
+ "eval_steps_per_second": 3.622,
177
+ "step": 200
178
+ },
179
+ {
180
+ "epoch": 2.1036269430051813,
181
+ "grad_norm": 2.684053421020508,
182
+ "learning_rate": 4.9252098203729154e-05,
183
+ "loss": 0.08770081400871277,
184
+ "step": 204
185
+ },
186
+ {
187
+ "epoch": 2.227979274611399,
188
+ "grad_norm": 21.072328567504883,
189
+ "learning_rate": 4.907428853779167e-05,
190
+ "loss": 0.1440110206604004,
191
+ "step": 216
192
+ },
193
+ {
194
+ "epoch": 2.3523316062176165,
195
+ "grad_norm": 3.221358299255371,
196
+ "learning_rate": 4.887792697727854e-05,
197
+ "loss": 0.11832749843597412,
198
+ "step": 228
199
+ },
200
+ {
201
+ "epoch": 2.476683937823834,
202
+ "grad_norm": 5.4783101081848145,
203
+ "learning_rate": 4.866316484042968e-05,
204
+ "loss": 0.0703568160533905,
205
+ "step": 240
206
+ },
207
+ {
208
+ "epoch": 2.5803108808290154,
209
+ "eval_accuracy": 0.9129593810444874,
210
+ "eval_f1": 0.7118012422360248,
211
+ "eval_loss": 0.0311787910759449,
212
+ "eval_precision": 0.6134903640256959,
213
+ "eval_recall": 0.8476331360946746,
214
+ "eval_runtime": 3.0166,
215
+ "eval_samples_per_second": 113.041,
216
+ "eval_steps_per_second": 3.646,
217
+ "step": 250
218
+ },
219
+ {
220
+ "epoch": 2.6010362694300517,
221
+ "grad_norm": 6.384552478790283,
222
+ "learning_rate": 4.8430167625158595e-05,
223
+ "loss": 0.09287564953168233,
224
+ "step": 252
225
+ },
226
+ {
227
+ "epoch": 2.7253886010362693,
228
+ "grad_norm": 4.217259407043457,
229
+ "learning_rate": 4.817911488151791e-05,
230
+ "loss": 0.08391642570495605,
231
+ "step": 264
232
+ },
233
+ {
234
+ "epoch": 2.849740932642487,
235
+ "grad_norm": 3.1646366119384766,
236
+ "learning_rate": 4.791020007333629e-05,
237
+ "loss": 0.061600317557652794,
238
+ "step": 276
239
+ },
240
+ {
241
+ "epoch": 2.9740932642487046,
242
+ "grad_norm": 6.243884563446045,
243
+ "learning_rate": 4.762363042913322e-05,
244
+ "loss": 0.05470449725786845,
245
+ "step": 288
246
+ },
247
+ {
248
+ "epoch": 3.093264248704663,
249
+ "grad_norm": 2.5704944133758545,
250
+ "learning_rate": 4.7319626782426564e-05,
251
+ "loss": 0.05606028437614441,
252
+ "step": 300
253
+ },
254
+ {
255
+ "epoch": 3.093264248704663,
256
+ "eval_accuracy": 0.7977648828712659,
257
+ "eval_f1": 0.5429740791268759,
258
+ "eval_loss": 0.03227110207080841,
259
+ "eval_precision": 0.39198949441891007,
260
+ "eval_recall": 0.8831360946745562,
261
+ "eval_runtime": 3.0347,
262
+ "eval_samples_per_second": 112.367,
263
+ "eval_steps_per_second": 3.625,
264
+ "step": 300
265
+ },
266
+ {
267
+ "epoch": 3.2176165803108807,
268
+ "grad_norm": 2.420072317123413,
269
+ "learning_rate": 4.699842340155598e-05,
270
+ "loss": 0.053138395150502525,
271
+ "step": 312
272
+ },
273
+ {
274
+ "epoch": 3.3419689119170983,
275
+ "grad_norm": 67.39490509033203,
276
+ "learning_rate": 4.666026780915334e-05,
277
+ "loss": 0.10085086027781169,
278
+ "step": 324
279
+ },
280
+ {
281
+ "epoch": 3.466321243523316,
282
+ "grad_norm": 0.7816835045814514,
283
+ "learning_rate": 4.630542059139924e-05,
284
+ "loss": 0.0743176539738973,
285
+ "step": 336
286
+ },
287
+ {
288
+ "epoch": 3.5906735751295336,
289
+ "grad_norm": 0.7841729521751404,
290
+ "learning_rate": 4.593415519721257e-05,
291
+ "loss": 0.05843483408292135,
292
+ "step": 348
293
+ },
294
+ {
295
+ "epoch": 3.61139896373057,
296
+ "eval_accuracy": 0.9329464861379755,
297
+ "eval_f1": 0.7535070140280562,
298
+ "eval_loss": 0.030000876635313034,
299
+ "eval_precision": 0.6869671132764921,
300
+ "eval_recall": 0.834319526627219,
301
+ "eval_runtime": 3.0255,
302
+ "eval_samples_per_second": 112.709,
303
+ "eval_steps_per_second": 3.636,
304
+ "step": 350
305
+ },
306
+ {
307
+ "epoch": 3.715025906735751,
308
+ "grad_norm": 5.054598331451416,
309
+ "learning_rate": 4.5546757727528035e-05,
310
+ "loss": 0.03325176239013672,
311
+ "step": 360
312
+ },
313
+ {
314
+ "epoch": 3.839378238341969,
315
+ "grad_norm": 0.5077673196792603,
316
+ "learning_rate": 4.5143526714823766e-05,
317
+ "loss": 0.036297169824441276,
318
+ "step": 372
319
+ },
320
+ {
321
+ "epoch": 3.9637305699481864,
322
+ "grad_norm": 0.8186505436897278,
323
+ "learning_rate": 4.472477289306913e-05,
324
+ "loss": 0.026280519862969715,
325
+ "step": 384
326
+ },
327
+ {
328
+ "epoch": 4.082901554404145,
329
+ "grad_norm": 0.9546571373939514,
330
+ "learning_rate": 4.429081895826997e-05,
331
+ "loss": 0.038141210873921715,
332
+ "step": 396
333
+ },
334
+ {
335
+ "epoch": 4.124352331606218,
336
+ "eval_accuracy": 0.9116698903932946,
337
+ "eval_f1": 0.7129798903107862,
338
+ "eval_loss": 0.02928924560546875,
339
+ "eval_precision": 0.6062176165803109,
340
+ "eval_recall": 0.8653846153846154,
341
+ "eval_runtime": 3.0233,
342
+ "eval_samples_per_second": 112.79,
343
+ "eval_steps_per_second": 3.638,
344
+ "step": 400
345
+ },
346
+ {
347
+ "epoch": 4.2072538860103625,
348
+ "grad_norm": 0.38423895835876465,
349
+ "learning_rate": 4.38419993197957e-05,
350
+ "loss": 0.023756501575311024,
351
+ "step": 408
352
+ },
353
+ {
354
+ "epoch": 4.331606217616581,
355
+ "grad_norm": 0.261392205953598,
356
+ "learning_rate": 4.337865984268001e-05,
357
+ "loss": 0.02620025227467219,
358
+ "step": 420
359
+ },
360
+ {
361
+ "epoch": 4.455958549222798,
362
+ "grad_norm": 2.0392212867736816,
363
+ "learning_rate": 4.290115758109375e-05,
364
+ "loss": 0.02181059867143631,
365
+ "step": 432
366
+ },
367
+ {
368
+ "epoch": 4.580310880829016,
369
+ "grad_norm": 0.4685036242008209,
370
+ "learning_rate": 4.240986050319524e-05,
371
+ "loss": 0.023431827624638874,
372
+ "step": 444
373
+ },
374
+ {
375
+ "epoch": 4.642487046632124,
376
+ "eval_accuracy": 0.9593810444874274,
377
+ "eval_f1": 0.828297715549005,
378
+ "eval_loss": 0.04263457655906677,
379
+ "eval_precision": 0.8252569750367107,
380
+ "eval_recall": 0.8313609467455622,
381
+ "eval_runtime": 3.0402,
382
+ "eval_samples_per_second": 112.164,
383
+ "eval_steps_per_second": 3.618,
384
+ "step": 450
385
+ },
386
+ {
387
+ "epoch": 4.704663212435233,
388
+ "grad_norm": 0.535210371017456,
389
+ "learning_rate": 4.19051472075703e-05,
390
+ "loss": 0.021484971046447754,
391
+ "step": 456
392
+ },
393
+ {
394
+ "epoch": 4.829015544041451,
395
+ "grad_norm": 3.4881668090820312,
396
+ "learning_rate": 4.138740663148019e-05,
397
+ "loss": 0.03922188033660253,
398
+ "step": 468
399
+ },
400
+ {
401
+ "epoch": 4.953367875647668,
402
+ "grad_norm": 0.809377908706665,
403
+ "learning_rate": 4.085703775114266e-05,
404
+ "loss": 0.018220727642377216,
405
+ "step": 480
406
+ },
407
+ {
408
+ "epoch": 5.072538860103627,
409
+ "grad_norm": 3.266059160232544,
410
+ "learning_rate": 4.0314449274276663e-05,
411
+ "loss": 0.015343919396400452,
412
+ "step": 492
413
+ },
414
+ {
415
+ "epoch": 5.155440414507772,
416
+ "eval_accuracy": 0.9574468085106383,
417
+ "eval_f1": 0.8162672476397966,
418
+ "eval_loss": 0.04887477308511734,
419
+ "eval_precision": 0.8017118402282454,
420
+ "eval_recall": 0.8313609467455622,
421
+ "eval_runtime": 3.0737,
422
+ "eval_samples_per_second": 110.94,
423
+ "eval_steps_per_second": 3.579,
424
+ "step": 500
425
+ },
426
+ {
427
+ "epoch": 5.196891191709844,
428
+ "grad_norm": 0.14335286617279053,
429
+ "learning_rate": 3.976005932514807e-05,
430
+ "loss": 0.007969564447800318,
431
+ "step": 504
432
+ },
433
+ {
434
+ "epoch": 5.321243523316062,
435
+ "grad_norm": 0.24810202419757843,
436
+ "learning_rate": 3.919429512235874e-05,
437
+ "loss": 0.007458047941327095,
438
+ "step": 516
439
+ },
440
+ {
441
+ "epoch": 5.44559585492228,
442
+ "grad_norm": 1.7369506359100342,
443
+ "learning_rate": 3.86175926496275e-05,
444
+ "loss": 0.012967783957719803,
445
+ "step": 528
446
+ },
447
+ {
448
+ "epoch": 5.569948186528498,
449
+ "grad_norm": 0.40230634808540344,
450
+ "learning_rate": 3.8030396319816585e-05,
451
+ "loss": 0.008079041416446367,
452
+ "step": 540
453
+ },
454
+ {
455
+ "epoch": 5.6735751295336785,
456
+ "eval_accuracy": 0.9591661293788953,
457
+ "eval_f1": 0.8310168251645941,
458
+ "eval_loss": 0.04243873059749603,
459
+ "eval_precision": 0.8219971056439942,
460
+ "eval_recall": 0.8402366863905325,
461
+ "eval_runtime": 3.0267,
462
+ "eval_samples_per_second": 112.662,
463
+ "eval_steps_per_second": 3.634,
464
+ "step": 550
465
+ },
466
+ {
467
+ "epoch": 5.694300518134715,
468
+ "grad_norm": 0.46429362893104553,
469
+ "learning_rate": 3.743315863246256e-05,
470
+ "loss": 0.005399468044439952,
471
+ "step": 552
472
+ },
473
+ {
474
+ "epoch": 5.818652849740933,
475
+ "grad_norm": 1.60989248752594,
476
+ "learning_rate": 3.682633982507543e-05,
477
+ "loss": 0.006222298989693324,
478
+ "step": 564
479
+ },
480
+ {
481
+ "epoch": 5.94300518134715,
482
+ "grad_norm": 0.5653553009033203,
483
+ "learning_rate": 3.621040751847497e-05,
484
+ "loss": 0.015771896888812382,
485
+ "step": 576
486
+ },
487
+ {
488
+ "epoch": 6.062176165803109,
489
+ "grad_norm": 0.34870776534080505,
490
+ "learning_rate": 3.5585836356437264e-05,
491
+ "loss": 0.0066079869866371155,
492
+ "step": 588
493
+ },
494
+ {
495
+ "epoch": 6.186528497409326,
496
+ "grad_norm": 0.2842996120452881,
497
+ "learning_rate": 3.495310763992928e-05,
498
+ "loss": 0.0051950619866450625,
499
+ "step": 600
500
+ },
501
+ {
502
+ "epoch": 6.186528497409326,
503
+ "eval_accuracy": 0.9580915538362347,
504
+ "eval_f1": 0.8308351177730193,
505
+ "eval_loss": 0.04230675473809242,
506
+ "eval_precision": 0.8027586206896552,
507
+ "eval_recall": 0.8609467455621301,
508
+ "eval_runtime": 3.0559,
509
+ "eval_samples_per_second": 111.586,
510
+ "eval_steps_per_second": 3.6,
511
+ "step": 600
512
+ },
513
+ {
514
+ "epoch": 6.310880829015544,
515
+ "grad_norm": 0.20262780785560608,
516
+ "learning_rate": 3.431270895621349e-05,
517
+ "loss": 0.0038215465222795806,
518
+ "step": 612
519
+ },
520
+ {
521
+ "epoch": 6.435233160621761,
522
+ "grad_norm": 0.08969880640506744,
523
+ "learning_rate": 3.3665133803108035e-05,
524
+ "loss": 0.002952239786585172,
525
+ "step": 624
526
+ },
527
+ {
528
+ "epoch": 6.5595854922279795,
529
+ "grad_norm": 7.001764297485352,
530
+ "learning_rate": 3.301088120869224e-05,
531
+ "loss": 0.006995424007376035,
532
+ "step": 636
533
+ },
534
+ {
535
+ "epoch": 6.683937823834197,
536
+ "grad_norm": 1.759656548500061,
537
+ "learning_rate": 3.235045534675045e-05,
538
+ "loss": 0.0032693296670913696,
539
+ "step": 648
540
+ },
541
+ {
542
+ "epoch": 6.704663212435233,
543
+ "eval_accuracy": 0.9537932516655921,
544
+ "eval_f1": 0.8183098591549295,
545
+ "eval_loss": 0.04099448770284653,
546
+ "eval_precision": 0.7809139784946236,
547
+ "eval_recall": 0.8594674556213018,
548
+ "eval_runtime": 3.0102,
549
+ "eval_samples_per_second": 113.281,
550
+ "eval_steps_per_second": 3.654,
551
+ "step": 650
552
+ },
553
+ {
554
+ "epoch": 6.808290155440415,
555
+ "grad_norm": 0.113290935754776,
556
+ "learning_rate": 3.168436514825052e-05,
557
+ "loss": 0.003636385935048262,
558
+ "step": 660
559
+ },
560
+ {
561
+ "epoch": 6.932642487046632,
562
+ "grad_norm": 0.4593416452407837,
563
+ "learning_rate": 3.101312390915634e-05,
564
+ "loss": 0.006359308337171872,
565
+ "step": 672
566
+ },
567
+ {
568
+ "epoch": 7.051813471502591,
569
+ "grad_norm": 0.5328609943389893,
570
+ "learning_rate": 3.0337248894876746e-05,
571
+ "loss": 0.0031539862975478172,
572
+ "step": 684
573
+ },
574
+ {
575
+ "epoch": 7.176165803108808,
576
+ "grad_norm": 0.13898538053035736,
577
+ "learning_rate": 2.9657260941655417e-05,
578
+ "loss": 0.0014738718358178933,
579
+ "step": 696
580
+ },
581
+ {
582
+ "epoch": 7.217616580310881,
583
+ "eval_accuracy": 0.9636793466580701,
584
+ "eval_f1": 0.8498542274052479,
585
+ "eval_loss": 0.041410233825445175,
586
+ "eval_precision": 0.8376436781609196,
587
+ "eval_recall": 0.8624260355029586,
588
+ "eval_runtime": 3.0173,
589
+ "eval_samples_per_second": 113.015,
590
+ "eval_steps_per_second": 3.646,
591
+ "step": 700
592
+ },
593
+ {
594
+ "epoch": 7.300518134715026,
595
+ "grad_norm": 0.020028043538331985,
596
+ "learning_rate": 2.8973684055209106e-05,
597
+ "loss": 0.0009494482850035032,
598
+ "step": 708
599
+ },
600
+ {
601
+ "epoch": 7.424870466321243,
602
+ "grad_norm": 0.038508884608745575,
603
+ "learning_rate": 2.8287045006923473e-05,
604
+ "loss": 0.002579918752113978,
605
+ "step": 720
606
+ },
607
+ {
608
+ "epoch": 7.549222797927461,
609
+ "grad_norm": 0.46958020329475403,
610
+ "learning_rate": 2.7597872927917613e-05,
611
+ "loss": 0.0010905038410176833,
612
+ "step": 732
613
+ },
614
+ {
615
+ "epoch": 7.6735751295336785,
616
+ "grad_norm": 0.1547262817621231,
617
+ "learning_rate": 2.6906698901290167e-05,
618
+ "loss": 0.0028232391923666,
619
+ "step": 744
620
+ },
621
+ {
622
+ "epoch": 7.7357512953367875,
623
+ "eval_accuracy": 0.957016978293574,
624
+ "eval_f1": 0.829302987197724,
625
+ "eval_loss": 0.04088299721479416,
626
+ "eval_precision": 0.7986301369863014,
627
+ "eval_recall": 0.8624260355029586,
628
+ "eval_runtime": 3.0234,
629
+ "eval_samples_per_second": 112.787,
630
+ "eval_steps_per_second": 3.638,
631
+ "step": 750
632
+ },
633
+ {
634
+ "epoch": 7.7979274611398965,
635
+ "grad_norm": 0.0519477054476738,
636
+ "learning_rate": 2.621405555286121e-05,
637
+ "loss": 0.0010418800326685111,
638
+ "step": 756
639
+ },
640
+ {
641
+ "epoch": 7.922279792746114,
642
+ "grad_norm": 1.0287493467330933,
643
+ "learning_rate": 2.5520476640725342e-05,
644
+ "loss": 0.001736572477966547,
645
+ "step": 768
646
+ },
647
+ {
648
+ "epoch": 8.041450777202073,
649
+ "grad_norm": 0.013289994560182095,
650
+ "learning_rate": 2.4826496643932183e-05,
651
+ "loss": 0.000329588889144361,
652
+ "step": 780
653
+ },
654
+ {
655
+ "epoch": 8.16580310880829,
656
+ "grad_norm": 0.011798656545579433,
657
+ "learning_rate": 2.413265035061131e-05,
658
+ "loss": 0.002016165293753147,
659
+ "step": 792
660
+ },
661
+ {
662
+ "epoch": 8.248704663212436,
663
+ "eval_accuracy": 0.9630346013324736,
664
+ "eval_f1": 0.8407079646017699,
665
+ "eval_loss": 0.05126392841339111,
666
+ "eval_precision": 0.8382352941176471,
667
+ "eval_recall": 0.8431952662721893,
668
+ "eval_runtime": 3.0399,
669
+ "eval_samples_per_second": 112.176,
670
+ "eval_steps_per_second": 3.619,
671
+ "step": 800
672
+ },
673
+ {
674
+ "epoch": 8.290155440414507,
675
+ "grad_norm": 0.016478369012475014,
676
+ "learning_rate": 2.343947244585902e-05,
677
+ "loss": 0.001583070649454991,
678
+ "step": 804
679
+ },
680
+ {
681
+ "epoch": 8.414507772020725,
682
+ "grad_norm": 0.08483371138572693,
683
+ "learning_rate": 2.2747497099704533e-05,
684
+ "loss": 0.0006651872924218575,
685
+ "step": 816
686
+ },
687
+ {
688
+ "epoch": 8.538860103626943,
689
+ "grad_norm": 0.02956315688788891,
690
+ "learning_rate": 2.2057257555472998e-05,
691
+ "loss": 0.001361117387811343,
692
+ "step": 828
693
+ },
694
+ {
695
+ "epoch": 8.663212435233161,
696
+ "grad_norm": 0.1703813225030899,
697
+ "learning_rate": 2.136928571886275e-05,
698
+ "loss": 0.0004198816216861208,
699
+ "step": 840
700
+ },
701
+ {
702
+ "epoch": 8.766839378238341,
703
+ "eval_accuracy": 0.9636793466580701,
704
+ "eval_f1": 0.851528384279476,
705
+ "eval_loss": 0.04630826786160469,
706
+ "eval_precision": 0.83810888252149,
707
+ "eval_recall": 0.8653846153846154,
708
+ "eval_runtime": 3.0066,
709
+ "eval_samples_per_second": 113.418,
710
+ "eval_steps_per_second": 3.659,
711
+ "step": 850
712
+ },
713
+ {
714
+ "epoch": 8.787564766839377,
715
+ "grad_norm": 0.08187761902809143,
716
+ "learning_rate": 2.068411174805327e-05,
717
+ "loss": 0.00021340335176015893,
718
+ "step": 852
719
+ },
720
+ {
721
+ "epoch": 8.911917098445596,
722
+ "grad_norm": 0.1566244512796402,
723
+ "learning_rate": 2.0002263645159892e-05,
724
+ "loss": 0.0004732525364185373,
725
+ "step": 864
726
+ },
727
+ {
728
+ "epoch": 9.031088082901555,
729
+ "grad_norm": 0.007061833515763283,
730
+ "learning_rate": 1.9324266849349848e-05,
731
+ "loss": 9.06402322774132e-05,
732
+ "step": 876
733
+ },
734
+ {
735
+ "epoch": 9.155440414507773,
736
+ "grad_norm": 0.010846185497939587,
737
+ "learning_rate": 1.8650643831933502e-05,
738
+ "loss": 0.00011390061505759756,
739
+ "step": 888
740
+ },
741
+ {
742
+ "epoch": 9.279792746113989,
743
+ "grad_norm": 0.03851994872093201,
744
+ "learning_rate": 1.798191369374258e-05,
745
+ "loss": 0.00019203778356313705,
746
+ "step": 900
747
+ },
748
+ {
749
+ "epoch": 9.279792746113989,
750
+ "eval_accuracy": 0.9649688373092629,
751
+ "eval_f1": 0.8546255506607928,
752
+ "eval_loss": 0.05020074546337128,
753
+ "eval_precision": 0.8483965014577259,
754
+ "eval_recall": 0.8609467455621301,
755
+ "eval_runtime": 3.0339,
756
+ "eval_samples_per_second": 112.397,
757
+ "eval_steps_per_second": 3.626,
758
+ "step": 900
759
+ },
760
+ {
761
+ "epoch": 9.404145077720207,
762
+ "grad_norm": 0.006672121584415436,
763
+ "learning_rate": 1.731859176510569e-05,
764
+ "loss": 9.344543407981594e-05,
765
+ "step": 912
766
+ },
767
+ {
768
+ "epoch": 9.528497409326425,
769
+ "grad_norm": 0.028979798778891563,
770
+ "learning_rate": 1.666118920872949e-05,
771
+ "loss": 0.0009806355616698663,
772
+ "step": 924
773
+ },
774
+ {
775
+ "epoch": 9.652849740932643,
776
+ "grad_norm": 0.010359357111155987,
777
+ "learning_rate": 1.6010212625791422e-05,
778
+ "loss": 0.00018761774602656564,
779
+ "step": 936
780
+ },
781
+ {
782
+ "epoch": 9.77720207253886,
783
+ "grad_norm": 0.0016930641140788794,
784
+ "learning_rate": 1.536616366554768e-05,
785
+ "loss": 0.00010497189941816032,
786
+ "step": 948
787
+ },
788
+ {
789
+ "epoch": 9.797927461139896,
790
+ "eval_accuracy": 0.9641091768751343,
791
+ "eval_f1": 0.849112426035503,
792
+ "eval_loss": 0.05002701282501221,
793
+ "eval_precision": 0.849112426035503,
794
+ "eval_recall": 0.849112426035503,
795
+ "eval_runtime": 3.0151,
796
+ "eval_samples_per_second": 113.098,
797
+ "eval_steps_per_second": 3.648,
798
+ "step": 950
799
+ },
800
+ {
801
+ "epoch": 9.901554404145077,
802
+ "grad_norm": 0.0343826562166214,
803
+ "learning_rate": 1.472953863875706e-05,
804
+ "loss": 8.524257767324646e-05,
805
+ "step": 960
806
+ },
807
+ {
808
+ "epoch": 10.020725388601036,
809
+ "grad_norm": 0.006111979950219393,
810
+ "learning_rate": 1.4100828135218829e-05,
811
+ "loss": 0.00012398386024869978,
812
+ "step": 972
813
+ },
814
+ {
815
+ "epoch": 10.145077720207254,
816
+ "grad_norm": 0.010606050491333008,
817
+ "learning_rate": 1.3480516645719172e-05,
818
+ "loss": 3.8064138304131724e-05,
819
+ "step": 984
820
+ },
821
+ {
822
+ "epoch": 10.26943005181347,
823
+ "grad_norm": 0.0009117216104641557,
824
+ "learning_rate": 1.2869082188677572e-05,
825
+ "loss": 0.0001223099728425344,
826
+ "step": 996
827
+ },
828
+ {
829
+ "epoch": 10.310880829015543,
830
+ "eval_accuracy": 0.9645390070921985,
831
+ "eval_f1": 0.8514412416851442,
832
+ "eval_loss": 0.05051806941628456,
833
+ "eval_precision": 0.8508124076809453,
834
+ "eval_recall": 0.8520710059171598,
835
+ "eval_runtime": 3.0202,
836
+ "eval_samples_per_second": 112.906,
837
+ "eval_steps_per_second": 3.642,
838
+ "step": 1000
839
+ },
840
+ {
841
+ "epoch": 10.393782383419689,
842
+ "grad_norm": 0.0027385777793824673,
843
+ "learning_rate": 1.2266995941780934e-05,
844
+ "loss": 7.653955738836278e-05,
845
+ "step": 1008
846
+ },
847
+ {
848
+ "epoch": 10.518134715025907,
849
+ "grad_norm": 0.00011921550321858376,
850
+ "learning_rate": 1.1674721878889171e-05,
851
+ "loss": 0.0001742592042622467,
852
+ "step": 1020
853
+ },
854
+ {
855
+ "epoch": 10.642487046632125,
856
+ "grad_norm": 0.001234560739248991,
857
+ "learning_rate": 1.1092716412492226e-05,
858
+ "loss": 7.672411447856575e-05,
859
+ "step": 1032
860
+ },
861
+ {
862
+ "epoch": 10.766839378238341,
863
+ "grad_norm": 0.0005847270367667079,
864
+ "learning_rate": 1.0521428041993848e-05,
865
+ "loss": 9.932015624751026e-06,
866
+ "step": 1044
867
+ },
868
+ {
869
+ "epoch": 10.82901554404145,
870
+ "eval_accuracy": 0.9643240919836664,
871
+ "eval_f1": 0.8508124076809453,
872
+ "eval_loss": 0.05043612793087959,
873
+ "eval_precision": 0.8495575221238938,
874
+ "eval_recall": 0.8520710059171598,
875
+ "eval_runtime": 3.024,
876
+ "eval_samples_per_second": 112.764,
877
+ "eval_steps_per_second": 3.638,
878
+ "step": 1050
879
+ },
880
+ {
881
+ "epoch": 10.89119170984456,
882
+ "grad_norm": 0.0753927081823349,
883
+ "learning_rate": 9.961297008093357e-06,
884
+ "loss": 0.00011471325221161048,
885
+ "step": 1056
886
+ },
887
+ {
888
+ "epoch": 11.010362694300518,
889
+ "grad_norm": 0.003958659712225199,
890
+ "learning_rate": 9.412754953531663e-06,
891
+ "loss": 1.709427791259562e-05,
892
+ "step": 1068
893
+ },
894
+ {
895
+ "epoch": 11.134715025906736,
896
+ "grad_norm": 0.001931481878273189,
897
+ "learning_rate": 8.876224590462873e-06,
898
+ "loss": 1.0689032933441922e-05,
899
+ "step": 1080
900
+ },
901
+ {
902
+ "epoch": 11.259067357512953,
903
+ "grad_norm": 0.012324509210884571,
904
+ "learning_rate": 8.352119374707978e-06,
905
+ "loss": 5.9023882689264916e-05,
906
+ "step": 1092
907
+ },
908
+ {
909
+ "epoch": 11.341968911917098,
910
+ "eval_accuracy": 0.9643240919836664,
911
+ "eval_f1": 0.8510324483775811,
912
+ "eval_loss": 0.050491511821746826,
913
+ "eval_precision": 0.8485294117647059,
914
+ "eval_recall": 0.8535502958579881,
915
+ "eval_runtime": 3.0303,
916
+ "eval_samples_per_second": 112.529,
917
+ "eval_steps_per_second": 3.63,
918
+ "step": 1100
919
+ },
920
+ {
921
+ "epoch": 11.38341968911917,
922
+ "grad_norm": 0.002621831139549613,
923
+ "learning_rate": 7.840843187141544e-06,
924
+ "loss": 5.975224000091354e-05,
925
+ "step": 1104
926
+ },
927
+ {
928
+ "epoch": 11.507772020725389,
929
+ "grad_norm": 0.0006910113152116537,
930
+ "learning_rate": 7.3427900224569606e-06,
931
+ "loss": 7.24908944296961e-05,
932
+ "step": 1116
933
+ },
934
+ {
935
+ "epoch": 11.632124352331607,
936
+ "grad_norm": 0.01195142138749361,
937
+ "learning_rate": 6.858343685550048e-06,
938
+ "loss": 3.2996586620962866e-05,
939
+ "step": 1128
940
+ },
941
+ {
942
+ "epoch": 11.756476683937823,
943
+ "grad_norm": 0.00024170154938474298,
944
+ "learning_rate": 6.387877495755065e-06,
945
+ "loss": 1.4914168180742612e-05,
946
+ "step": 1140
947
+ },
948
+ {
949
+ "epoch": 11.860103626943005,
950
+ "eval_accuracy": 0.9645390070921985,
951
+ "eval_f1": 0.8514412416851442,
952
+ "eval_loss": 0.05080362409353256,
953
+ "eval_precision": 0.8508124076809453,
954
+ "eval_recall": 0.8520710059171598,
955
+ "eval_runtime": 3.0169,
956
+ "eval_samples_per_second": 113.03,
957
+ "eval_steps_per_second": 3.646,
958
+ "step": 1150
959
+ },
960
+ {
961
+ "epoch": 11.880829015544041,
962
+ "grad_norm": 0.00013972603483125567,
963
+ "learning_rate": 5.93175399916101e-06,
964
+ "loss": 0.00017348669158915678,
965
+ "step": 1152
966
+ },
967
+ {
968
+ "epoch": 12.0,
969
+ "grad_norm": 0.014590195380151272,
970
+ "learning_rate": 5.490324689229834e-06,
971
+ "loss": 6.250141692968707e-05,
972
+ "step": 1164
973
+ },
974
+ {
975
+ "epoch": 12.124352331606218,
976
+ "grad_norm": 0.002015709411352873,
977
+ "learning_rate": 5.063929735931985e-06,
978
+ "loss": 1.4617435226682574e-05,
979
+ "step": 1176
980
+ },
981
+ {
982
+ "epoch": 12.248704663212436,
983
+ "grad_norm": 0.0005398832727223635,
984
+ "learning_rate": 4.652897723607913e-06,
985
+ "loss": 1.6918191249715164e-05,
986
+ "step": 1188
987
+ },
988
+ {
989
+ "epoch": 12.373056994818652,
990
+ "grad_norm": 0.0006741413380950689,
991
+ "learning_rate": 4.257545397757623e-06,
992
+ "loss": 4.216020170133561e-05,
993
+ "step": 1200
994
+ },
995
+ {
996
+ "epoch": 12.373056994818652,
997
+ "eval_accuracy": 0.9645390070921985,
998
+ "eval_f1": 0.8514412416851442,
999
+ "eval_loss": 0.05097455903887749,
1000
+ "eval_precision": 0.8508124076809453,
1001
+ "eval_recall": 0.8520710059171598,
1002
+ "eval_runtime": 3.0165,
1003
+ "eval_samples_per_second": 113.043,
1004
+ "eval_steps_per_second": 3.647,
1005
+ "step": 1200
1006
+ },
1007
+ {
1008
+ "epoch": 12.49740932642487,
1009
+ "grad_norm": 0.0007238159305416048,
1010
+ "learning_rate": 3.878177420953291e-06,
1011
+ "loss": 5.365341106274476e-05,
1012
+ "step": 1212
1013
+ },
1014
+ {
1015
+ "epoch": 12.621761658031089,
1016
+ "grad_norm": 0.002653967123478651,
1017
+ "learning_rate": 3.5150861380631755e-06,
1018
+ "loss": 3.87875431139643e-05,
1019
+ "step": 1224
1020
+ },
1021
+ {
1022
+ "epoch": 12.746113989637305,
1023
+ "grad_norm": 0.0012897817650809884,
1024
+ "learning_rate": 3.1685513509676606e-06,
1025
+ "loss": 1.6210615285672247e-05,
1026
+ "step": 1236
1027
+ },
1028
+ {
1029
+ "epoch": 12.870466321243523,
1030
+ "grad_norm": 0.010877900756895542,
1031
+ "learning_rate": 2.838840102941015e-06,
1032
+ "loss": 4.428145863736669e-05,
1033
+ "step": 1248
1034
+ },
1035
+ {
1036
+ "epoch": 12.89119170984456,
1037
+ "eval_accuracy": 0.9645390070921985,
1038
+ "eval_f1": 0.8514412416851442,
1039
+ "eval_loss": 0.05106152221560478,
1040
+ "eval_precision": 0.8508124076809453,
1041
+ "eval_recall": 0.8520710059171598,
1042
+ "eval_runtime": 3.0194,
1043
+ "eval_samples_per_second": 112.937,
1044
+ "eval_steps_per_second": 3.643,
1045
+ "step": 1250
1046
+ },
1047
+ {
1048
+ "epoch": 12.994818652849741,
1049
+ "grad_norm": 0.004842772614210844,
1050
+ "learning_rate": 2.52620647286512e-06,
1051
+ "loss": 4.890701772334675e-05,
1052
+ "step": 1260
1053
+ },
1054
+ {
1055
+ "epoch": 13.1139896373057,
1056
+ "grad_norm": 0.005717657506465912,
1057
+ "learning_rate": 2.2308913794336583e-06,
1058
+ "loss": 8.704484207555652e-05,
1059
+ "step": 1272
1060
+ },
1061
+ {
1062
+ "epoch": 13.238341968911918,
1063
+ "grad_norm": 0.0031942392233759165,
1064
+ "learning_rate": 1.953122395497711e-06,
1065
+ "loss": 3.180075388324136e-05,
1066
+ "step": 1284
1067
+ },
1068
+ {
1069
+ "epoch": 13.362694300518134,
1070
+ "grad_norm": 0.006513912230730057,
1071
+ "learning_rate": 1.6931135726957315e-06,
1072
+ "loss": 1.0696584164785842e-05,
1073
+ "step": 1296
1074
+ },
1075
+ {
1076
+ "epoch": 13.404145077720207,
1077
+ "eval_accuracy": 0.9645390070921985,
1078
+ "eval_f1": 0.8514412416851442,
1079
+ "eval_loss": 0.05115008354187012,
1080
+ "eval_precision": 0.8508124076809453,
1081
+ "eval_recall": 0.8520710059171598,
1082
+ "eval_runtime": 3.0366,
1083
+ "eval_samples_per_second": 112.298,
1084
+ "eval_steps_per_second": 3.623,
1085
+ "step": 1300
1086
+ },
1087
+ {
1088
+ "epoch": 13.487046632124352,
1089
+ "grad_norm": 0.03252461925148964,
1090
+ "learning_rate": 1.4510652765031829e-06,
1091
+ "loss": 4.1350130535041295e-05,
1092
+ "step": 1308
1093
+ },
1094
+ {
1095
+ "epoch": 13.61139896373057,
1096
+ "grad_norm": 0.0007789316587150097,
1097
+ "learning_rate": 1.2271640318288397e-06,
1098
+ "loss": 9.945468870379651e-06,
1099
+ "step": 1320
1100
+ },
1101
+ {
1102
+ "epoch": 13.735751295336787,
1103
+ "grad_norm": 0.005432860925793648,
1104
+ "learning_rate": 1.0215823792767615e-06,
1105
+ "loss": 2.4644163204357028e-05,
1106
+ "step": 1332
1107
+ },
1108
+ {
1109
+ "epoch": 13.860103626943005,
1110
+ "grad_norm": 0.07517717033624649,
1111
+ "learning_rate": 8.344787421847217e-07,
1112
+ "loss": 0.00010969570333448549,
1113
+ "step": 1344
1114
+ },
1115
+ {
1116
+ "epoch": 13.922279792746114,
1117
+ "eval_accuracy": 0.9645390070921985,
1118
+ "eval_f1": 0.8514412416851442,
1119
+ "eval_loss": 0.05114062502980232,
1120
+ "eval_precision": 0.8508124076809453,
1121
+ "eval_recall": 0.8520710059171598,
1122
+ "eval_runtime": 3.0311,
1123
+ "eval_samples_per_second": 112.5,
1124
+ "eval_steps_per_second": 3.629,
1125
+ "step": 1350
1126
+ },
1127
+ {
1128
+ "epoch": 13.984455958549223,
1129
+ "grad_norm": 0.004264460410922766,
1130
+ "learning_rate": 6.659973045415685e-07,
1131
+ "loss": 7.129671575967222e-05,
1132
+ "step": 1356
1133
+ },
1134
+ {
1135
+ "epoch": 14.103626943005182,
1136
+ "grad_norm": 0.00019057428289670497,
1137
+ "learning_rate": 5.162678998775567e-07,
1138
+ "loss": 1.6013045145276312e-05,
1139
+ "step": 1368
1140
+ },
1141
+ {
1142
+ "epoch": 14.2279792746114,
1143
+ "grad_norm": 0.008129872381687164,
1144
+ "learning_rate": 3.8540591121327686e-07,
1145
+ "loss": 3.3401636756025255e-05,
1146
+ "step": 1380
1147
+ },
1148
+ {
1149
+ "epoch": 14.352331606217616,
1150
+ "grad_norm": 0.03226028382778168,
1151
+ "learning_rate": 2.7351218214433647e-07,
1152
+ "loss": 7.967630517669022e-05,
1153
+ "step": 1392
1154
+ },
1155
+ {
1156
+ "epoch": 14.435233160621761,
1157
+ "eval_accuracy": 0.9645390070921985,
1158
+ "eval_f1": 0.8514412416851442,
1159
+ "eval_loss": 0.05112980678677559,
1160
+ "eval_precision": 0.8508124076809453,
1161
+ "eval_recall": 0.8520710059171598,
1162
+ "eval_runtime": 3.0382,
1163
+ "eval_samples_per_second": 112.238,
1164
+ "eval_steps_per_second": 3.621,
1165
+ "step": 1400
1166
+ }
1167
+ ],
1168
+ "logging_steps": 12,
1169
+ "max_steps": 1455,
1170
+ "num_input_tokens_seen": 0,
1171
+ "num_train_epochs": 15,
1172
+ "save_steps": 50,
1173
+ "stateful_callbacks": {
1174
+ "EarlyStoppingCallback": {
1175
+ "args": {
1176
+ "early_stopping_patience": 10,
1177
+ "early_stopping_threshold": 0.0
1178
+ },
1179
+ "attributes": {
1180
+ "early_stopping_patience_counter": 10
1181
+ }
1182
+ },
1183
+ "TrainerControl": {
1184
+ "args": {
1185
+ "should_epoch_stop": false,
1186
+ "should_evaluate": false,
1187
+ "should_log": false,
1188
+ "should_save": true,
1189
+ "should_training_stop": true
1190
+ },
1191
+ "attributes": {}
1192
+ }
1193
+ },
1194
+ "total_flos": 2.3548810040103936e+16,
1195
+ "train_batch_size": 8,
1196
+ "trial_name": null,
1197
+ "trial_params": null
1198
+ }
checkpoint-1400/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:32e421217e1eea826c4025f111c9f4fa0b9d53a8bca71632cae9ad6aed6b26f4
3
+ size 5201
checkpoint-900/config.json ADDED
@@ -0,0 +1,99 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "architectures": [
3
+ "ModernBertForTokenClassification"
4
+ ],
5
+ "attention_bias": false,
6
+ "attention_dropout": 0.0,
7
+ "bos_token_id": 50281,
8
+ "classifier_activation": "gelu",
9
+ "classifier_bias": false,
10
+ "classifier_dropout": 0.3,
11
+ "classifier_pooling": "mean",
12
+ "cls_token_id": 50281,
13
+ "decoder_bias": true,
14
+ "deterministic_flash_attn": false,
15
+ "dtype": "float32",
16
+ "embedding_dropout": 0.0,
17
+ "eos_token_id": 50282,
18
+ "global_attn_every_n_layers": 3,
19
+ "gradient_checkpointing": false,
20
+ "hidden_activation": "gelu",
21
+ "hidden_size": 1024,
22
+ "id2label": {
23
+ "0": "O",
24
+ "1": "FP",
25
+ "2": "RP",
26
+ "3": "RV",
27
+ "4": "PW"
28
+ },
29
+ "initializer_cutoff_factor": 2.0,
30
+ "initializer_range": 0.02,
31
+ "intermediate_size": 2624,
32
+ "label2id": {
33
+ "FP": 1,
34
+ "O": 0,
35
+ "PW": 4,
36
+ "RP": 2,
37
+ "RV": 3
38
+ },
39
+ "layer_norm_eps": 1e-05,
40
+ "layer_types": [
41
+ "full_attention",
42
+ "sliding_attention",
43
+ "sliding_attention",
44
+ "full_attention",
45
+ "sliding_attention",
46
+ "sliding_attention",
47
+ "full_attention",
48
+ "sliding_attention",
49
+ "sliding_attention",
50
+ "full_attention",
51
+ "sliding_attention",
52
+ "sliding_attention",
53
+ "full_attention",
54
+ "sliding_attention",
55
+ "sliding_attention",
56
+ "full_attention",
57
+ "sliding_attention",
58
+ "sliding_attention",
59
+ "full_attention",
60
+ "sliding_attention",
61
+ "sliding_attention",
62
+ "full_attention",
63
+ "sliding_attention",
64
+ "sliding_attention",
65
+ "full_attention",
66
+ "sliding_attention",
67
+ "sliding_attention",
68
+ "full_attention"
69
+ ],
70
+ "local_attention": 128,
71
+ "max_position_embeddings": 8192,
72
+ "mlp_bias": false,
73
+ "mlp_dropout": 0.0,
74
+ "model_type": "modernbert",
75
+ "norm_bias": false,
76
+ "norm_eps": 1e-05,
77
+ "num_attention_heads": 16,
78
+ "num_hidden_layers": 28,
79
+ "pad_token_id": 50283,
80
+ "position_embedding_type": "absolute",
81
+ "repad_logits_with_grad": false,
82
+ "rope_parameters": {
83
+ "full_attention": {
84
+ "rope_theta": 160000.0,
85
+ "rope_type": "default"
86
+ },
87
+ "sliding_attention": {
88
+ "rope_theta": 10000.0,
89
+ "rope_type": "default"
90
+ }
91
+ },
92
+ "sep_token_id": 50282,
93
+ "sparse_pred_ignore_index": -100,
94
+ "sparse_prediction": false,
95
+ "tie_word_embeddings": true,
96
+ "transformers_version": "5.0.0",
97
+ "use_cache": false,
98
+ "vocab_size": 50368
99
+ }
checkpoint-900/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d0821184b2d65c0626e9be2d035c7a860180d6fb63ab1b2ea98bb119972eefa2
3
+ size 1583363940
checkpoint-900/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0c6b652664dc5ea0bcaac98d62b2e77c77253b1c023e4b3c2d96a97f68614fea
3
+ size 1112419851
checkpoint-900/rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0ee75aede4debbc6b85d9c083fb35d50c7b3ad23c0f1c0733c939b26dd633dfb
3
+ size 14709
checkpoint-900/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:779730cd96cf9cb6808d4bee932cd31f0febd2dab73c0ff6dabc9fccc57c2591
3
+ size 1465
checkpoint-900/tokenizer.json ADDED
The diff for this file is too large to render. See raw diff
 
checkpoint-900/tokenizer_config.json ADDED
@@ -0,0 +1,16 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "backend": "tokenizers",
3
+ "clean_up_tokenization_spaces": true,
4
+ "cls_token": "[CLS]",
5
+ "is_local": false,
6
+ "mask_token": "[MASK]",
7
+ "model_input_names": [
8
+ "input_ids",
9
+ "attention_mask"
10
+ ],
11
+ "model_max_length": 8192,
12
+ "pad_token": "[PAD]",
13
+ "sep_token": "[SEP]",
14
+ "tokenizer_class": "TokenizersBackend",
15
+ "unk_token": "[UNK]"
16
+ }
checkpoint-900/trainer_state.json ADDED
@@ -0,0 +1,791 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_global_step": 900,
3
+ "best_metric": 0.8546255506607928,
4
+ "best_model_checkpoint": "/content/drive/MyDrive/expE_large_oversample_rv/checkpoint-900",
5
+ "epoch": 9.279792746113989,
6
+ "eval_steps": 50,
7
+ "global_step": 900,
8
+ "is_hyper_param_search": false,
9
+ "is_local_process_zero": true,
10
+ "is_world_process_zero": true,
11
+ "log_history": [
12
+ {
13
+ "epoch": 0.010362694300518135,
14
+ "grad_norm": 19.391862869262695,
15
+ "learning_rate": 0.0,
16
+ "loss": 0.8294010162353516,
17
+ "step": 1
18
+ },
19
+ {
20
+ "epoch": 0.12435233160621761,
21
+ "grad_norm": 7.022524833679199,
22
+ "learning_rate": 5.670103092783505e-06,
23
+ "loss": 0.6434053074229847,
24
+ "step": 12
25
+ },
26
+ {
27
+ "epoch": 0.24870466321243523,
28
+ "grad_norm": 2.0728249549865723,
29
+ "learning_rate": 1.1855670103092783e-05,
30
+ "loss": 0.5182687441507975,
31
+ "step": 24
32
+ },
33
+ {
34
+ "epoch": 0.37305699481865284,
35
+ "grad_norm": 8.758841514587402,
36
+ "learning_rate": 1.8041237113402062e-05,
37
+ "loss": 0.39159242312113446,
38
+ "step": 36
39
+ },
40
+ {
41
+ "epoch": 0.49740932642487046,
42
+ "grad_norm": 10.565062522888184,
43
+ "learning_rate": 2.422680412371134e-05,
44
+ "loss": 0.31471407413482666,
45
+ "step": 48
46
+ },
47
+ {
48
+ "epoch": 0.5181347150259067,
49
+ "eval_accuracy": 0.7721899849559424,
50
+ "eval_f1": 0.4335260115606937,
51
+ "eval_loss": 0.05006009340286255,
52
+ "eval_precision": 0.32142857142857145,
53
+ "eval_recall": 0.665680473372781,
54
+ "eval_runtime": 5.6283,
55
+ "eval_samples_per_second": 60.586,
56
+ "eval_steps_per_second": 1.954,
57
+ "step": 50
58
+ },
59
+ {
60
+ "epoch": 0.6217616580310881,
61
+ "grad_norm": 5.744125843048096,
62
+ "learning_rate": 3.0412371134020617e-05,
63
+ "loss": 0.28232719500859577,
64
+ "step": 60
65
+ },
66
+ {
67
+ "epoch": 0.7461139896373057,
68
+ "grad_norm": 4.099153518676758,
69
+ "learning_rate": 3.6597938144329896e-05,
70
+ "loss": 0.22093822558720908,
71
+ "step": 72
72
+ },
73
+ {
74
+ "epoch": 0.8704663212435233,
75
+ "grad_norm": 12.179475784301758,
76
+ "learning_rate": 4.2783505154639175e-05,
77
+ "loss": 0.2065215309460958,
78
+ "step": 84
79
+ },
80
+ {
81
+ "epoch": 0.9948186528497409,
82
+ "grad_norm": 8.685556411743164,
83
+ "learning_rate": 4.8969072164948454e-05,
84
+ "loss": 0.16811863581339517,
85
+ "step": 96
86
+ },
87
+ {
88
+ "epoch": 1.0310880829015545,
89
+ "eval_accuracy": 0.7461852568235547,
90
+ "eval_f1": 0.4638047138047138,
91
+ "eval_loss": 0.03645005077123642,
92
+ "eval_precision": 0.3241176470588235,
93
+ "eval_recall": 0.8150887573964497,
94
+ "eval_runtime": 3.0466,
95
+ "eval_samples_per_second": 111.926,
96
+ "eval_steps_per_second": 3.611,
97
+ "step": 100
98
+ },
99
+ {
100
+ "epoch": 1.1139896373056994,
101
+ "grad_norm": 33.46553421020508,
102
+ "learning_rate": 4.999331054325924e-05,
103
+ "loss": 0.15307321151097616,
104
+ "step": 108
105
+ },
106
+ {
107
+ "epoch": 1.238341968911917,
108
+ "grad_norm": 5.198877811431885,
109
+ "learning_rate": 4.996762857385083e-05,
110
+ "loss": 0.1403733491897583,
111
+ "step": 120
112
+ },
113
+ {
114
+ "epoch": 1.3626943005181347,
115
+ "grad_norm": 1.8454481363296509,
116
+ "learning_rate": 4.9922706292838655e-05,
117
+ "loss": 0.13873055577278137,
118
+ "step": 132
119
+ },
120
+ {
121
+ "epoch": 1.4870466321243523,
122
+ "grad_norm": 2.0861477851867676,
123
+ "learning_rate": 4.9858578317794894e-05,
124
+ "loss": 0.11305248737335205,
125
+ "step": 144
126
+ },
127
+ {
128
+ "epoch": 1.549222797927461,
129
+ "eval_accuracy": 0.895766172361917,
130
+ "eval_f1": 0.6706515242080096,
131
+ "eval_loss": 0.029485274106264114,
132
+ "eval_precision": 0.5626880641925778,
133
+ "eval_recall": 0.8298816568047337,
134
+ "eval_runtime": 3.0006,
135
+ "eval_samples_per_second": 113.645,
136
+ "eval_steps_per_second": 3.666,
137
+ "step": 150
138
+ },
139
+ {
140
+ "epoch": 1.61139896373057,
141
+ "grad_norm": 6.362898349761963,
142
+ "learning_rate": 4.9775294066397274e-05,
143
+ "loss": 0.11805283029874165,
144
+ "step": 156
145
+ },
146
+ {
147
+ "epoch": 1.7357512953367875,
148
+ "grad_norm": 2.3624749183654785,
149
+ "learning_rate": 4.967291771834727e-05,
150
+ "loss": 0.11144719521204631,
151
+ "step": 168
152
+ },
153
+ {
154
+ "epoch": 1.8601036269430051,
155
+ "grad_norm": 0.9627630710601807,
156
+ "learning_rate": 4.955152816591259e-05,
157
+ "loss": 0.0819690078496933,
158
+ "step": 180
159
+ },
160
+ {
161
+ "epoch": 1.9844559585492227,
162
+ "grad_norm": 15.513121604919434,
163
+ "learning_rate": 4.9411218953131964e-05,
164
+ "loss": 0.10844972729682922,
165
+ "step": 192
166
+ },
167
+ {
168
+ "epoch": 2.062176165803109,
169
+ "eval_accuracy": 0.8074360627552117,
170
+ "eval_f1": 0.5410536307546274,
171
+ "eval_loss": 0.030047627165913582,
172
+ "eval_precision": 0.39832285115303984,
173
+ "eval_recall": 0.8431952662721893,
174
+ "eval_runtime": 3.037,
175
+ "eval_samples_per_second": 112.283,
176
+ "eval_steps_per_second": 3.622,
177
+ "step": 200
178
+ },
179
+ {
180
+ "epoch": 2.1036269430051813,
181
+ "grad_norm": 2.684053421020508,
182
+ "learning_rate": 4.9252098203729154e-05,
183
+ "loss": 0.08770081400871277,
184
+ "step": 204
185
+ },
186
+ {
187
+ "epoch": 2.227979274611399,
188
+ "grad_norm": 21.072328567504883,
189
+ "learning_rate": 4.907428853779167e-05,
190
+ "loss": 0.1440110206604004,
191
+ "step": 216
192
+ },
193
+ {
194
+ "epoch": 2.3523316062176165,
195
+ "grad_norm": 3.221358299255371,
196
+ "learning_rate": 4.887792697727854e-05,
197
+ "loss": 0.11832749843597412,
198
+ "step": 228
199
+ },
200
+ {
201
+ "epoch": 2.476683937823834,
202
+ "grad_norm": 5.4783101081848145,
203
+ "learning_rate": 4.866316484042968e-05,
204
+ "loss": 0.0703568160533905,
205
+ "step": 240
206
+ },
207
+ {
208
+ "epoch": 2.5803108808290154,
209
+ "eval_accuracy": 0.9129593810444874,
210
+ "eval_f1": 0.7118012422360248,
211
+ "eval_loss": 0.0311787910759449,
212
+ "eval_precision": 0.6134903640256959,
213
+ "eval_recall": 0.8476331360946746,
214
+ "eval_runtime": 3.0166,
215
+ "eval_samples_per_second": 113.041,
216
+ "eval_steps_per_second": 3.646,
217
+ "step": 250
218
+ },
219
+ {
220
+ "epoch": 2.6010362694300517,
221
+ "grad_norm": 6.384552478790283,
222
+ "learning_rate": 4.8430167625158595e-05,
223
+ "loss": 0.09287564953168233,
224
+ "step": 252
225
+ },
226
+ {
227
+ "epoch": 2.7253886010362693,
228
+ "grad_norm": 4.217259407043457,
229
+ "learning_rate": 4.817911488151791e-05,
230
+ "loss": 0.08391642570495605,
231
+ "step": 264
232
+ },
233
+ {
234
+ "epoch": 2.849740932642487,
235
+ "grad_norm": 3.1646366119384766,
236
+ "learning_rate": 4.791020007333629e-05,
237
+ "loss": 0.061600317557652794,
238
+ "step": 276
239
+ },
240
+ {
241
+ "epoch": 2.9740932642487046,
242
+ "grad_norm": 6.243884563446045,
243
+ "learning_rate": 4.762363042913322e-05,
244
+ "loss": 0.05470449725786845,
245
+ "step": 288
246
+ },
247
+ {
248
+ "epoch": 3.093264248704663,
249
+ "grad_norm": 2.5704944133758545,
250
+ "learning_rate": 4.7319626782426564e-05,
251
+ "loss": 0.05606028437614441,
252
+ "step": 300
253
+ },
254
+ {
255
+ "epoch": 3.093264248704663,
256
+ "eval_accuracy": 0.7977648828712659,
257
+ "eval_f1": 0.5429740791268759,
258
+ "eval_loss": 0.03227110207080841,
259
+ "eval_precision": 0.39198949441891007,
260
+ "eval_recall": 0.8831360946745562,
261
+ "eval_runtime": 3.0347,
262
+ "eval_samples_per_second": 112.367,
263
+ "eval_steps_per_second": 3.625,
264
+ "step": 300
265
+ },
266
+ {
267
+ "epoch": 3.2176165803108807,
268
+ "grad_norm": 2.420072317123413,
269
+ "learning_rate": 4.699842340155598e-05,
270
+ "loss": 0.053138395150502525,
271
+ "step": 312
272
+ },
273
+ {
274
+ "epoch": 3.3419689119170983,
275
+ "grad_norm": 67.39490509033203,
276
+ "learning_rate": 4.666026780915334e-05,
277
+ "loss": 0.10085086027781169,
278
+ "step": 324
279
+ },
280
+ {
281
+ "epoch": 3.466321243523316,
282
+ "grad_norm": 0.7816835045814514,
283
+ "learning_rate": 4.630542059139924e-05,
284
+ "loss": 0.0743176539738973,
285
+ "step": 336
286
+ },
287
+ {
288
+ "epoch": 3.5906735751295336,
289
+ "grad_norm": 0.7841729521751404,
290
+ "learning_rate": 4.593415519721257e-05,
291
+ "loss": 0.05843483408292135,
292
+ "step": 348
293
+ },
294
+ {
295
+ "epoch": 3.61139896373057,
296
+ "eval_accuracy": 0.9329464861379755,
297
+ "eval_f1": 0.7535070140280562,
298
+ "eval_loss": 0.030000876635313034,
299
+ "eval_precision": 0.6869671132764921,
300
+ "eval_recall": 0.834319526627219,
301
+ "eval_runtime": 3.0255,
302
+ "eval_samples_per_second": 112.709,
303
+ "eval_steps_per_second": 3.636,
304
+ "step": 350
305
+ },
306
+ {
307
+ "epoch": 3.715025906735751,
308
+ "grad_norm": 5.054598331451416,
309
+ "learning_rate": 4.5546757727528035e-05,
310
+ "loss": 0.03325176239013672,
311
+ "step": 360
312
+ },
313
+ {
314
+ "epoch": 3.839378238341969,
315
+ "grad_norm": 0.5077673196792603,
316
+ "learning_rate": 4.5143526714823766e-05,
317
+ "loss": 0.036297169824441276,
318
+ "step": 372
319
+ },
320
+ {
321
+ "epoch": 3.9637305699481864,
322
+ "grad_norm": 0.8186505436897278,
323
+ "learning_rate": 4.472477289306913e-05,
324
+ "loss": 0.026280519862969715,
325
+ "step": 384
326
+ },
327
+ {
328
+ "epoch": 4.082901554404145,
329
+ "grad_norm": 0.9546571373939514,
330
+ "learning_rate": 4.429081895826997e-05,
331
+ "loss": 0.038141210873921715,
332
+ "step": 396
333
+ },
334
+ {
335
+ "epoch": 4.124352331606218,
336
+ "eval_accuracy": 0.9116698903932946,
337
+ "eval_f1": 0.7129798903107862,
338
+ "eval_loss": 0.02928924560546875,
339
+ "eval_precision": 0.6062176165803109,
340
+ "eval_recall": 0.8653846153846154,
341
+ "eval_runtime": 3.0233,
342
+ "eval_samples_per_second": 112.79,
343
+ "eval_steps_per_second": 3.638,
344
+ "step": 400
345
+ },
346
+ {
347
+ "epoch": 4.2072538860103625,
348
+ "grad_norm": 0.38423895835876465,
349
+ "learning_rate": 4.38419993197957e-05,
350
+ "loss": 0.023756501575311024,
351
+ "step": 408
352
+ },
353
+ {
354
+ "epoch": 4.331606217616581,
355
+ "grad_norm": 0.261392205953598,
356
+ "learning_rate": 4.337865984268001e-05,
357
+ "loss": 0.02620025227467219,
358
+ "step": 420
359
+ },
360
+ {
361
+ "epoch": 4.455958549222798,
362
+ "grad_norm": 2.0392212867736816,
363
+ "learning_rate": 4.290115758109375e-05,
364
+ "loss": 0.02181059867143631,
365
+ "step": 432
366
+ },
367
+ {
368
+ "epoch": 4.580310880829016,
369
+ "grad_norm": 0.4685036242008209,
370
+ "learning_rate": 4.240986050319524e-05,
371
+ "loss": 0.023431827624638874,
372
+ "step": 444
373
+ },
374
+ {
375
+ "epoch": 4.642487046632124,
376
+ "eval_accuracy": 0.9593810444874274,
377
+ "eval_f1": 0.828297715549005,
378
+ "eval_loss": 0.04263457655906677,
379
+ "eval_precision": 0.8252569750367107,
380
+ "eval_recall": 0.8313609467455622,
381
+ "eval_runtime": 3.0402,
382
+ "eval_samples_per_second": 112.164,
383
+ "eval_steps_per_second": 3.618,
384
+ "step": 450
385
+ },
386
+ {
387
+ "epoch": 4.704663212435233,
388
+ "grad_norm": 0.535210371017456,
389
+ "learning_rate": 4.19051472075703e-05,
390
+ "loss": 0.021484971046447754,
391
+ "step": 456
392
+ },
393
+ {
394
+ "epoch": 4.829015544041451,
395
+ "grad_norm": 3.4881668090820312,
396
+ "learning_rate": 4.138740663148019e-05,
397
+ "loss": 0.03922188033660253,
398
+ "step": 468
399
+ },
400
+ {
401
+ "epoch": 4.953367875647668,
402
+ "grad_norm": 0.809377908706665,
403
+ "learning_rate": 4.085703775114266e-05,
404
+ "loss": 0.018220727642377216,
405
+ "step": 480
406
+ },
407
+ {
408
+ "epoch": 5.072538860103627,
409
+ "grad_norm": 3.266059160232544,
410
+ "learning_rate": 4.0314449274276663e-05,
411
+ "loss": 0.015343919396400452,
412
+ "step": 492
413
+ },
414
+ {
415
+ "epoch": 5.155440414507772,
416
+ "eval_accuracy": 0.9574468085106383,
417
+ "eval_f1": 0.8162672476397966,
418
+ "eval_loss": 0.04887477308511734,
419
+ "eval_precision": 0.8017118402282454,
420
+ "eval_recall": 0.8313609467455622,
421
+ "eval_runtime": 3.0737,
422
+ "eval_samples_per_second": 110.94,
423
+ "eval_steps_per_second": 3.579,
424
+ "step": 500
425
+ },
426
+ {
427
+ "epoch": 5.196891191709844,
428
+ "grad_norm": 0.14335286617279053,
429
+ "learning_rate": 3.976005932514807e-05,
430
+ "loss": 0.007969564447800318,
431
+ "step": 504
432
+ },
433
+ {
434
+ "epoch": 5.321243523316062,
435
+ "grad_norm": 0.24810202419757843,
436
+ "learning_rate": 3.919429512235874e-05,
437
+ "loss": 0.007458047941327095,
438
+ "step": 516
439
+ },
440
+ {
441
+ "epoch": 5.44559585492228,
442
+ "grad_norm": 1.7369506359100342,
443
+ "learning_rate": 3.86175926496275e-05,
444
+ "loss": 0.012967783957719803,
445
+ "step": 528
446
+ },
447
+ {
448
+ "epoch": 5.569948186528498,
449
+ "grad_norm": 0.40230634808540344,
450
+ "learning_rate": 3.8030396319816585e-05,
451
+ "loss": 0.008079041416446367,
452
+ "step": 540
453
+ },
454
+ {
455
+ "epoch": 5.6735751295336785,
456
+ "eval_accuracy": 0.9591661293788953,
457
+ "eval_f1": 0.8310168251645941,
458
+ "eval_loss": 0.04243873059749603,
459
+ "eval_precision": 0.8219971056439942,
460
+ "eval_recall": 0.8402366863905325,
461
+ "eval_runtime": 3.0267,
462
+ "eval_samples_per_second": 112.662,
463
+ "eval_steps_per_second": 3.634,
464
+ "step": 550
465
+ },
466
+ {
467
+ "epoch": 5.694300518134715,
468
+ "grad_norm": 0.46429362893104553,
469
+ "learning_rate": 3.743315863246256e-05,
470
+ "loss": 0.005399468044439952,
471
+ "step": 552
472
+ },
473
+ {
474
+ "epoch": 5.818652849740933,
475
+ "grad_norm": 1.60989248752594,
476
+ "learning_rate": 3.682633982507543e-05,
477
+ "loss": 0.006222298989693324,
478
+ "step": 564
479
+ },
480
+ {
481
+ "epoch": 5.94300518134715,
482
+ "grad_norm": 0.5653553009033203,
483
+ "learning_rate": 3.621040751847497e-05,
484
+ "loss": 0.015771896888812382,
485
+ "step": 576
486
+ },
487
+ {
488
+ "epoch": 6.062176165803109,
489
+ "grad_norm": 0.34870776534080505,
490
+ "learning_rate": 3.5585836356437264e-05,
491
+ "loss": 0.0066079869866371155,
492
+ "step": 588
493
+ },
494
+ {
495
+ "epoch": 6.186528497409326,
496
+ "grad_norm": 0.2842996120452881,
497
+ "learning_rate": 3.495310763992928e-05,
498
+ "loss": 0.0051950619866450625,
499
+ "step": 600
500
+ },
501
+ {
502
+ "epoch": 6.186528497409326,
503
+ "eval_accuracy": 0.9580915538362347,
504
+ "eval_f1": 0.8308351177730193,
505
+ "eval_loss": 0.04230675473809242,
506
+ "eval_precision": 0.8027586206896552,
507
+ "eval_recall": 0.8609467455621301,
508
+ "eval_runtime": 3.0559,
509
+ "eval_samples_per_second": 111.586,
510
+ "eval_steps_per_second": 3.6,
511
+ "step": 600
512
+ },
513
+ {
514
+ "epoch": 6.310880829015544,
515
+ "grad_norm": 0.20262780785560608,
516
+ "learning_rate": 3.431270895621349e-05,
517
+ "loss": 0.0038215465222795806,
518
+ "step": 612
519
+ },
520
+ {
521
+ "epoch": 6.435233160621761,
522
+ "grad_norm": 0.08969880640506744,
523
+ "learning_rate": 3.3665133803108035e-05,
524
+ "loss": 0.002952239786585172,
525
+ "step": 624
526
+ },
527
+ {
528
+ "epoch": 6.5595854922279795,
529
+ "grad_norm": 7.001764297485352,
530
+ "learning_rate": 3.301088120869224e-05,
531
+ "loss": 0.006995424007376035,
532
+ "step": 636
533
+ },
534
+ {
535
+ "epoch": 6.683937823834197,
536
+ "grad_norm": 1.759656548500061,
537
+ "learning_rate": 3.235045534675045e-05,
538
+ "loss": 0.0032693296670913696,
539
+ "step": 648
540
+ },
541
+ {
542
+ "epoch": 6.704663212435233,
543
+ "eval_accuracy": 0.9537932516655921,
544
+ "eval_f1": 0.8183098591549295,
545
+ "eval_loss": 0.04099448770284653,
546
+ "eval_precision": 0.7809139784946236,
547
+ "eval_recall": 0.8594674556213018,
548
+ "eval_runtime": 3.0102,
549
+ "eval_samples_per_second": 113.281,
550
+ "eval_steps_per_second": 3.654,
551
+ "step": 650
552
+ },
553
+ {
554
+ "epoch": 6.808290155440415,
555
+ "grad_norm": 0.113290935754776,
556
+ "learning_rate": 3.168436514825052e-05,
557
+ "loss": 0.003636385935048262,
558
+ "step": 660
559
+ },
560
+ {
561
+ "epoch": 6.932642487046632,
562
+ "grad_norm": 0.4593416452407837,
563
+ "learning_rate": 3.101312390915634e-05,
564
+ "loss": 0.006359308337171872,
565
+ "step": 672
566
+ },
567
+ {
568
+ "epoch": 7.051813471502591,
569
+ "grad_norm": 0.5328609943389893,
570
+ "learning_rate": 3.0337248894876746e-05,
571
+ "loss": 0.0031539862975478172,
572
+ "step": 684
573
+ },
574
+ {
575
+ "epoch": 7.176165803108808,
576
+ "grad_norm": 0.13898538053035736,
577
+ "learning_rate": 2.9657260941655417e-05,
578
+ "loss": 0.0014738718358178933,
579
+ "step": 696
580
+ },
581
+ {
582
+ "epoch": 7.217616580310881,
583
+ "eval_accuracy": 0.9636793466580701,
584
+ "eval_f1": 0.8498542274052479,
585
+ "eval_loss": 0.041410233825445175,
586
+ "eval_precision": 0.8376436781609196,
587
+ "eval_recall": 0.8624260355029586,
588
+ "eval_runtime": 3.0173,
589
+ "eval_samples_per_second": 113.015,
590
+ "eval_steps_per_second": 3.646,
591
+ "step": 700
592
+ },
593
+ {
594
+ "epoch": 7.300518134715026,
595
+ "grad_norm": 0.020028043538331985,
596
+ "learning_rate": 2.8973684055209106e-05,
597
+ "loss": 0.0009494482850035032,
598
+ "step": 708
599
+ },
600
+ {
601
+ "epoch": 7.424870466321243,
602
+ "grad_norm": 0.038508884608745575,
603
+ "learning_rate": 2.8287045006923473e-05,
604
+ "loss": 0.002579918752113978,
605
+ "step": 720
606
+ },
607
+ {
608
+ "epoch": 7.549222797927461,
609
+ "grad_norm": 0.46958020329475403,
610
+ "learning_rate": 2.7597872927917613e-05,
611
+ "loss": 0.0010905038410176833,
612
+ "step": 732
613
+ },
614
+ {
615
+ "epoch": 7.6735751295336785,
616
+ "grad_norm": 0.1547262817621231,
617
+ "learning_rate": 2.6906698901290167e-05,
618
+ "loss": 0.0028232391923666,
619
+ "step": 744
620
+ },
621
+ {
622
+ "epoch": 7.7357512953367875,
623
+ "eval_accuracy": 0.957016978293574,
624
+ "eval_f1": 0.829302987197724,
625
+ "eval_loss": 0.04088299721479416,
626
+ "eval_precision": 0.7986301369863014,
627
+ "eval_recall": 0.8624260355029586,
628
+ "eval_runtime": 3.0234,
629
+ "eval_samples_per_second": 112.787,
630
+ "eval_steps_per_second": 3.638,
631
+ "step": 750
632
+ },
633
+ {
634
+ "epoch": 7.7979274611398965,
635
+ "grad_norm": 0.0519477054476738,
636
+ "learning_rate": 2.621405555286121e-05,
637
+ "loss": 0.0010418800326685111,
638
+ "step": 756
639
+ },
640
+ {
641
+ "epoch": 7.922279792746114,
642
+ "grad_norm": 1.0287493467330933,
643
+ "learning_rate": 2.5520476640725342e-05,
644
+ "loss": 0.001736572477966547,
645
+ "step": 768
646
+ },
647
+ {
648
+ "epoch": 8.041450777202073,
649
+ "grad_norm": 0.013289994560182095,
650
+ "learning_rate": 2.4826496643932183e-05,
651
+ "loss": 0.000329588889144361,
652
+ "step": 780
653
+ },
654
+ {
655
+ "epoch": 8.16580310880829,
656
+ "grad_norm": 0.011798656545579433,
657
+ "learning_rate": 2.413265035061131e-05,
658
+ "loss": 0.002016165293753147,
659
+ "step": 792
660
+ },
661
+ {
662
+ "epoch": 8.248704663212436,
663
+ "eval_accuracy": 0.9630346013324736,
664
+ "eval_f1": 0.8407079646017699,
665
+ "eval_loss": 0.05126392841339111,
666
+ "eval_precision": 0.8382352941176471,
667
+ "eval_recall": 0.8431952662721893,
668
+ "eval_runtime": 3.0399,
669
+ "eval_samples_per_second": 112.176,
670
+ "eval_steps_per_second": 3.619,
671
+ "step": 800
672
+ },
673
+ {
674
+ "epoch": 8.290155440414507,
675
+ "grad_norm": 0.016478369012475014,
676
+ "learning_rate": 2.343947244585902e-05,
677
+ "loss": 0.001583070649454991,
678
+ "step": 804
679
+ },
680
+ {
681
+ "epoch": 8.414507772020725,
682
+ "grad_norm": 0.08483371138572693,
683
+ "learning_rate": 2.2747497099704533e-05,
684
+ "loss": 0.0006651872924218575,
685
+ "step": 816
686
+ },
687
+ {
688
+ "epoch": 8.538860103626943,
689
+ "grad_norm": 0.02956315688788891,
690
+ "learning_rate": 2.2057257555472998e-05,
691
+ "loss": 0.001361117387811343,
692
+ "step": 828
693
+ },
694
+ {
695
+ "epoch": 8.663212435233161,
696
+ "grad_norm": 0.1703813225030899,
697
+ "learning_rate": 2.136928571886275e-05,
698
+ "loss": 0.0004198816216861208,
699
+ "step": 840
700
+ },
701
+ {
702
+ "epoch": 8.766839378238341,
703
+ "eval_accuracy": 0.9636793466580701,
704
+ "eval_f1": 0.851528384279476,
705
+ "eval_loss": 0.04630826786160469,
706
+ "eval_precision": 0.83810888252149,
707
+ "eval_recall": 0.8653846153846154,
708
+ "eval_runtime": 3.0066,
709
+ "eval_samples_per_second": 113.418,
710
+ "eval_steps_per_second": 3.659,
711
+ "step": 850
712
+ },
713
+ {
714
+ "epoch": 8.787564766839377,
715
+ "grad_norm": 0.08187761902809143,
716
+ "learning_rate": 2.068411174805327e-05,
717
+ "loss": 0.00021340335176015893,
718
+ "step": 852
719
+ },
720
+ {
721
+ "epoch": 8.911917098445596,
722
+ "grad_norm": 0.1566244512796402,
723
+ "learning_rate": 2.0002263645159892e-05,
724
+ "loss": 0.0004732525364185373,
725
+ "step": 864
726
+ },
727
+ {
728
+ "epoch": 9.031088082901555,
729
+ "grad_norm": 0.007061833515763283,
730
+ "learning_rate": 1.9324266849349848e-05,
731
+ "loss": 9.06402322774132e-05,
732
+ "step": 876
733
+ },
734
+ {
735
+ "epoch": 9.155440414507773,
736
+ "grad_norm": 0.010846185497939587,
737
+ "learning_rate": 1.8650643831933502e-05,
738
+ "loss": 0.00011390061505759756,
739
+ "step": 888
740
+ },
741
+ {
742
+ "epoch": 9.279792746113989,
743
+ "grad_norm": 0.03851994872093201,
744
+ "learning_rate": 1.798191369374258e-05,
745
+ "loss": 0.00019203778356313705,
746
+ "step": 900
747
+ },
748
+ {
749
+ "epoch": 9.279792746113989,
750
+ "eval_accuracy": 0.9649688373092629,
751
+ "eval_f1": 0.8546255506607928,
752
+ "eval_loss": 0.05020074546337128,
753
+ "eval_precision": 0.8483965014577259,
754
+ "eval_recall": 0.8609467455621301,
755
+ "eval_runtime": 3.0339,
756
+ "eval_samples_per_second": 112.397,
757
+ "eval_steps_per_second": 3.626,
758
+ "step": 900
759
+ }
760
+ ],
761
+ "logging_steps": 12,
762
+ "max_steps": 1455,
763
+ "num_input_tokens_seen": 0,
764
+ "num_train_epochs": 15,
765
+ "save_steps": 50,
766
+ "stateful_callbacks": {
767
+ "EarlyStoppingCallback": {
768
+ "args": {
769
+ "early_stopping_patience": 10,
770
+ "early_stopping_threshold": 0.0
771
+ },
772
+ "attributes": {
773
+ "early_stopping_patience_counter": 0
774
+ }
775
+ },
776
+ "TrainerControl": {
777
+ "args": {
778
+ "should_epoch_stop": false,
779
+ "should_evaluate": false,
780
+ "should_log": false,
781
+ "should_save": true,
782
+ "should_training_stop": false
783
+ },
784
+ "attributes": {}
785
+ }
786
+ },
787
+ "total_flos": 1.5138520740066816e+16,
788
+ "train_batch_size": 8,
789
+ "trial_name": null,
790
+ "trial_params": null
791
+ }
checkpoint-900/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:32e421217e1eea826c4025f111c9f4fa0b9d53a8bca71632cae9ad6aed6b26f4
3
+ size 5201
config.json ADDED
@@ -0,0 +1,99 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "architectures": [
3
+ "ModernBertForTokenClassification"
4
+ ],
5
+ "attention_bias": false,
6
+ "attention_dropout": 0.0,
7
+ "bos_token_id": 50281,
8
+ "classifier_activation": "gelu",
9
+ "classifier_bias": false,
10
+ "classifier_dropout": 0.3,
11
+ "classifier_pooling": "mean",
12
+ "cls_token_id": 50281,
13
+ "decoder_bias": true,
14
+ "deterministic_flash_attn": false,
15
+ "dtype": "float32",
16
+ "embedding_dropout": 0.0,
17
+ "eos_token_id": 50282,
18
+ "global_attn_every_n_layers": 3,
19
+ "gradient_checkpointing": false,
20
+ "hidden_activation": "gelu",
21
+ "hidden_size": 1024,
22
+ "id2label": {
23
+ "0": "O",
24
+ "1": "FP",
25
+ "2": "RP",
26
+ "3": "RV",
27
+ "4": "PW"
28
+ },
29
+ "initializer_cutoff_factor": 2.0,
30
+ "initializer_range": 0.02,
31
+ "intermediate_size": 2624,
32
+ "label2id": {
33
+ "FP": 1,
34
+ "O": 0,
35
+ "PW": 4,
36
+ "RP": 2,
37
+ "RV": 3
38
+ },
39
+ "layer_norm_eps": 1e-05,
40
+ "layer_types": [
41
+ "full_attention",
42
+ "sliding_attention",
43
+ "sliding_attention",
44
+ "full_attention",
45
+ "sliding_attention",
46
+ "sliding_attention",
47
+ "full_attention",
48
+ "sliding_attention",
49
+ "sliding_attention",
50
+ "full_attention",
51
+ "sliding_attention",
52
+ "sliding_attention",
53
+ "full_attention",
54
+ "sliding_attention",
55
+ "sliding_attention",
56
+ "full_attention",
57
+ "sliding_attention",
58
+ "sliding_attention",
59
+ "full_attention",
60
+ "sliding_attention",
61
+ "sliding_attention",
62
+ "full_attention",
63
+ "sliding_attention",
64
+ "sliding_attention",
65
+ "full_attention",
66
+ "sliding_attention",
67
+ "sliding_attention",
68
+ "full_attention"
69
+ ],
70
+ "local_attention": 128,
71
+ "max_position_embeddings": 8192,
72
+ "mlp_bias": false,
73
+ "mlp_dropout": 0.0,
74
+ "model_type": "modernbert",
75
+ "norm_bias": false,
76
+ "norm_eps": 1e-05,
77
+ "num_attention_heads": 16,
78
+ "num_hidden_layers": 28,
79
+ "pad_token_id": 50283,
80
+ "position_embedding_type": "absolute",
81
+ "repad_logits_with_grad": false,
82
+ "rope_parameters": {
83
+ "full_attention": {
84
+ "rope_theta": 160000.0,
85
+ "rope_type": "default"
86
+ },
87
+ "sliding_attention": {
88
+ "rope_theta": 10000.0,
89
+ "rope_type": "default"
90
+ }
91
+ },
92
+ "sep_token_id": 50282,
93
+ "sparse_pred_ignore_index": -100,
94
+ "sparse_prediction": false,
95
+ "tie_word_embeddings": true,
96
+ "transformers_version": "5.0.0",
97
+ "use_cache": false,
98
+ "vocab_size": 50368
99
+ }
confusion_matrix.png ADDED

Git LFS Details

  • SHA256: a51146bf9bcb4ded40c6d832dd0b491660dfd8042de3e728e8e3b8c840998c71
  • Pointer size: 131 Bytes
  • Size of remote file: 116 kB
model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d0821184b2d65c0626e9be2d035c7a860180d6fb63ab1b2ea98bb119972eefa2
3
+ size 1583363940