Hugging Face's logo

hoangcung165
/
scanning-model-bert

Text Classification
TensorBoard
Safetensors
mpnet
Trained with AutoTrain
Model card Files Metrics Community
hoangcung165 commited on
Commit
7e9ed69
·
verified ·
1 Parent(s): a3a64aa

Upload folder using huggingface_hub

Browse files
Files changed (18) hide show
  1. README.md +36 -0
  2. checkpoint-7473/config.json +54 -0
  3. checkpoint-7473/model.safetensors +3 -0
  4. checkpoint-7473/optimizer.pt +3 -0
  5. checkpoint-7473/rng_state.pth +3 -0
  6. checkpoint-7473/scheduler.pt +3 -0
  7. checkpoint-7473/trainer_state.json +2182 -0
  8. checkpoint-7473/training_args.bin +3 -0
  9. config.json +54 -0
  10. model.safetensors +3 -0
  11. runs/Nov17_06-15-53_r-hoangcung165-test-train-gwo6rh86-54315-0a1vk/events.out.tfevents.1731824155.r-hoangcung165-test-train-gwo6rh86-54315-0a1vk.2400.0 +2 -2
  12. runs/Nov17_06-15-53_r-hoangcung165-test-train-gwo6rh86-54315-0a1vk/events.out.tfevents.1731863830.r-hoangcung165-test-train-gwo6rh86-54315-0a1vk.2400.1 +3 -0
  13. special_tokens_map.json +51 -0
  14. tokenizer.json +0 -0
  15. tokenizer_config.json +72 -0
  16. training_args.bin +3 -0
  17. training_params.json +30 -0
  18. vocab.txt +0 -0
README.md ADDED
@@ -0,0 +1,36 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+
2
+ ---
3
+ tags:
4
+ - autotrain
5
+ - text-classification
6
+ base_model: sentence-transformers/all-mpnet-base-v2
7
+ widget:
8
+ - text: "I love AutoTrain"
9
+ ---
10
+
11
+ # Model Trained Using AutoTrain
12
+
13
+ - Problem type: Text Classification
14
+
15
+ ## Validation Metrics
16
+ loss: 0.8851153254508972
17
+
18
+ f1_macro: 0.6845343178671808
19
+
20
+ f1_micro: 0.6850662384584504
21
+
22
+ f1_weighted: 0.6834454222699836
23
+
24
+ precision_macro: 0.6899698563739772
25
+
26
+ precision_micro: 0.6850662384584504
27
+
28
+ precision_weighted: 0.6888476966796808
29
+
30
+ recall_macro: 0.6860535953088197
31
+
32
+ recall_micro: 0.6850662384584504
33
+
34
+ recall_weighted: 0.6850662384584504
35
+
36
+ accuracy: 0.6850662384584504
checkpoint-7473/config.json ADDED
@@ -0,0 +1,54 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "sentence-transformers/all-mpnet-base-v2",
3
+ "_num_labels": 12,
4
+ "architectures": [
5
+ "MPNetForSequenceClassification"
6
+ ],
7
+ "attention_probs_dropout_prob": 0.1,
8
+ "bos_token_id": 0,
9
+ "eos_token_id": 2,
10
+ "hidden_act": "gelu",
11
+ "hidden_dropout_prob": 0.1,
12
+ "hidden_size": 768,
13
+ "id2label": {
14
+ "0": "Ph\u00e2n Li\u1ec7t C\u1ea3m X\u00fac - Nh\u1eb9",
15
+ "1": "Ph\u00e2n Li\u1ec7t C\u1ea3m X\u00fac - N\u1eb7ng",
16
+ "2": "Ph\u00e2n Li\u1ec7t C\u1ea3m X\u00fac - Tr\u1ea7m Tr\u1ecdng",
17
+ "3": "R\u1ed1i Lo\u1ea1n Lo \u00c2u - Nh\u1eb9",
18
+ "4": "R\u1ed1i Lo\u1ea1n Lo \u00c2u - N\u1eb7ng",
19
+ "5": "R\u1ed1i Lo\u1ea1n Lo \u00c2u - Tr\u1ea7m Tr\u1ecdng",
20
+ "6": "Tr\u1ea7m C\u1ea3m - Nh\u1eb9",
21
+ "7": "Tr\u1ea7m C\u1ea3m - N\u1eb7ng",
22
+ "8": "Tr\u1ea7m C\u1ea3m - Tr\u1ea7m Tr\u1ecdng",
23
+ "9": "T\u1ef1 K\u1ef7 - Nh\u1eb9",
24
+ "10": "T\u1ef1 K\u1ef7 - N\u1eb7ng",
25
+ "11": "T\u1ef1 K\u1ef7 - Tr\u1ea7m Tr\u1ecdng"
26
+ },
27
+ "initializer_range": 0.02,
28
+ "intermediate_size": 3072,
29
+ "label2id": {
30
+ "Ph\u00e2n Li\u1ec7t C\u1ea3m X\u00fac - Nh\u1eb9": 0,
31
+ "Ph\u00e2n Li\u1ec7t C\u1ea3m X\u00fac - N\u1eb7ng": 1,
32
+ "Ph\u00e2n Li\u1ec7t C\u1ea3m X\u00fac - Tr\u1ea7m Tr\u1ecdng": 2,
33
+ "R\u1ed1i Lo\u1ea1n Lo \u00c2u - Nh\u1eb9": 3,
34
+ "R\u1ed1i Lo\u1ea1n Lo \u00c2u - N\u1eb7ng": 4,
35
+ "R\u1ed1i Lo\u1ea1n Lo \u00c2u - Tr\u1ea7m Tr\u1ecdng": 5,
36
+ "Tr\u1ea7m C\u1ea3m - Nh\u1eb9": 6,
37
+ "Tr\u1ea7m C\u1ea3m - N\u1eb7ng": 7,
38
+ "Tr\u1ea7m C\u1ea3m - Tr\u1ea7m Tr\u1ecdng": 8,
39
+ "T\u1ef1 K\u1ef7 - Nh\u1eb9": 9,
40
+ "T\u1ef1 K\u1ef7 - N\u1eb7ng": 10,
41
+ "T\u1ef1 K\u1ef7 - Tr\u1ea7m Tr\u1ecdng": 11
42
+ },
43
+ "layer_norm_eps": 1e-05,
44
+ "max_position_embeddings": 514,
45
+ "model_type": "mpnet",
46
+ "num_attention_heads": 12,
47
+ "num_hidden_layers": 12,
48
+ "pad_token_id": 1,
49
+ "problem_type": "single_label_classification",
50
+ "relative_attention_num_buckets": 32,
51
+ "torch_dtype": "float32",
52
+ "transformers_version": "4.46.2",
53
+ "vocab_size": 30527
54
+ }
checkpoint-7473/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d6f8750d7d265db701a13545b5afb36dc72592e2dc0a435717b0d2a8e7af1eb5
3
+ size 438005960
checkpoint-7473/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:2c86c23c5cd5792eb7f47406ab1b459d2a2b5ac4ed7c2b6e56c341dac6398528
3
+ size 876127994
checkpoint-7473/rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3eb8e412d026325e72324b340b54cb88cea5801b6bfe8491a25b1d663e1c6fb9
3
+ size 13990
checkpoint-7473/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:68b198871321d5f81135dc530f694ba687b450554cfebe5dd8071eda393fba93
3
+ size 1064
checkpoint-7473/trainer_state.json ADDED
@@ -0,0 +1,2182 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": 0.8851153254508972,
3
+ "best_model_checkpoint": "scanning-model-bert/checkpoint-7473",
4
+ "epoch": 3.0,
5
+ "eval_steps": 500,
6
+ "global_step": 7473,
7
+ "is_hyper_param_search": false,
8
+ "is_local_process_zero": true,
9
+ "is_world_process_zero": true,
10
+ "log_history": [
11
+ {
12
+ "epoch": 0.010036130068245684,
13
+ "grad_norm": 1.1764743328094482,
14
+ "learning_rate": 1.6711229946524065e-06,
15
+ "loss": 2.4853,
16
+ "step": 25
17
+ },
18
+ {
19
+ "epoch": 0.02007226013649137,
20
+ "grad_norm": 1.1972541809082031,
21
+ "learning_rate": 3.342245989304813e-06,
22
+ "loss": 2.4787,
23
+ "step": 50
24
+ },
25
+ {
26
+ "epoch": 0.030108390204737052,
27
+ "grad_norm": 1.1974806785583496,
28
+ "learning_rate": 5.013368983957219e-06,
29
+ "loss": 2.4831,
30
+ "step": 75
31
+ },
32
+ {
33
+ "epoch": 0.04014452027298274,
34
+ "grad_norm": 1.139513373374939,
35
+ "learning_rate": 6.684491978609626e-06,
36
+ "loss": 2.474,
37
+ "step": 100
38
+ },
39
+ {
40
+ "epoch": 0.05018065034122842,
41
+ "grad_norm": 1.4917333126068115,
42
+ "learning_rate": 8.355614973262032e-06,
43
+ "loss": 2.4656,
44
+ "step": 125
45
+ },
46
+ {
47
+ "epoch": 0.060216780409474105,
48
+ "grad_norm": 2.7864933013916016,
49
+ "learning_rate": 1.0026737967914438e-05,
50
+ "loss": 2.4514,
51
+ "step": 150
52
+ },
53
+ {
54
+ "epoch": 0.0702529104777198,
55
+ "grad_norm": 1.9038527011871338,
56
+ "learning_rate": 1.1697860962566845e-05,
57
+ "loss": 2.4267,
58
+ "step": 175
59
+ },
60
+ {
61
+ "epoch": 0.08028904054596547,
62
+ "grad_norm": 1.842796802520752,
63
+ "learning_rate": 1.3368983957219252e-05,
64
+ "loss": 2.3909,
65
+ "step": 200
66
+ },
67
+ {
68
+ "epoch": 0.09032517061421116,
69
+ "grad_norm": 1.5834379196166992,
70
+ "learning_rate": 1.5040106951871657e-05,
71
+ "loss": 2.4023,
72
+ "step": 225
73
+ },
74
+ {
75
+ "epoch": 0.10036130068245684,
76
+ "grad_norm": 1.8323915004730225,
77
+ "learning_rate": 1.6711229946524065e-05,
78
+ "loss": 2.3552,
79
+ "step": 250
80
+ },
81
+ {
82
+ "epoch": 0.11039743075070253,
83
+ "grad_norm": 6.8411173820495605,
84
+ "learning_rate": 1.8382352941176472e-05,
85
+ "loss": 2.3116,
86
+ "step": 275
87
+ },
88
+ {
89
+ "epoch": 0.12043356081894821,
90
+ "grad_norm": 3.661825656890869,
91
+ "learning_rate": 2.0053475935828877e-05,
92
+ "loss": 2.2989,
93
+ "step": 300
94
+ },
95
+ {
96
+ "epoch": 0.1304696908871939,
97
+ "grad_norm": 2.0745482444763184,
98
+ "learning_rate": 2.1724598930481285e-05,
99
+ "loss": 2.3182,
100
+ "step": 325
101
+ },
102
+ {
103
+ "epoch": 0.1405058209554396,
104
+ "grad_norm": 14.493195533752441,
105
+ "learning_rate": 2.339572192513369e-05,
106
+ "loss": 2.2371,
107
+ "step": 350
108
+ },
109
+ {
110
+ "epoch": 0.15054195102368526,
111
+ "grad_norm": 4.515138149261475,
112
+ "learning_rate": 2.5066844919786097e-05,
113
+ "loss": 2.1893,
114
+ "step": 375
115
+ },
116
+ {
117
+ "epoch": 0.16057808109193095,
118
+ "grad_norm": 4.537893295288086,
119
+ "learning_rate": 2.6737967914438505e-05,
120
+ "loss": 2.1654,
121
+ "step": 400
122
+ },
123
+ {
124
+ "epoch": 0.17061421116017664,
125
+ "grad_norm": 6.065027236938477,
126
+ "learning_rate": 2.8409090909090912e-05,
127
+ "loss": 2.0774,
128
+ "step": 425
129
+ },
130
+ {
131
+ "epoch": 0.18065034122842233,
132
+ "grad_norm": 3.3011465072631836,
133
+ "learning_rate": 3.0080213903743313e-05,
134
+ "loss": 2.1148,
135
+ "step": 450
136
+ },
137
+ {
138
+ "epoch": 0.190686471296668,
139
+ "grad_norm": 3.5985777378082275,
140
+ "learning_rate": 3.1751336898395725e-05,
141
+ "loss": 2.1128,
142
+ "step": 475
143
+ },
144
+ {
145
+ "epoch": 0.20072260136491368,
146
+ "grad_norm": 10.858529090881348,
147
+ "learning_rate": 3.342245989304813e-05,
148
+ "loss": 2.0861,
149
+ "step": 500
150
+ },
151
+ {
152
+ "epoch": 0.21075873143315937,
153
+ "grad_norm": 5.10444450378418,
154
+ "learning_rate": 3.509358288770054e-05,
155
+ "loss": 2.0483,
156
+ "step": 525
157
+ },
158
+ {
159
+ "epoch": 0.22079486150140507,
160
+ "grad_norm": 2.763580560684204,
161
+ "learning_rate": 3.6764705882352945e-05,
162
+ "loss": 1.9238,
163
+ "step": 550
164
+ },
165
+ {
166
+ "epoch": 0.23083099156965076,
167
+ "grad_norm": 3.7232158184051514,
168
+ "learning_rate": 3.843582887700535e-05,
169
+ "loss": 1.8426,
170
+ "step": 575
171
+ },
172
+ {
173
+ "epoch": 0.24086712163789642,
174
+ "grad_norm": 2.688575267791748,
175
+ "learning_rate": 4.0106951871657754e-05,
176
+ "loss": 1.87,
177
+ "step": 600
178
+ },
179
+ {
180
+ "epoch": 0.25090325170614214,
181
+ "grad_norm": 3.5620744228363037,
182
+ "learning_rate": 4.1778074866310165e-05,
183
+ "loss": 1.8417,
184
+ "step": 625
185
+ },
186
+ {
187
+ "epoch": 0.2609393817743878,
188
+ "grad_norm": 3.834071397781372,
189
+ "learning_rate": 4.344919786096257e-05,
190
+ "loss": 1.9035,
191
+ "step": 650
192
+ },
193
+ {
194
+ "epoch": 0.27097551184263347,
195
+ "grad_norm": 6.310031414031982,
196
+ "learning_rate": 4.512032085561498e-05,
197
+ "loss": 1.83,
198
+ "step": 675
199
+ },
200
+ {
201
+ "epoch": 0.2810116419108792,
202
+ "grad_norm": 10.144730567932129,
203
+ "learning_rate": 4.679144385026738e-05,
204
+ "loss": 1.8859,
205
+ "step": 700
206
+ },
207
+ {
208
+ "epoch": 0.29104777197912485,
209
+ "grad_norm": 6.981380939483643,
210
+ "learning_rate": 4.846256684491979e-05,
211
+ "loss": 1.9067,
212
+ "step": 725
213
+ },
214
+ {
215
+ "epoch": 0.3010839020473705,
216
+ "grad_norm": 6.095215797424316,
217
+ "learning_rate": 4.9985130111524165e-05,
218
+ "loss": 1.6395,
219
+ "step": 750
220
+ },
221
+ {
222
+ "epoch": 0.31112003211561623,
223
+ "grad_norm": 9.925978660583496,
224
+ "learning_rate": 4.979925650557621e-05,
225
+ "loss": 1.7119,
226
+ "step": 775
227
+ },
228
+ {
229
+ "epoch": 0.3211561621838619,
230
+ "grad_norm": 4.9185309410095215,
231
+ "learning_rate": 4.961338289962826e-05,
232
+ "loss": 1.6727,
233
+ "step": 800
234
+ },
235
+ {
236
+ "epoch": 0.3311922922521076,
237
+ "grad_norm": 5.485770225524902,
238
+ "learning_rate": 4.94275092936803e-05,
239
+ "loss": 1.7179,
240
+ "step": 825
241
+ },
242
+ {
243
+ "epoch": 0.3412284223203533,
244
+ "grad_norm": 4.848090648651123,
245
+ "learning_rate": 4.9241635687732344e-05,
246
+ "loss": 1.7757,
247
+ "step": 850
248
+ },
249
+ {
250
+ "epoch": 0.35126455238859894,
251
+ "grad_norm": 4.81159782409668,
252
+ "learning_rate": 4.905576208178439e-05,
253
+ "loss": 1.7111,
254
+ "step": 875
255
+ },
256
+ {
257
+ "epoch": 0.36130068245684466,
258
+ "grad_norm": 9.039434432983398,
259
+ "learning_rate": 4.886988847583644e-05,
260
+ "loss": 1.658,
261
+ "step": 900
262
+ },
263
+ {
264
+ "epoch": 0.3713368125250903,
265
+ "grad_norm": 4.779675483703613,
266
+ "learning_rate": 4.868401486988848e-05,
267
+ "loss": 1.7113,
268
+ "step": 925
269
+ },
270
+ {
271
+ "epoch": 0.381372942593336,
272
+ "grad_norm": 3.4945783615112305,
273
+ "learning_rate": 4.849814126394052e-05,
274
+ "loss": 1.5576,
275
+ "step": 950
276
+ },
277
+ {
278
+ "epoch": 0.3914090726615817,
279
+ "grad_norm": 7.573122978210449,
280
+ "learning_rate": 4.8312267657992566e-05,
281
+ "loss": 1.5566,
282
+ "step": 975
283
+ },
284
+ {
285
+ "epoch": 0.40144520272982737,
286
+ "grad_norm": 13.348621368408203,
287
+ "learning_rate": 4.812639405204461e-05,
288
+ "loss": 1.6002,
289
+ "step": 1000
290
+ },
291
+ {
292
+ "epoch": 0.4114813327980731,
293
+ "grad_norm": 3.8586764335632324,
294
+ "learning_rate": 4.794052044609665e-05,
295
+ "loss": 1.4931,
296
+ "step": 1025
297
+ },
298
+ {
299
+ "epoch": 0.42151746286631875,
300
+ "grad_norm": 8.584020614624023,
301
+ "learning_rate": 4.77546468401487e-05,
302
+ "loss": 1.58,
303
+ "step": 1050
304
+ },
305
+ {
306
+ "epoch": 0.4315535929345644,
307
+ "grad_norm": 9.650528907775879,
308
+ "learning_rate": 4.7568773234200745e-05,
309
+ "loss": 1.5622,
310
+ "step": 1075
311
+ },
312
+ {
313
+ "epoch": 0.44158972300281013,
314
+ "grad_norm": 5.387539863586426,
315
+ "learning_rate": 4.7382899628252794e-05,
316
+ "loss": 1.5855,
317
+ "step": 1100
318
+ },
319
+ {
320
+ "epoch": 0.4516258530710558,
321
+ "grad_norm": 5.179405689239502,
322
+ "learning_rate": 4.719702602230484e-05,
323
+ "loss": 1.5094,
324
+ "step": 1125
325
+ },
326
+ {
327
+ "epoch": 0.4616619831393015,
328
+ "grad_norm": 12.544023513793945,
329
+ "learning_rate": 4.701115241635688e-05,
330
+ "loss": 1.5518,
331
+ "step": 1150
332
+ },
333
+ {
334
+ "epoch": 0.4716981132075472,
335
+ "grad_norm": 5.561625003814697,
336
+ "learning_rate": 4.682527881040892e-05,
337
+ "loss": 1.5438,
338
+ "step": 1175
339
+ },
340
+ {
341
+ "epoch": 0.48173424327579284,
342
+ "grad_norm": 5.334999084472656,
343
+ "learning_rate": 4.6639405204460966e-05,
344
+ "loss": 1.4815,
345
+ "step": 1200
346
+ },
347
+ {
348
+ "epoch": 0.49177037334403856,
349
+ "grad_norm": 5.292882919311523,
350
+ "learning_rate": 4.645353159851301e-05,
351
+ "loss": 1.5336,
352
+ "step": 1225
353
+ },
354
+ {
355
+ "epoch": 0.5018065034122843,
356
+ "grad_norm": 4.656703948974609,
357
+ "learning_rate": 4.626765799256505e-05,
358
+ "loss": 1.4317,
359
+ "step": 1250
360
+ },
361
+ {
362
+ "epoch": 0.5118426334805299,
363
+ "grad_norm": 6.756280422210693,
364
+ "learning_rate": 4.60817843866171e-05,
365
+ "loss": 1.4975,
366
+ "step": 1275
367
+ },
368
+ {
369
+ "epoch": 0.5218787635487756,
370
+ "grad_norm": 6.566856861114502,
371
+ "learning_rate": 4.589591078066915e-05,
372
+ "loss": 1.4193,
373
+ "step": 1300
374
+ },
375
+ {
376
+ "epoch": 0.5319148936170213,
377
+ "grad_norm": 14.7003812789917,
378
+ "learning_rate": 4.5710037174721195e-05,
379
+ "loss": 1.3279,
380
+ "step": 1325
381
+ },
382
+ {
383
+ "epoch": 0.5419510236852669,
384
+ "grad_norm": 6.0828633308410645,
385
+ "learning_rate": 4.552416356877324e-05,
386
+ "loss": 1.508,
387
+ "step": 1350
388
+ },
389
+ {
390
+ "epoch": 0.5519871537535126,
391
+ "grad_norm": 4.892170429229736,
392
+ "learning_rate": 4.533828996282528e-05,
393
+ "loss": 1.3173,
394
+ "step": 1375
395
+ },
396
+ {
397
+ "epoch": 0.5620232838217584,
398
+ "grad_norm": 6.631337642669678,
399
+ "learning_rate": 4.5152416356877324e-05,
400
+ "loss": 1.3692,
401
+ "step": 1400
402
+ },
403
+ {
404
+ "epoch": 0.572059413890004,
405
+ "grad_norm": 10.353917121887207,
406
+ "learning_rate": 4.496654275092937e-05,
407
+ "loss": 1.3371,
408
+ "step": 1425
409
+ },
410
+ {
411
+ "epoch": 0.5820955439582497,
412
+ "grad_norm": 6.489056587219238,
413
+ "learning_rate": 4.478066914498141e-05,
414
+ "loss": 1.5097,
415
+ "step": 1450
416
+ },
417
+ {
418
+ "epoch": 0.5921316740264954,
419
+ "grad_norm": 3.911562442779541,
420
+ "learning_rate": 4.459479553903346e-05,
421
+ "loss": 1.3946,
422
+ "step": 1475
423
+ },
424
+ {
425
+ "epoch": 0.602167804094741,
426
+ "grad_norm": 7.033605098724365,
427
+ "learning_rate": 4.440892193308551e-05,
428
+ "loss": 1.3581,
429
+ "step": 1500
430
+ },
431
+ {
432
+ "epoch": 0.6122039341629868,
433
+ "grad_norm": 24.549911499023438,
434
+ "learning_rate": 4.422304832713755e-05,
435
+ "loss": 1.4775,
436
+ "step": 1525
437
+ },
438
+ {
439
+ "epoch": 0.6222400642312325,
440
+ "grad_norm": 5.103108882904053,
441
+ "learning_rate": 4.4037174721189596e-05,
442
+ "loss": 1.4047,
443
+ "step": 1550
444
+ },
445
+ {
446
+ "epoch": 0.6322761942994781,
447
+ "grad_norm": 6.033442974090576,
448
+ "learning_rate": 4.385130111524164e-05,
449
+ "loss": 1.3348,
450
+ "step": 1575
451
+ },
452
+ {
453
+ "epoch": 0.6423123243677238,
454
+ "grad_norm": 2.9465322494506836,
455
+ "learning_rate": 4.366542750929368e-05,
456
+ "loss": 1.3134,
457
+ "step": 1600
458
+ },
459
+ {
460
+ "epoch": 0.6523484544359694,
461
+ "grad_norm": 6.447948455810547,
462
+ "learning_rate": 4.3479553903345725e-05,
463
+ "loss": 1.3922,
464
+ "step": 1625
465
+ },
466
+ {
467
+ "epoch": 0.6623845845042152,
468
+ "grad_norm": 8.688138961791992,
469
+ "learning_rate": 4.329368029739777e-05,
470
+ "loss": 1.3806,
471
+ "step": 1650
472
+ },
473
+ {
474
+ "epoch": 0.6724207145724609,
475
+ "grad_norm": 8.916844367980957,
476
+ "learning_rate": 4.310780669144982e-05,
477
+ "loss": 1.4429,
478
+ "step": 1675
479
+ },
480
+ {
481
+ "epoch": 0.6824568446407066,
482
+ "grad_norm": 5.320075035095215,
483
+ "learning_rate": 4.292193308550186e-05,
484
+ "loss": 1.4155,
485
+ "step": 1700
486
+ },
487
+ {
488
+ "epoch": 0.6924929747089522,
489
+ "grad_norm": 7.1480302810668945,
490
+ "learning_rate": 4.2736059479553904e-05,
491
+ "loss": 1.3846,
492
+ "step": 1725
493
+ },
494
+ {
495
+ "epoch": 0.7025291047771979,
496
+ "grad_norm": 6.049670696258545,
497
+ "learning_rate": 4.255018587360595e-05,
498
+ "loss": 1.3116,
499
+ "step": 1750
500
+ },
501
+ {
502
+ "epoch": 0.7125652348454437,
503
+ "grad_norm": 9.660329818725586,
504
+ "learning_rate": 4.2364312267657996e-05,
505
+ "loss": 1.1697,
506
+ "step": 1775
507
+ },
508
+ {
509
+ "epoch": 0.7226013649136893,
510
+ "grad_norm": 3.5257277488708496,
511
+ "learning_rate": 4.217843866171004e-05,
512
+ "loss": 1.368,
513
+ "step": 1800
514
+ },
515
+ {
516
+ "epoch": 0.732637494981935,
517
+ "grad_norm": 5.34630823135376,
518
+ "learning_rate": 4.199256505576208e-05,
519
+ "loss": 1.3725,
520
+ "step": 1825
521
+ },
522
+ {
523
+ "epoch": 0.7426736250501806,
524
+ "grad_norm": 6.894128799438477,
525
+ "learning_rate": 4.1806691449814125e-05,
526
+ "loss": 1.1919,
527
+ "step": 1850
528
+ },
529
+ {
530
+ "epoch": 0.7527097551184263,
531
+ "grad_norm": 5.004602432250977,
532
+ "learning_rate": 4.1620817843866175e-05,
533
+ "loss": 1.2233,
534
+ "step": 1875
535
+ },
536
+ {
537
+ "epoch": 0.762745885186672,
538
+ "grad_norm": 7.174907684326172,
539
+ "learning_rate": 4.143494423791822e-05,
540
+ "loss": 1.3805,
541
+ "step": 1900
542
+ },
543
+ {
544
+ "epoch": 0.7727820152549177,
545
+ "grad_norm": 5.870824337005615,
546
+ "learning_rate": 4.124907063197026e-05,
547
+ "loss": 1.2306,
548
+ "step": 1925
549
+ },
550
+ {
551
+ "epoch": 0.7828181453231634,
552
+ "grad_norm": 13.816654205322266,
553
+ "learning_rate": 4.1063197026022304e-05,
554
+ "loss": 1.33,
555
+ "step": 1950
556
+ },
557
+ {
558
+ "epoch": 0.7928542753914091,
559
+ "grad_norm": 6.733251571655273,
560
+ "learning_rate": 4.0877323420074354e-05,
561
+ "loss": 1.2333,
562
+ "step": 1975
563
+ },
564
+ {
565
+ "epoch": 0.8028904054596547,
566
+ "grad_norm": 6.878822326660156,
567
+ "learning_rate": 4.06914498141264e-05,
568
+ "loss": 1.1058,
569
+ "step": 2000
570
+ },
571
+ {
572
+ "epoch": 0.8129265355279004,
573
+ "grad_norm": 5.60182523727417,
574
+ "learning_rate": 4.050557620817844e-05,
575
+ "loss": 1.3247,
576
+ "step": 2025
577
+ },
578
+ {
579
+ "epoch": 0.8229626655961462,
580
+ "grad_norm": 5.8014655113220215,
581
+ "learning_rate": 4.031970260223048e-05,
582
+ "loss": 1.2512,
583
+ "step": 2050
584
+ },
585
+ {
586
+ "epoch": 0.8329987956643918,
587
+ "grad_norm": 8.3912992477417,
588
+ "learning_rate": 4.013382899628253e-05,
589
+ "loss": 1.2348,
590
+ "step": 2075
591
+ },
592
+ {
593
+ "epoch": 0.8430349257326375,
594
+ "grad_norm": 4.4156270027160645,
595
+ "learning_rate": 3.9947955390334576e-05,
596
+ "loss": 1.2844,
597
+ "step": 2100
598
+ },
599
+ {
600
+ "epoch": 0.8530710558008832,
601
+ "grad_norm": 12.413702011108398,
602
+ "learning_rate": 3.976208178438662e-05,
603
+ "loss": 1.192,
604
+ "step": 2125
605
+ },
606
+ {
607
+ "epoch": 0.8631071858691288,
608
+ "grad_norm": 5.390079498291016,
609
+ "learning_rate": 3.957620817843866e-05,
610
+ "loss": 1.2788,
611
+ "step": 2150
612
+ },
613
+ {
614
+ "epoch": 0.8731433159373746,
615
+ "grad_norm": 12.286083221435547,
616
+ "learning_rate": 3.9390334572490705e-05,
617
+ "loss": 1.272,
618
+ "step": 2175
619
+ },
620
+ {
621
+ "epoch": 0.8831794460056203,
622
+ "grad_norm": 3.3497118949890137,
623
+ "learning_rate": 3.9204460966542755e-05,
624
+ "loss": 1.2199,
625
+ "step": 2200
626
+ },
627
+ {
628
+ "epoch": 0.8932155760738659,
629
+ "grad_norm": 16.91133689880371,
630
+ "learning_rate": 3.90185873605948e-05,
631
+ "loss": 1.1223,
632
+ "step": 2225
633
+ },
634
+ {
635
+ "epoch": 0.9032517061421116,
636
+ "grad_norm": 4.592910289764404,
637
+ "learning_rate": 3.883271375464684e-05,
638
+ "loss": 1.2841,
639
+ "step": 2250
640
+ },
641
+ {
642
+ "epoch": 0.9132878362103573,
643
+ "grad_norm": 7.878051280975342,
644
+ "learning_rate": 3.864684014869889e-05,
645
+ "loss": 1.1558,
646
+ "step": 2275
647
+ },
648
+ {
649
+ "epoch": 0.923323966278603,
650
+ "grad_norm": 6.925229072570801,
651
+ "learning_rate": 3.8460966542750934e-05,
652
+ "loss": 1.1348,
653
+ "step": 2300
654
+ },
655
+ {
656
+ "epoch": 0.9333600963468487,
657
+ "grad_norm": 6.3386359214782715,
658
+ "learning_rate": 3.8275092936802977e-05,
659
+ "loss": 1.2382,
660
+ "step": 2325
661
+ },
662
+ {
663
+ "epoch": 0.9433962264150944,
664
+ "grad_norm": 6.875855922698975,
665
+ "learning_rate": 3.808921933085502e-05,
666
+ "loss": 1.0859,
667
+ "step": 2350
668
+ },
669
+ {
670
+ "epoch": 0.95343235648334,
671
+ "grad_norm": 5.8908305168151855,
672
+ "learning_rate": 3.790334572490706e-05,
673
+ "loss": 1.2598,
674
+ "step": 2375
675
+ },
676
+ {
677
+ "epoch": 0.9634684865515857,
678
+ "grad_norm": 10.769807815551758,
679
+ "learning_rate": 3.7717472118959106e-05,
680
+ "loss": 1.1719,
681
+ "step": 2400
682
+ },
683
+ {
684
+ "epoch": 0.9735046166198313,
685
+ "grad_norm": 4.296354293823242,
686
+ "learning_rate": 3.7531598513011155e-05,
687
+ "loss": 1.099,
688
+ "step": 2425
689
+ },
690
+ {
691
+ "epoch": 0.9835407466880771,
692
+ "grad_norm": 6.258970260620117,
693
+ "learning_rate": 3.73457249070632e-05,
694
+ "loss": 1.2804,
695
+ "step": 2450
696
+ },
697
+ {
698
+ "epoch": 0.9935768767563228,
699
+ "grad_norm": 5.117431163787842,
700
+ "learning_rate": 3.715985130111525e-05,
701
+ "loss": 1.1818,
702
+ "step": 2475
703
+ },
704
+ {
705
+ "epoch": 1.0,
706
+ "eval_accuracy": 0.6057808109193095,
707
+ "eval_f1_macro": 0.5983023090636659,
708
+ "eval_f1_micro": 0.6057808109193095,
709
+ "eval_f1_weighted": 0.5969621030224332,
710
+ "eval_loss": 1.129994511604309,
711
+ "eval_precision_macro": 0.6126689914548499,
712
+ "eval_precision_micro": 0.6057808109193095,
713
+ "eval_precision_weighted": 0.6110410923243241,
714
+ "eval_recall_macro": 0.6066459481100611,
715
+ "eval_recall_micro": 0.6057808109193095,
716
+ "eval_recall_weighted": 0.6057808109193095,
717
+ "eval_runtime": 797.1394,
718
+ "eval_samples_per_second": 6.25,
719
+ "eval_steps_per_second": 0.391,
720
+ "step": 2491
721
+ },
722
+ {
723
+ "epoch": 1.0036130068245686,
724
+ "grad_norm": 6.499170303344727,
725
+ "learning_rate": 3.697397769516729e-05,
726
+ "loss": 1.1431,
727
+ "step": 2500
728
+ },
729
+ {
730
+ "epoch": 1.0136491368928142,
731
+ "grad_norm": 9.59432601928711,
732
+ "learning_rate": 3.6788104089219334e-05,
733
+ "loss": 1.1769,
734
+ "step": 2525
735
+ },
736
+ {
737
+ "epoch": 1.0236852669610599,
738
+ "grad_norm": 19.949962615966797,
739
+ "learning_rate": 3.660223048327138e-05,
740
+ "loss": 1.0948,
741
+ "step": 2550
742
+ },
743
+ {
744
+ "epoch": 1.0337213970293055,
745
+ "grad_norm": 14.18024730682373,
746
+ "learning_rate": 3.641635687732342e-05,
747
+ "loss": 1.2609,
748
+ "step": 2575
749
+ },
750
+ {
751
+ "epoch": 1.0437575270975512,
752
+ "grad_norm": 11.269380569458008,
753
+ "learning_rate": 3.623048327137546e-05,
754
+ "loss": 1.0704,
755
+ "step": 2600
756
+ },
757
+ {
758
+ "epoch": 1.0537936571657969,
759
+ "grad_norm": 5.695765018463135,
760
+ "learning_rate": 3.6044609665427506e-05,
761
+ "loss": 1.0063,
762
+ "step": 2625
763
+ },
764
+ {
765
+ "epoch": 1.0638297872340425,
766
+ "grad_norm": 14.094305038452148,
767
+ "learning_rate": 3.5858736059479556e-05,
768
+ "loss": 1.2925,
769
+ "step": 2650
770
+ },
771
+ {
772
+ "epoch": 1.0738659173022882,
773
+ "grad_norm": 12.211389541625977,
774
+ "learning_rate": 3.5672862453531606e-05,
775
+ "loss": 1.0322,
776
+ "step": 2675
777
+ },
778
+ {
779
+ "epoch": 1.0839020473705339,
780
+ "grad_norm": 13.703721046447754,
781
+ "learning_rate": 3.548698884758365e-05,
782
+ "loss": 1.1343,
783
+ "step": 2700
784
+ },
785
+ {
786
+ "epoch": 1.0939381774387795,
787
+ "grad_norm": 9.662201881408691,
788
+ "learning_rate": 3.530111524163569e-05,
789
+ "loss": 1.0711,
790
+ "step": 2725
791
+ },
792
+ {
793
+ "epoch": 1.1039743075070252,
794
+ "grad_norm": 5.344089508056641,
795
+ "learning_rate": 3.5115241635687735e-05,
796
+ "loss": 1.2139,
797
+ "step": 2750
798
+ },
799
+ {
800
+ "epoch": 1.114010437575271,
801
+ "grad_norm": 8.1133394241333,
802
+ "learning_rate": 3.492936802973978e-05,
803
+ "loss": 1.1404,
804
+ "step": 2775
805
+ },
806
+ {
807
+ "epoch": 1.1240465676435167,
808
+ "grad_norm": 8.59904670715332,
809
+ "learning_rate": 3.474349442379182e-05,
810
+ "loss": 1.2904,
811
+ "step": 2800
812
+ },
813
+ {
814
+ "epoch": 1.1340826977117624,
815
+ "grad_norm": 7.583441257476807,
816
+ "learning_rate": 3.4557620817843864e-05,
817
+ "loss": 1.1021,
818
+ "step": 2825
819
+ },
820
+ {
821
+ "epoch": 1.144118827780008,
822
+ "grad_norm": 4.938253879547119,
823
+ "learning_rate": 3.4371747211895914e-05,
824
+ "loss": 1.0878,
825
+ "step": 2850
826
+ },
827
+ {
828
+ "epoch": 1.1541549578482537,
829
+ "grad_norm": 9.23643684387207,
830
+ "learning_rate": 3.418587360594796e-05,
831
+ "loss": 1.1033,
832
+ "step": 2875
833
+ },
834
+ {
835
+ "epoch": 1.1641910879164994,
836
+ "grad_norm": 6.702592372894287,
837
+ "learning_rate": 3.4000000000000007e-05,
838
+ "loss": 1.0389,
839
+ "step": 2900
840
+ },
841
+ {
842
+ "epoch": 1.174227217984745,
843
+ "grad_norm": 7.392385959625244,
844
+ "learning_rate": 3.381412639405205e-05,
845
+ "loss": 1.0155,
846
+ "step": 2925
847
+ },
848
+ {
849
+ "epoch": 1.1842633480529907,
850
+ "grad_norm": 27.505746841430664,
851
+ "learning_rate": 3.362825278810409e-05,
852
+ "loss": 1.2215,
853
+ "step": 2950
854
+ },
855
+ {
856
+ "epoch": 1.1942994781212364,
857
+ "grad_norm": 14.125563621520996,
858
+ "learning_rate": 3.3442379182156136e-05,
859
+ "loss": 1.0795,
860
+ "step": 2975
861
+ },
862
+ {
863
+ "epoch": 1.204335608189482,
864
+ "grad_norm": 9.185543060302734,
865
+ "learning_rate": 3.325650557620818e-05,
866
+ "loss": 1.1799,
867
+ "step": 3000
868
+ },
869
+ {
870
+ "epoch": 1.2143717382577277,
871
+ "grad_norm": 9.01848316192627,
872
+ "learning_rate": 3.307063197026022e-05,
873
+ "loss": 1.212,
874
+ "step": 3025
875
+ },
876
+ {
877
+ "epoch": 1.2244078683259736,
878
+ "grad_norm": 9.692822456359863,
879
+ "learning_rate": 3.288475836431227e-05,
880
+ "loss": 1.0784,
881
+ "step": 3050
882
+ },
883
+ {
884
+ "epoch": 1.2344439983942193,
885
+ "grad_norm": 20.630401611328125,
886
+ "learning_rate": 3.2698884758364314e-05,
887
+ "loss": 1.2005,
888
+ "step": 3075
889
+ },
890
+ {
891
+ "epoch": 1.244480128462465,
892
+ "grad_norm": 8.891132354736328,
893
+ "learning_rate": 3.251301115241636e-05,
894
+ "loss": 1.1043,
895
+ "step": 3100
896
+ },
897
+ {
898
+ "epoch": 1.2545162585307106,
899
+ "grad_norm": 7.9135308265686035,
900
+ "learning_rate": 3.232713754646841e-05,
901
+ "loss": 1.006,
902
+ "step": 3125
903
+ },
904
+ {
905
+ "epoch": 1.2645523885989562,
906
+ "grad_norm": 7.054498195648193,
907
+ "learning_rate": 3.214126394052045e-05,
908
+ "loss": 1.0986,
909
+ "step": 3150
910
+ },
911
+ {
912
+ "epoch": 1.274588518667202,
913
+ "grad_norm": 9.929298400878906,
914
+ "learning_rate": 3.195539033457249e-05,
915
+ "loss": 1.0762,
916
+ "step": 3175
917
+ },
918
+ {
919
+ "epoch": 1.2846246487354476,
920
+ "grad_norm": 7.586366653442383,
921
+ "learning_rate": 3.1769516728624536e-05,
922
+ "loss": 0.9468,
923
+ "step": 3200
924
+ },
925
+ {
926
+ "epoch": 1.2946607788036932,
927
+ "grad_norm": 14.547595024108887,
928
+ "learning_rate": 3.158364312267658e-05,
929
+ "loss": 1.0565,
930
+ "step": 3225
931
+ },
932
+ {
933
+ "epoch": 1.304696908871939,
934
+ "grad_norm": 6.774365425109863,
935
+ "learning_rate": 3.139776951672863e-05,
936
+ "loss": 1.0749,
937
+ "step": 3250
938
+ },
939
+ {
940
+ "epoch": 1.3147330389401848,
941
+ "grad_norm": 6.633537292480469,
942
+ "learning_rate": 3.121189591078067e-05,
943
+ "loss": 0.9786,
944
+ "step": 3275
945
+ },
946
+ {
947
+ "epoch": 1.3247691690084302,
948
+ "grad_norm": 9.134368896484375,
949
+ "learning_rate": 3.1026022304832715e-05,
950
+ "loss": 1.1634,
951
+ "step": 3300
952
+ },
953
+ {
954
+ "epoch": 1.334805299076676,
955
+ "grad_norm": 5.2640228271484375,
956
+ "learning_rate": 3.084014869888476e-05,
957
+ "loss": 1.1088,
958
+ "step": 3325
959
+ },
960
+ {
961
+ "epoch": 1.3448414291449218,
962
+ "grad_norm": 12.46059513092041,
963
+ "learning_rate": 3.065427509293681e-05,
964
+ "loss": 0.9776,
965
+ "step": 3350
966
+ },
967
+ {
968
+ "epoch": 1.3548775592131674,
969
+ "grad_norm": 10.393540382385254,
970
+ "learning_rate": 3.046840148698885e-05,
971
+ "loss": 1.083,
972
+ "step": 3375
973
+ },
974
+ {
975
+ "epoch": 1.364913689281413,
976
+ "grad_norm": 9.62662124633789,
977
+ "learning_rate": 3.0282527881040894e-05,
978
+ "loss": 1.0039,
979
+ "step": 3400
980
+ },
981
+ {
982
+ "epoch": 1.3749498193496588,
983
+ "grad_norm": 6.393442630767822,
984
+ "learning_rate": 3.009665427509294e-05,
985
+ "loss": 1.0858,
986
+ "step": 3425
987
+ },
988
+ {
989
+ "epoch": 1.3849859494179044,
990
+ "grad_norm": 6.460680961608887,
991
+ "learning_rate": 2.9910780669144983e-05,
992
+ "loss": 0.8974,
993
+ "step": 3450
994
+ },
995
+ {
996
+ "epoch": 1.39502207948615,
997
+ "grad_norm": 19.161996841430664,
998
+ "learning_rate": 2.9724907063197026e-05,
999
+ "loss": 0.9708,
1000
+ "step": 3475
1001
+ },
1002
+ {
1003
+ "epoch": 1.4050582095543958,
1004
+ "grad_norm": 5.401683330535889,
1005
+ "learning_rate": 2.9539033457249073e-05,
1006
+ "loss": 0.8682,
1007
+ "step": 3500
1008
+ },
1009
+ {
1010
+ "epoch": 1.4150943396226414,
1011
+ "grad_norm": 10.808605194091797,
1012
+ "learning_rate": 2.9353159851301116e-05,
1013
+ "loss": 0.9252,
1014
+ "step": 3525
1015
+ },
1016
+ {
1017
+ "epoch": 1.4251304696908873,
1018
+ "grad_norm": 15.710016250610352,
1019
+ "learning_rate": 2.916728624535316e-05,
1020
+ "loss": 1.1788,
1021
+ "step": 3550
1022
+ },
1023
+ {
1024
+ "epoch": 1.4351665997591327,
1025
+ "grad_norm": 8.340784072875977,
1026
+ "learning_rate": 2.8981412639405202e-05,
1027
+ "loss": 1.0412,
1028
+ "step": 3575
1029
+ },
1030
+ {
1031
+ "epoch": 1.4452027298273786,
1032
+ "grad_norm": 6.99077033996582,
1033
+ "learning_rate": 2.879553903345725e-05,
1034
+ "loss": 1.0227,
1035
+ "step": 3600
1036
+ },
1037
+ {
1038
+ "epoch": 1.4552388598956243,
1039
+ "grad_norm": 6.2852959632873535,
1040
+ "learning_rate": 2.8609665427509298e-05,
1041
+ "loss": 0.8811,
1042
+ "step": 3625
1043
+ },
1044
+ {
1045
+ "epoch": 1.46527498996387,
1046
+ "grad_norm": 6.518800258636475,
1047
+ "learning_rate": 2.842379182156134e-05,
1048
+ "loss": 1.0829,
1049
+ "step": 3650
1050
+ },
1051
+ {
1052
+ "epoch": 1.4753111200321156,
1053
+ "grad_norm": 12.849925994873047,
1054
+ "learning_rate": 2.8237918215613384e-05,
1055
+ "loss": 1.1139,
1056
+ "step": 3675
1057
+ },
1058
+ {
1059
+ "epoch": 1.4853472501003613,
1060
+ "grad_norm": 8.003408432006836,
1061
+ "learning_rate": 2.805204460966543e-05,
1062
+ "loss": 1.0784,
1063
+ "step": 3700
1064
+ },
1065
+ {
1066
+ "epoch": 1.495383380168607,
1067
+ "grad_norm": 4.201417922973633,
1068
+ "learning_rate": 2.7866171003717473e-05,
1069
+ "loss": 1.1086,
1070
+ "step": 3725
1071
+ },
1072
+ {
1073
+ "epoch": 1.5054195102368526,
1074
+ "grad_norm": 10.88537311553955,
1075
+ "learning_rate": 2.7680297397769516e-05,
1076
+ "loss": 1.0667,
1077
+ "step": 3750
1078
+ },
1079
+ {
1080
+ "epoch": 1.5154556403050985,
1081
+ "grad_norm": 5.831775665283203,
1082
+ "learning_rate": 2.749442379182156e-05,
1083
+ "loss": 1.1309,
1084
+ "step": 3775
1085
+ },
1086
+ {
1087
+ "epoch": 1.525491770373344,
1088
+ "grad_norm": 15.371512413024902,
1089
+ "learning_rate": 2.7308550185873606e-05,
1090
+ "loss": 0.991,
1091
+ "step": 3800
1092
+ },
1093
+ {
1094
+ "epoch": 1.5355279004415898,
1095
+ "grad_norm": 12.852242469787598,
1096
+ "learning_rate": 2.7122676579925656e-05,
1097
+ "loss": 0.9607,
1098
+ "step": 3825
1099
+ },
1100
+ {
1101
+ "epoch": 1.5455640305098353,
1102
+ "grad_norm": 6.148252964019775,
1103
+ "learning_rate": 2.69368029739777e-05,
1104
+ "loss": 1.13,
1105
+ "step": 3850
1106
+ },
1107
+ {
1108
+ "epoch": 1.5556001605780811,
1109
+ "grad_norm": 9.267879486083984,
1110
+ "learning_rate": 2.675092936802974e-05,
1111
+ "loss": 1.0981,
1112
+ "step": 3875
1113
+ },
1114
+ {
1115
+ "epoch": 1.5656362906463268,
1116
+ "grad_norm": 3.6907107830047607,
1117
+ "learning_rate": 2.6565055762081788e-05,
1118
+ "loss": 1.123,
1119
+ "step": 3900
1120
+ },
1121
+ {
1122
+ "epoch": 1.5756724207145725,
1123
+ "grad_norm": 13.81143569946289,
1124
+ "learning_rate": 2.637918215613383e-05,
1125
+ "loss": 0.9826,
1126
+ "step": 3925
1127
+ },
1128
+ {
1129
+ "epoch": 1.5857085507828181,
1130
+ "grad_norm": 10.353007316589355,
1131
+ "learning_rate": 2.6193308550185874e-05,
1132
+ "loss": 1.0302,
1133
+ "step": 3950
1134
+ },
1135
+ {
1136
+ "epoch": 1.5957446808510638,
1137
+ "grad_norm": 6.6601457595825195,
1138
+ "learning_rate": 2.6007434944237917e-05,
1139
+ "loss": 0.9897,
1140
+ "step": 3975
1141
+ },
1142
+ {
1143
+ "epoch": 1.6057808109193095,
1144
+ "grad_norm": 10.227742195129395,
1145
+ "learning_rate": 2.5821561338289963e-05,
1146
+ "loss": 0.939,
1147
+ "step": 4000
1148
+ },
1149
+ {
1150
+ "epoch": 1.6158169409875551,
1151
+ "grad_norm": 12.598037719726562,
1152
+ "learning_rate": 2.5635687732342007e-05,
1153
+ "loss": 1.0387,
1154
+ "step": 4025
1155
+ },
1156
+ {
1157
+ "epoch": 1.625853071055801,
1158
+ "grad_norm": 10.454833030700684,
1159
+ "learning_rate": 2.5449814126394056e-05,
1160
+ "loss": 0.9812,
1161
+ "step": 4050
1162
+ },
1163
+ {
1164
+ "epoch": 1.6358892011240465,
1165
+ "grad_norm": 7.645059585571289,
1166
+ "learning_rate": 2.52639405204461e-05,
1167
+ "loss": 1.009,
1168
+ "step": 4075
1169
+ },
1170
+ {
1171
+ "epoch": 1.6459253311922923,
1172
+ "grad_norm": 6.201138973236084,
1173
+ "learning_rate": 2.5078066914498142e-05,
1174
+ "loss": 1.0306,
1175
+ "step": 4100
1176
+ },
1177
+ {
1178
+ "epoch": 1.6559614612605378,
1179
+ "grad_norm": 12.0375394821167,
1180
+ "learning_rate": 2.489219330855019e-05,
1181
+ "loss": 1.013,
1182
+ "step": 4125
1183
+ },
1184
+ {
1185
+ "epoch": 1.6659975913287837,
1186
+ "grad_norm": 8.109641075134277,
1187
+ "learning_rate": 2.4706319702602232e-05,
1188
+ "loss": 0.8755,
1189
+ "step": 4150
1190
+ },
1191
+ {
1192
+ "epoch": 1.6760337213970293,
1193
+ "grad_norm": 13.750313758850098,
1194
+ "learning_rate": 2.4520446096654275e-05,
1195
+ "loss": 0.9282,
1196
+ "step": 4175
1197
+ },
1198
+ {
1199
+ "epoch": 1.686069851465275,
1200
+ "grad_norm": 3.3968727588653564,
1201
+ "learning_rate": 2.433457249070632e-05,
1202
+ "loss": 1.0483,
1203
+ "step": 4200
1204
+ },
1205
+ {
1206
+ "epoch": 1.6961059815335207,
1207
+ "grad_norm": 8.283858299255371,
1208
+ "learning_rate": 2.4148698884758368e-05,
1209
+ "loss": 1.1133,
1210
+ "step": 4225
1211
+ },
1212
+ {
1213
+ "epoch": 1.7061421116017663,
1214
+ "grad_norm": 5.217039108276367,
1215
+ "learning_rate": 2.396282527881041e-05,
1216
+ "loss": 0.8489,
1217
+ "step": 4250
1218
+ },
1219
+ {
1220
+ "epoch": 1.7161782416700122,
1221
+ "grad_norm": 4.49629545211792,
1222
+ "learning_rate": 2.3776951672862454e-05,
1223
+ "loss": 0.743,
1224
+ "step": 4275
1225
+ },
1226
+ {
1227
+ "epoch": 1.7262143717382576,
1228
+ "grad_norm": 7.107280254364014,
1229
+ "learning_rate": 2.35910780669145e-05,
1230
+ "loss": 0.9398,
1231
+ "step": 4300
1232
+ },
1233
+ {
1234
+ "epoch": 1.7362505018065035,
1235
+ "grad_norm": 3.9583284854888916,
1236
+ "learning_rate": 2.3405204460966543e-05,
1237
+ "loss": 0.961,
1238
+ "step": 4325
1239
+ },
1240
+ {
1241
+ "epoch": 1.746286631874749,
1242
+ "grad_norm": 10.223085403442383,
1243
+ "learning_rate": 2.321933085501859e-05,
1244
+ "loss": 1.1148,
1245
+ "step": 4350
1246
+ },
1247
+ {
1248
+ "epoch": 1.7563227619429949,
1249
+ "grad_norm": 7.453252792358398,
1250
+ "learning_rate": 2.3033457249070632e-05,
1251
+ "loss": 1.0246,
1252
+ "step": 4375
1253
+ },
1254
+ {
1255
+ "epoch": 1.7663588920112403,
1256
+ "grad_norm": 4.716647148132324,
1257
+ "learning_rate": 2.284758364312268e-05,
1258
+ "loss": 0.929,
1259
+ "step": 4400
1260
+ },
1261
+ {
1262
+ "epoch": 1.7763950220794862,
1263
+ "grad_norm": 6.915979862213135,
1264
+ "learning_rate": 2.2661710037174722e-05,
1265
+ "loss": 0.9241,
1266
+ "step": 4425
1267
+ },
1268
+ {
1269
+ "epoch": 1.7864311521477318,
1270
+ "grad_norm": 15.455207824707031,
1271
+ "learning_rate": 2.2475836431226765e-05,
1272
+ "loss": 0.9634,
1273
+ "step": 4450
1274
+ },
1275
+ {
1276
+ "epoch": 1.7964672822159775,
1277
+ "grad_norm": 14.850927352905273,
1278
+ "learning_rate": 2.228996282527881e-05,
1279
+ "loss": 0.9322,
1280
+ "step": 4475
1281
+ },
1282
+ {
1283
+ "epoch": 1.8065034122842232,
1284
+ "grad_norm": 7.982198715209961,
1285
+ "learning_rate": 2.2104089219330858e-05,
1286
+ "loss": 0.9881,
1287
+ "step": 4500
1288
+ },
1289
+ {
1290
+ "epoch": 1.8165395423524688,
1291
+ "grad_norm": 5.206161022186279,
1292
+ "learning_rate": 2.19182156133829e-05,
1293
+ "loss": 1.0773,
1294
+ "step": 4525
1295
+ },
1296
+ {
1297
+ "epoch": 1.8265756724207147,
1298
+ "grad_norm": 5.201132297515869,
1299
+ "learning_rate": 2.1732342007434944e-05,
1300
+ "loss": 0.9938,
1301
+ "step": 4550
1302
+ },
1303
+ {
1304
+ "epoch": 1.8366118024889602,
1305
+ "grad_norm": 7.117099761962891,
1306
+ "learning_rate": 2.154646840148699e-05,
1307
+ "loss": 1.0281,
1308
+ "step": 4575
1309
+ },
1310
+ {
1311
+ "epoch": 1.846647932557206,
1312
+ "grad_norm": 8.35094165802002,
1313
+ "learning_rate": 2.1360594795539036e-05,
1314
+ "loss": 0.922,
1315
+ "step": 4600
1316
+ },
1317
+ {
1318
+ "epoch": 1.8566840626254515,
1319
+ "grad_norm": 8.167677879333496,
1320
+ "learning_rate": 2.117472118959108e-05,
1321
+ "loss": 0.9462,
1322
+ "step": 4625
1323
+ },
1324
+ {
1325
+ "epoch": 1.8667201926936974,
1326
+ "grad_norm": 10.520957946777344,
1327
+ "learning_rate": 2.0988847583643123e-05,
1328
+ "loss": 1.0836,
1329
+ "step": 4650
1330
+ },
1331
+ {
1332
+ "epoch": 1.876756322761943,
1333
+ "grad_norm": 34.1854248046875,
1334
+ "learning_rate": 2.0802973977695166e-05,
1335
+ "loss": 0.9133,
1336
+ "step": 4675
1337
+ },
1338
+ {
1339
+ "epoch": 1.8867924528301887,
1340
+ "grad_norm": 6.081179618835449,
1341
+ "learning_rate": 2.0617100371747215e-05,
1342
+ "loss": 0.983,
1343
+ "step": 4700
1344
+ },
1345
+ {
1346
+ "epoch": 1.8968285828984344,
1347
+ "grad_norm": 10.599027633666992,
1348
+ "learning_rate": 2.043122676579926e-05,
1349
+ "loss": 0.9998,
1350
+ "step": 4725
1351
+ },
1352
+ {
1353
+ "epoch": 1.90686471296668,
1354
+ "grad_norm": 9.57131290435791,
1355
+ "learning_rate": 2.02453531598513e-05,
1356
+ "loss": 0.9191,
1357
+ "step": 4750
1358
+ },
1359
+ {
1360
+ "epoch": 1.9169008430349257,
1361
+ "grad_norm": 6.110264778137207,
1362
+ "learning_rate": 2.0059479553903344e-05,
1363
+ "loss": 0.8917,
1364
+ "step": 4775
1365
+ },
1366
+ {
1367
+ "epoch": 1.9269369731031714,
1368
+ "grad_norm": 10.64118766784668,
1369
+ "learning_rate": 1.9873605947955394e-05,
1370
+ "loss": 0.9154,
1371
+ "step": 4800
1372
+ },
1373
+ {
1374
+ "epoch": 1.9369731031714172,
1375
+ "grad_norm": 18.218952178955078,
1376
+ "learning_rate": 1.9687732342007437e-05,
1377
+ "loss": 0.9216,
1378
+ "step": 4825
1379
+ },
1380
+ {
1381
+ "epoch": 1.9470092332396627,
1382
+ "grad_norm": 5.614959716796875,
1383
+ "learning_rate": 1.950185873605948e-05,
1384
+ "loss": 1.0401,
1385
+ "step": 4850
1386
+ },
1387
+ {
1388
+ "epoch": 1.9570453633079086,
1389
+ "grad_norm": 9.277515411376953,
1390
+ "learning_rate": 1.9315985130111523e-05,
1391
+ "loss": 1.1864,
1392
+ "step": 4875
1393
+ },
1394
+ {
1395
+ "epoch": 1.967081493376154,
1396
+ "grad_norm": 14.814881324768066,
1397
+ "learning_rate": 1.913011152416357e-05,
1398
+ "loss": 0.9891,
1399
+ "step": 4900
1400
+ },
1401
+ {
1402
+ "epoch": 1.9771176234444,
1403
+ "grad_norm": 10.074165344238281,
1404
+ "learning_rate": 1.8944237918215616e-05,
1405
+ "loss": 0.9461,
1406
+ "step": 4925
1407
+ },
1408
+ {
1409
+ "epoch": 1.9871537535126456,
1410
+ "grad_norm": 8.03749942779541,
1411
+ "learning_rate": 1.875836431226766e-05,
1412
+ "loss": 0.8771,
1413
+ "step": 4950
1414
+ },
1415
+ {
1416
+ "epoch": 1.9971898835808912,
1417
+ "grad_norm": 3.215528726577759,
1418
+ "learning_rate": 1.8572490706319702e-05,
1419
+ "loss": 0.8763,
1420
+ "step": 4975
1421
+ },
1422
+ {
1423
+ "epoch": 2.0,
1424
+ "eval_accuracy": 0.6547571256523484,
1425
+ "eval_f1_macro": 0.6500225040595868,
1426
+ "eval_f1_micro": 0.6547571256523484,
1427
+ "eval_f1_weighted": 0.6488410398009357,
1428
+ "eval_loss": 0.9818471670150757,
1429
+ "eval_precision_macro": 0.6640770883783581,
1430
+ "eval_precision_micro": 0.6547571256523484,
1431
+ "eval_precision_weighted": 0.662792396194005,
1432
+ "eval_recall_macro": 0.6557225665163219,
1433
+ "eval_recall_micro": 0.6547571256523484,
1434
+ "eval_recall_weighted": 0.6547571256523484,
1435
+ "eval_runtime": 851.1654,
1436
+ "eval_samples_per_second": 5.853,
1437
+ "eval_steps_per_second": 0.367,
1438
+ "step": 4982
1439
+ },
1440
+ {
1441
+ "epoch": 2.007226013649137,
1442
+ "grad_norm": 8.038765907287598,
1443
+ "learning_rate": 1.838661710037175e-05,
1444
+ "loss": 0.9551,
1445
+ "step": 5000
1446
+ },
1447
+ {
1448
+ "epoch": 2.0172621437173825,
1449
+ "grad_norm": 7.78813362121582,
1450
+ "learning_rate": 1.820074349442379e-05,
1451
+ "loss": 0.8925,
1452
+ "step": 5025
1453
+ },
1454
+ {
1455
+ "epoch": 2.0272982737856284,
1456
+ "grad_norm": 7.722963333129883,
1457
+ "learning_rate": 1.8014869888475838e-05,
1458
+ "loss": 0.8987,
1459
+ "step": 5050
1460
+ },
1461
+ {
1462
+ "epoch": 2.037334403853874,
1463
+ "grad_norm": 10.702392578125,
1464
+ "learning_rate": 1.782899628252788e-05,
1465
+ "loss": 0.8756,
1466
+ "step": 5075
1467
+ },
1468
+ {
1469
+ "epoch": 2.0473705339221198,
1470
+ "grad_norm": 5.403848171234131,
1471
+ "learning_rate": 1.7643122676579927e-05,
1472
+ "loss": 0.9061,
1473
+ "step": 5100
1474
+ },
1475
+ {
1476
+ "epoch": 2.057406663990365,
1477
+ "grad_norm": 10.365225791931152,
1478
+ "learning_rate": 1.745724907063197e-05,
1479
+ "loss": 0.8466,
1480
+ "step": 5125
1481
+ },
1482
+ {
1483
+ "epoch": 2.067442794058611,
1484
+ "grad_norm": 3.015634775161743,
1485
+ "learning_rate": 1.7271375464684017e-05,
1486
+ "loss": 0.7796,
1487
+ "step": 5150
1488
+ },
1489
+ {
1490
+ "epoch": 2.0774789241268565,
1491
+ "grad_norm": 11.44224739074707,
1492
+ "learning_rate": 1.708550185873606e-05,
1493
+ "loss": 0.7642,
1494
+ "step": 5175
1495
+ },
1496
+ {
1497
+ "epoch": 2.0875150541951024,
1498
+ "grad_norm": 8.612204551696777,
1499
+ "learning_rate": 1.6899628252788106e-05,
1500
+ "loss": 0.8178,
1501
+ "step": 5200
1502
+ },
1503
+ {
1504
+ "epoch": 2.097551184263348,
1505
+ "grad_norm": 13.947504997253418,
1506
+ "learning_rate": 1.671375464684015e-05,
1507
+ "loss": 0.7342,
1508
+ "step": 5225
1509
+ },
1510
+ {
1511
+ "epoch": 2.1075873143315937,
1512
+ "grad_norm": 5.7223663330078125,
1513
+ "learning_rate": 1.6527881040892192e-05,
1514
+ "loss": 1.0052,
1515
+ "step": 5250
1516
+ },
1517
+ {
1518
+ "epoch": 2.1176234443998396,
1519
+ "grad_norm": 6.79681921005249,
1520
+ "learning_rate": 1.634200743494424e-05,
1521
+ "loss": 0.858,
1522
+ "step": 5275
1523
+ },
1524
+ {
1525
+ "epoch": 2.127659574468085,
1526
+ "grad_norm": 10.765701293945312,
1527
+ "learning_rate": 1.6156133828996285e-05,
1528
+ "loss": 0.7969,
1529
+ "step": 5300
1530
+ },
1531
+ {
1532
+ "epoch": 2.137695704536331,
1533
+ "grad_norm": 5.018877029418945,
1534
+ "learning_rate": 1.5970260223048328e-05,
1535
+ "loss": 0.7923,
1536
+ "step": 5325
1537
+ },
1538
+ {
1539
+ "epoch": 2.1477318346045764,
1540
+ "grad_norm": 10.32058334350586,
1541
+ "learning_rate": 1.578438661710037e-05,
1542
+ "loss": 0.9131,
1543
+ "step": 5350
1544
+ },
1545
+ {
1546
+ "epoch": 2.1577679646728223,
1547
+ "grad_norm": 4.6106061935424805,
1548
+ "learning_rate": 1.5598513011152417e-05,
1549
+ "loss": 0.8826,
1550
+ "step": 5375
1551
+ },
1552
+ {
1553
+ "epoch": 2.1678040947410677,
1554
+ "grad_norm": 10.13034439086914,
1555
+ "learning_rate": 1.5412639405204464e-05,
1556
+ "loss": 0.8108,
1557
+ "step": 5400
1558
+ },
1559
+ {
1560
+ "epoch": 2.1778402248093136,
1561
+ "grad_norm": 10.367287635803223,
1562
+ "learning_rate": 1.5226765799256507e-05,
1563
+ "loss": 0.8552,
1564
+ "step": 5425
1565
+ },
1566
+ {
1567
+ "epoch": 2.187876354877559,
1568
+ "grad_norm": 13.695893287658691,
1569
+ "learning_rate": 1.504089219330855e-05,
1570
+ "loss": 0.8305,
1571
+ "step": 5450
1572
+ },
1573
+ {
1574
+ "epoch": 2.197912484945805,
1575
+ "grad_norm": 7.799850940704346,
1576
+ "learning_rate": 1.4855018587360594e-05,
1577
+ "loss": 0.9426,
1578
+ "step": 5475
1579
+ },
1580
+ {
1581
+ "epoch": 2.2079486150140504,
1582
+ "grad_norm": 7.746633529663086,
1583
+ "learning_rate": 1.4669144981412641e-05,
1584
+ "loss": 0.8498,
1585
+ "step": 5500
1586
+ },
1587
+ {
1588
+ "epoch": 2.2179847450822963,
1589
+ "grad_norm": 3.6280875205993652,
1590
+ "learning_rate": 1.4483271375464686e-05,
1591
+ "loss": 0.7575,
1592
+ "step": 5525
1593
+ },
1594
+ {
1595
+ "epoch": 2.228020875150542,
1596
+ "grad_norm": 12.937957763671875,
1597
+ "learning_rate": 1.4297397769516729e-05,
1598
+ "loss": 0.7432,
1599
+ "step": 5550
1600
+ },
1601
+ {
1602
+ "epoch": 2.2380570052187876,
1603
+ "grad_norm": 12.451675415039062,
1604
+ "learning_rate": 1.4111524163568773e-05,
1605
+ "loss": 0.8524,
1606
+ "step": 5575
1607
+ },
1608
+ {
1609
+ "epoch": 2.2480931352870335,
1610
+ "grad_norm": 6.163950443267822,
1611
+ "learning_rate": 1.3925650557620818e-05,
1612
+ "loss": 0.7289,
1613
+ "step": 5600
1614
+ },
1615
+ {
1616
+ "epoch": 2.258129265355279,
1617
+ "grad_norm": 13.072674751281738,
1618
+ "learning_rate": 1.3739776951672864e-05,
1619
+ "loss": 0.863,
1620
+ "step": 5625
1621
+ },
1622
+ {
1623
+ "epoch": 2.268165395423525,
1624
+ "grad_norm": 8.630057334899902,
1625
+ "learning_rate": 1.3553903345724907e-05,
1626
+ "loss": 0.8568,
1627
+ "step": 5650
1628
+ },
1629
+ {
1630
+ "epoch": 2.2782015254917702,
1631
+ "grad_norm": 4.900511741638184,
1632
+ "learning_rate": 1.3368029739776952e-05,
1633
+ "loss": 0.8374,
1634
+ "step": 5675
1635
+ },
1636
+ {
1637
+ "epoch": 2.288237655560016,
1638
+ "grad_norm": 9.796160697937012,
1639
+ "learning_rate": 1.3182156133828997e-05,
1640
+ "loss": 0.7818,
1641
+ "step": 5700
1642
+ },
1643
+ {
1644
+ "epoch": 2.2982737856282616,
1645
+ "grad_norm": 7.636236667633057,
1646
+ "learning_rate": 1.299628252788104e-05,
1647
+ "loss": 0.9105,
1648
+ "step": 5725
1649
+ },
1650
+ {
1651
+ "epoch": 2.3083099156965075,
1652
+ "grad_norm": 9.415719032287598,
1653
+ "learning_rate": 1.2810408921933086e-05,
1654
+ "loss": 0.8107,
1655
+ "step": 5750
1656
+ },
1657
+ {
1658
+ "epoch": 2.318346045764753,
1659
+ "grad_norm": 8.357912063598633,
1660
+ "learning_rate": 1.2624535315985131e-05,
1661
+ "loss": 0.7744,
1662
+ "step": 5775
1663
+ },
1664
+ {
1665
+ "epoch": 2.328382175832999,
1666
+ "grad_norm": 8.900328636169434,
1667
+ "learning_rate": 1.2438661710037176e-05,
1668
+ "loss": 0.8166,
1669
+ "step": 5800
1670
+ },
1671
+ {
1672
+ "epoch": 2.3384183059012447,
1673
+ "grad_norm": 9.10429859161377,
1674
+ "learning_rate": 1.225278810408922e-05,
1675
+ "loss": 0.8207,
1676
+ "step": 5825
1677
+ },
1678
+ {
1679
+ "epoch": 2.34845443596949,
1680
+ "grad_norm": 9.185796737670898,
1681
+ "learning_rate": 1.2066914498141265e-05,
1682
+ "loss": 0.8174,
1683
+ "step": 5850
1684
+ },
1685
+ {
1686
+ "epoch": 2.358490566037736,
1687
+ "grad_norm": 8.934030532836914,
1688
+ "learning_rate": 1.188104089219331e-05,
1689
+ "loss": 0.8926,
1690
+ "step": 5875
1691
+ },
1692
+ {
1693
+ "epoch": 2.3685266961059814,
1694
+ "grad_norm": 7.495822906494141,
1695
+ "learning_rate": 1.1695167286245355e-05,
1696
+ "loss": 0.7944,
1697
+ "step": 5900
1698
+ },
1699
+ {
1700
+ "epoch": 2.3785628261742273,
1701
+ "grad_norm": 4.582374095916748,
1702
+ "learning_rate": 1.1509293680297398e-05,
1703
+ "loss": 0.8671,
1704
+ "step": 5925
1705
+ },
1706
+ {
1707
+ "epoch": 2.3885989562424728,
1708
+ "grad_norm": 12.170183181762695,
1709
+ "learning_rate": 1.1323420074349444e-05,
1710
+ "loss": 0.9016,
1711
+ "step": 5950
1712
+ },
1713
+ {
1714
+ "epoch": 2.3986350863107186,
1715
+ "grad_norm": 3.8928985595703125,
1716
+ "learning_rate": 1.1137546468401487e-05,
1717
+ "loss": 0.8334,
1718
+ "step": 5975
1719
+ },
1720
+ {
1721
+ "epoch": 2.408671216378964,
1722
+ "grad_norm": 11.172091484069824,
1723
+ "learning_rate": 1.0951672862453533e-05,
1724
+ "loss": 0.7836,
1725
+ "step": 6000
1726
+ },
1727
+ {
1728
+ "epoch": 2.41870734644721,
1729
+ "grad_norm": 6.816530227661133,
1730
+ "learning_rate": 1.0765799256505576e-05,
1731
+ "loss": 0.8845,
1732
+ "step": 6025
1733
+ },
1734
+ {
1735
+ "epoch": 2.4287434765154554,
1736
+ "grad_norm": 7.811573505401611,
1737
+ "learning_rate": 1.0579925650557623e-05,
1738
+ "loss": 0.8348,
1739
+ "step": 6050
1740
+ },
1741
+ {
1742
+ "epoch": 2.4387796065837013,
1743
+ "grad_norm": 11.93806266784668,
1744
+ "learning_rate": 1.0394052044609666e-05,
1745
+ "loss": 0.8122,
1746
+ "step": 6075
1747
+ },
1748
+ {
1749
+ "epoch": 2.448815736651947,
1750
+ "grad_norm": 2.4777565002441406,
1751
+ "learning_rate": 1.020817843866171e-05,
1752
+ "loss": 0.8031,
1753
+ "step": 6100
1754
+ },
1755
+ {
1756
+ "epoch": 2.4588518667201926,
1757
+ "grad_norm": 15.819439888000488,
1758
+ "learning_rate": 1.0022304832713755e-05,
1759
+ "loss": 0.7895,
1760
+ "step": 6125
1761
+ },
1762
+ {
1763
+ "epoch": 2.4688879967884385,
1764
+ "grad_norm": 6.035953998565674,
1765
+ "learning_rate": 9.8364312267658e-06,
1766
+ "loss": 0.8489,
1767
+ "step": 6150
1768
+ },
1769
+ {
1770
+ "epoch": 2.478924126856684,
1771
+ "grad_norm": 4.4093427658081055,
1772
+ "learning_rate": 9.650557620817845e-06,
1773
+ "loss": 0.7383,
1774
+ "step": 6175
1775
+ },
1776
+ {
1777
+ "epoch": 2.48896025692493,
1778
+ "grad_norm": 18.50689125061035,
1779
+ "learning_rate": 9.46468401486989e-06,
1780
+ "loss": 0.8581,
1781
+ "step": 6200
1782
+ },
1783
+ {
1784
+ "epoch": 2.4989963869931753,
1785
+ "grad_norm": 13.305225372314453,
1786
+ "learning_rate": 9.278810408921934e-06,
1787
+ "loss": 0.6993,
1788
+ "step": 6225
1789
+ },
1790
+ {
1791
+ "epoch": 2.509032517061421,
1792
+ "grad_norm": 11.539031028747559,
1793
+ "learning_rate": 9.092936802973979e-06,
1794
+ "loss": 0.8224,
1795
+ "step": 6250
1796
+ },
1797
+ {
1798
+ "epoch": 2.519068647129667,
1799
+ "grad_norm": 12.486981391906738,
1800
+ "learning_rate": 8.907063197026022e-06,
1801
+ "loss": 0.7293,
1802
+ "step": 6275
1803
+ },
1804
+ {
1805
+ "epoch": 2.5291047771979125,
1806
+ "grad_norm": 9.520222663879395,
1807
+ "learning_rate": 8.721189591078068e-06,
1808
+ "loss": 0.7651,
1809
+ "step": 6300
1810
+ },
1811
+ {
1812
+ "epoch": 2.539140907266158,
1813
+ "grad_norm": 8.691309928894043,
1814
+ "learning_rate": 8.535315985130111e-06,
1815
+ "loss": 0.7133,
1816
+ "step": 6325
1817
+ },
1818
+ {
1819
+ "epoch": 2.549177037334404,
1820
+ "grad_norm": 10.589598655700684,
1821
+ "learning_rate": 8.349442379182158e-06,
1822
+ "loss": 0.7406,
1823
+ "step": 6350
1824
+ },
1825
+ {
1826
+ "epoch": 2.5592131674026497,
1827
+ "grad_norm": 9.358379364013672,
1828
+ "learning_rate": 8.1635687732342e-06,
1829
+ "loss": 0.7483,
1830
+ "step": 6375
1831
+ },
1832
+ {
1833
+ "epoch": 2.569249297470895,
1834
+ "grad_norm": 4.712927341461182,
1835
+ "learning_rate": 7.977695167286247e-06,
1836
+ "loss": 0.7921,
1837
+ "step": 6400
1838
+ },
1839
+ {
1840
+ "epoch": 2.579285427539141,
1841
+ "grad_norm": 10.725476264953613,
1842
+ "learning_rate": 7.79182156133829e-06,
1843
+ "loss": 0.7842,
1844
+ "step": 6425
1845
+ },
1846
+ {
1847
+ "epoch": 2.5893215576073865,
1848
+ "grad_norm": 5.546850204467773,
1849
+ "learning_rate": 7.605947955390335e-06,
1850
+ "loss": 0.797,
1851
+ "step": 6450
1852
+ },
1853
+ {
1854
+ "epoch": 2.5993576876756324,
1855
+ "grad_norm": 9.5972261428833,
1856
+ "learning_rate": 7.420074349442379e-06,
1857
+ "loss": 0.8795,
1858
+ "step": 6475
1859
+ },
1860
+ {
1861
+ "epoch": 2.609393817743878,
1862
+ "grad_norm": 18.99161720275879,
1863
+ "learning_rate": 7.234200743494423e-06,
1864
+ "loss": 0.8776,
1865
+ "step": 6500
1866
+ },
1867
+ {
1868
+ "epoch": 2.6194299478121237,
1869
+ "grad_norm": 12.693746566772461,
1870
+ "learning_rate": 7.048327137546469e-06,
1871
+ "loss": 0.921,
1872
+ "step": 6525
1873
+ },
1874
+ {
1875
+ "epoch": 2.6294660778803696,
1876
+ "grad_norm": 7.012732028961182,
1877
+ "learning_rate": 6.862453531598513e-06,
1878
+ "loss": 0.9028,
1879
+ "step": 6550
1880
+ },
1881
+ {
1882
+ "epoch": 2.639502207948615,
1883
+ "grad_norm": 3.836461305618286,
1884
+ "learning_rate": 6.676579925650558e-06,
1885
+ "loss": 0.7473,
1886
+ "step": 6575
1887
+ },
1888
+ {
1889
+ "epoch": 2.6495383380168605,
1890
+ "grad_norm": 5.454137802124023,
1891
+ "learning_rate": 6.490706319702602e-06,
1892
+ "loss": 0.8219,
1893
+ "step": 6600
1894
+ },
1895
+ {
1896
+ "epoch": 2.6595744680851063,
1897
+ "grad_norm": 9.578851699829102,
1898
+ "learning_rate": 6.304832713754647e-06,
1899
+ "loss": 0.7777,
1900
+ "step": 6625
1901
+ },
1902
+ {
1903
+ "epoch": 2.669610598153352,
1904
+ "grad_norm": 8.398209571838379,
1905
+ "learning_rate": 6.1189591078066915e-06,
1906
+ "loss": 0.8382,
1907
+ "step": 6650
1908
+ },
1909
+ {
1910
+ "epoch": 2.6796467282215977,
1911
+ "grad_norm": 7.7773566246032715,
1912
+ "learning_rate": 5.933085501858736e-06,
1913
+ "loss": 0.8239,
1914
+ "step": 6675
1915
+ },
1916
+ {
1917
+ "epoch": 2.6896828582898435,
1918
+ "grad_norm": 17.32167625427246,
1919
+ "learning_rate": 5.747211895910781e-06,
1920
+ "loss": 0.8283,
1921
+ "step": 6700
1922
+ },
1923
+ {
1924
+ "epoch": 2.699718988358089,
1925
+ "grad_norm": 9.905217170715332,
1926
+ "learning_rate": 5.561338289962826e-06,
1927
+ "loss": 0.7463,
1928
+ "step": 6725
1929
+ },
1930
+ {
1931
+ "epoch": 2.709755118426335,
1932
+ "grad_norm": 8.677200317382812,
1933
+ "learning_rate": 5.37546468401487e-06,
1934
+ "loss": 0.658,
1935
+ "step": 6750
1936
+ },
1937
+ {
1938
+ "epoch": 2.7197912484945803,
1939
+ "grad_norm": 12.012971878051758,
1940
+ "learning_rate": 5.189591078066915e-06,
1941
+ "loss": 0.7865,
1942
+ "step": 6775
1943
+ },
1944
+ {
1945
+ "epoch": 2.729827378562826,
1946
+ "grad_norm": 13.909863471984863,
1947
+ "learning_rate": 5.00371747211896e-06,
1948
+ "loss": 0.863,
1949
+ "step": 6800
1950
+ },
1951
+ {
1952
+ "epoch": 2.739863508631072,
1953
+ "grad_norm": 2.4488868713378906,
1954
+ "learning_rate": 4.817843866171004e-06,
1955
+ "loss": 0.7535,
1956
+ "step": 6825
1957
+ },
1958
+ {
1959
+ "epoch": 2.7498996386993175,
1960
+ "grad_norm": 12.209616661071777,
1961
+ "learning_rate": 4.631970260223048e-06,
1962
+ "loss": 0.8308,
1963
+ "step": 6850
1964
+ },
1965
+ {
1966
+ "epoch": 2.759935768767563,
1967
+ "grad_norm": 10.368229866027832,
1968
+ "learning_rate": 4.446096654275093e-06,
1969
+ "loss": 0.8772,
1970
+ "step": 6875
1971
+ },
1972
+ {
1973
+ "epoch": 2.769971898835809,
1974
+ "grad_norm": 10.47776985168457,
1975
+ "learning_rate": 4.260223048327138e-06,
1976
+ "loss": 0.7404,
1977
+ "step": 6900
1978
+ },
1979
+ {
1980
+ "epoch": 2.7800080289040547,
1981
+ "grad_norm": 7.025723934173584,
1982
+ "learning_rate": 4.0743494423791824e-06,
1983
+ "loss": 0.7694,
1984
+ "step": 6925
1985
+ },
1986
+ {
1987
+ "epoch": 2.7900441589723,
1988
+ "grad_norm": 16.20415687561035,
1989
+ "learning_rate": 3.888475836431227e-06,
1990
+ "loss": 0.8484,
1991
+ "step": 6950
1992
+ },
1993
+ {
1994
+ "epoch": 2.800080289040546,
1995
+ "grad_norm": 7.123861312866211,
1996
+ "learning_rate": 3.7026022304832714e-06,
1997
+ "loss": 0.778,
1998
+ "step": 6975
1999
+ },
2000
+ {
2001
+ "epoch": 2.8101164191087915,
2002
+ "grad_norm": 14.885129928588867,
2003
+ "learning_rate": 3.516728624535316e-06,
2004
+ "loss": 0.8011,
2005
+ "step": 7000
2006
+ },
2007
+ {
2008
+ "epoch": 2.8201525491770374,
2009
+ "grad_norm": 5.704489231109619,
2010
+ "learning_rate": 3.330855018587361e-06,
2011
+ "loss": 0.8135,
2012
+ "step": 7025
2013
+ },
2014
+ {
2015
+ "epoch": 2.830188679245283,
2016
+ "grad_norm": 15.538228988647461,
2017
+ "learning_rate": 3.1449814126394056e-06,
2018
+ "loss": 0.7954,
2019
+ "step": 7050
2020
+ },
2021
+ {
2022
+ "epoch": 2.8402248093135287,
2023
+ "grad_norm": 13.824156761169434,
2024
+ "learning_rate": 2.95910780669145e-06,
2025
+ "loss": 0.8175,
2026
+ "step": 7075
2027
+ },
2028
+ {
2029
+ "epoch": 2.8502609393817746,
2030
+ "grad_norm": 12.763619422912598,
2031
+ "learning_rate": 2.7732342007434946e-06,
2032
+ "loss": 0.9212,
2033
+ "step": 7100
2034
+ },
2035
+ {
2036
+ "epoch": 2.86029706945002,
2037
+ "grad_norm": 8.71843433380127,
2038
+ "learning_rate": 2.587360594795539e-06,
2039
+ "loss": 0.8323,
2040
+ "step": 7125
2041
+ },
2042
+ {
2043
+ "epoch": 2.8703331995182655,
2044
+ "grad_norm": 8.395557403564453,
2045
+ "learning_rate": 2.4014869888475835e-06,
2046
+ "loss": 0.7226,
2047
+ "step": 7150
2048
+ },
2049
+ {
2050
+ "epoch": 2.8803693295865114,
2051
+ "grad_norm": 2.9431092739105225,
2052
+ "learning_rate": 2.2156133828996283e-06,
2053
+ "loss": 0.7787,
2054
+ "step": 7175
2055
+ },
2056
+ {
2057
+ "epoch": 2.8904054596547573,
2058
+ "grad_norm": 6.241610527038574,
2059
+ "learning_rate": 2.029739776951673e-06,
2060
+ "loss": 0.7049,
2061
+ "step": 7200
2062
+ },
2063
+ {
2064
+ "epoch": 2.9004415897230027,
2065
+ "grad_norm": 8.485464096069336,
2066
+ "learning_rate": 1.8438661710037175e-06,
2067
+ "loss": 0.7828,
2068
+ "step": 7225
2069
+ },
2070
+ {
2071
+ "epoch": 2.9104777197912486,
2072
+ "grad_norm": 12.730948448181152,
2073
+ "learning_rate": 1.6579925650557622e-06,
2074
+ "loss": 0.7128,
2075
+ "step": 7250
2076
+ },
2077
+ {
2078
+ "epoch": 2.920513849859494,
2079
+ "grad_norm": 9.1871337890625,
2080
+ "learning_rate": 1.4721189591078069e-06,
2081
+ "loss": 0.8205,
2082
+ "step": 7275
2083
+ },
2084
+ {
2085
+ "epoch": 2.93054997992774,
2086
+ "grad_norm": 8.444822311401367,
2087
+ "learning_rate": 1.2862453531598514e-06,
2088
+ "loss": 0.7645,
2089
+ "step": 7300
2090
+ },
2091
+ {
2092
+ "epoch": 2.9405861099959854,
2093
+ "grad_norm": 8.305177688598633,
2094
+ "learning_rate": 1.100371747211896e-06,
2095
+ "loss": 0.6978,
2096
+ "step": 7325
2097
+ },
2098
+ {
2099
+ "epoch": 2.9506222400642312,
2100
+ "grad_norm": 8.048552513122559,
2101
+ "learning_rate": 9.144981412639406e-07,
2102
+ "loss": 0.7673,
2103
+ "step": 7350
2104
+ },
2105
+ {
2106
+ "epoch": 2.960658370132477,
2107
+ "grad_norm": 7.963881492614746,
2108
+ "learning_rate": 7.286245353159852e-07,
2109
+ "loss": 0.6996,
2110
+ "step": 7375
2111
+ },
2112
+ {
2113
+ "epoch": 2.9706945002007226,
2114
+ "grad_norm": 3.0167317390441895,
2115
+ "learning_rate": 5.427509293680298e-07,
2116
+ "loss": 0.7399,
2117
+ "step": 7400
2118
+ },
2119
+ {
2120
+ "epoch": 2.9807306302689685,
2121
+ "grad_norm": 10.247928619384766,
2122
+ "learning_rate": 3.568773234200744e-07,
2123
+ "loss": 0.7349,
2124
+ "step": 7425
2125
+ },
2126
+ {
2127
+ "epoch": 2.990766760337214,
2128
+ "grad_norm": 6.888391494750977,
2129
+ "learning_rate": 1.7100371747211895e-07,
2130
+ "loss": 0.8048,
2131
+ "step": 7450
2132
+ },
2133
+ {
2134
+ "epoch": 3.0,
2135
+ "eval_accuracy": 0.6850662384584504,
2136
+ "eval_f1_macro": 0.6845343178671808,
2137
+ "eval_f1_micro": 0.6850662384584504,
2138
+ "eval_f1_weighted": 0.6834454222699836,
2139
+ "eval_loss": 0.8851153254508972,
2140
+ "eval_precision_macro": 0.6899698563739772,
2141
+ "eval_precision_micro": 0.6850662384584504,
2142
+ "eval_precision_weighted": 0.6888476966796808,
2143
+ "eval_recall_macro": 0.6860535953088197,
2144
+ "eval_recall_micro": 0.6850662384584504,
2145
+ "eval_recall_weighted": 0.6850662384584504,
2146
+ "eval_runtime": 814.4881,
2147
+ "eval_samples_per_second": 6.117,
2148
+ "eval_steps_per_second": 0.383,
2149
+ "step": 7473
2150
+ }
2151
+ ],
2152
+ "logging_steps": 25,
2153
+ "max_steps": 7473,
2154
+ "num_input_tokens_seen": 0,
2155
+ "num_train_epochs": 3,
2156
+ "save_steps": 500,
2157
+ "stateful_callbacks": {
2158
+ "EarlyStoppingCallback": {
2159
+ "args": {
2160
+ "early_stopping_patience": 5,
2161
+ "early_stopping_threshold": 0.01
2162
+ },
2163
+ "attributes": {
2164
+ "early_stopping_patience_counter": 0
2165
+ }
2166
+ },
2167
+ "TrainerControl": {
2168
+ "args": {
2169
+ "should_epoch_stop": false,
2170
+ "should_evaluate": false,
2171
+ "should_log": false,
2172
+ "should_save": true,
2173
+ "should_training_stop": true
2174
+ },
2175
+ "attributes": {}
2176
+ }
2177
+ },
2178
+ "total_flos": 3932218859904000.0,
2179
+ "train_batch_size": 8,
2180
+ "trial_name": null,
2181
+ "trial_params": null
2182
+ }
checkpoint-7473/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:aa8e0b2fe76e94cceaa03b7f4e4ef2c0f0fff2c9b1dcb64d4179cfaeca5eed8d
3
+ size 5304
config.json ADDED
@@ -0,0 +1,54 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "sentence-transformers/all-mpnet-base-v2",
3
+ "_num_labels": 12,
4
+ "architectures": [
5
+ "MPNetForSequenceClassification"
6
+ ],
7
+ "attention_probs_dropout_prob": 0.1,
8
+ "bos_token_id": 0,
9
+ "eos_token_id": 2,
10
+ "hidden_act": "gelu",
11
+ "hidden_dropout_prob": 0.1,
12
+ "hidden_size": 768,
13
+ "id2label": {
14
+ "0": "Ph\u00e2n Li\u1ec7t C\u1ea3m X\u00fac - Nh\u1eb9",
15
+ "1": "Ph\u00e2n Li\u1ec7t C\u1ea3m X\u00fac - N\u1eb7ng",
16
+ "2": "Ph\u00e2n Li\u1ec7t C\u1ea3m X\u00fac - Tr\u1ea7m Tr\u1ecdng",
17
+ "3": "R\u1ed1i Lo\u1ea1n Lo \u00c2u - Nh\u1eb9",
18
+ "4": "R\u1ed1i Lo\u1ea1n Lo \u00c2u - N\u1eb7ng",
19
+ "5": "R\u1ed1i Lo\u1ea1n Lo \u00c2u - Tr\u1ea7m Tr\u1ecdng",
20
+ "6": "Tr\u1ea7m C\u1ea3m - Nh\u1eb9",
21
+ "7": "Tr\u1ea7m C\u1ea3m - N\u1eb7ng",
22
+ "8": "Tr\u1ea7m C\u1ea3m - Tr\u1ea7m Tr\u1ecdng",
23
+ "9": "T\u1ef1 K\u1ef7 - Nh\u1eb9",
24
+ "10": "T\u1ef1 K\u1ef7 - N\u1eb7ng",
25
+ "11": "T\u1ef1 K\u1ef7 - Tr\u1ea7m Tr\u1ecdng"
26
+ },
27
+ "initializer_range": 0.02,
28
+ "intermediate_size": 3072,
29
+ "label2id": {
30
+ "Ph\u00e2n Li\u1ec7t C\u1ea3m X\u00fac - Nh\u1eb9": 0,
31
+ "Ph\u00e2n Li\u1ec7t C\u1ea3m X\u00fac - N\u1eb7ng": 1,
32
+ "Ph\u00e2n Li\u1ec7t C\u1ea3m X\u00fac - Tr\u1ea7m Tr\u1ecdng": 2,
33
+ "R\u1ed1i Lo\u1ea1n Lo \u00c2u - Nh\u1eb9": 3,
34
+ "R\u1ed1i Lo\u1ea1n Lo \u00c2u - N\u1eb7ng": 4,
35
+ "R\u1ed1i Lo\u1ea1n Lo \u00c2u - Tr\u1ea7m Tr\u1ecdng": 5,
36
+ "Tr\u1ea7m C\u1ea3m - Nh\u1eb9": 6,
37
+ "Tr\u1ea7m C\u1ea3m - N\u1eb7ng": 7,
38
+ "Tr\u1ea7m C\u1ea3m - Tr\u1ea7m Tr\u1ecdng": 8,
39
+ "T\u1ef1 K\u1ef7 - Nh\u1eb9": 9,
40
+ "T\u1ef1 K\u1ef7 - N\u1eb7ng": 10,
41
+ "T\u1ef1 K\u1ef7 - Tr\u1ea7m Tr\u1ecdng": 11
42
+ },
43
+ "layer_norm_eps": 1e-05,
44
+ "max_position_embeddings": 514,
45
+ "model_type": "mpnet",
46
+ "num_attention_heads": 12,
47
+ "num_hidden_layers": 12,
48
+ "pad_token_id": 1,
49
+ "problem_type": "single_label_classification",
50
+ "relative_attention_num_buckets": 32,
51
+ "torch_dtype": "float32",
52
+ "transformers_version": "4.46.2",
53
+ "vocab_size": 30527
54
+ }
model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d6f8750d7d265db701a13545b5afb36dc72592e2dc0a435717b0d2a8e7af1eb5
3
+ size 438005960
runs/Nov17_06-15-53_r-hoangcung165-test-train-gwo6rh86-54315-0a1vk/events.out.tfevents.1731824155.r-hoangcung165-test-train-gwo6rh86-54315-0a1vk.2400.0 CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:6b174e5047a27e03fb3102ec99ecbf60f932ab67e635e7c8728d38d2b97c5217
3
- size 70234
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:2262c7ffadfe912f8ea81ec15c74eb8bbe0db54f16bf8ae0d365106e1e83168c
3
+ size 72054
runs/Nov17_06-15-53_r-hoangcung165-test-train-gwo6rh86-54315-0a1vk/events.out.tfevents.1731863830.r-hoangcung165-test-train-gwo6rh86-54315-0a1vk.2400.1 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7cb4fe149a88638a344d739614655cbaec27b511b7f1787e1da8863fd0f01b2d
3
+ size 921
special_tokens_map.json ADDED
@@ -0,0 +1,51 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "bos_token": {
3
+ "content": "<s>",
4
+ "lstrip": false,
5
+ "normalized": false,
6
+ "rstrip": false,
7
+ "single_word": false
8
+ },
9
+ "cls_token": {
10
+ "content": "<s>",
11
+ "lstrip": false,
12
+ "normalized": false,
13
+ "rstrip": false,
14
+ "single_word": false
15
+ },
16
+ "eos_token": {
17
+ "content": "</s>",
18
+ "lstrip": false,
19
+ "normalized": false,
20
+ "rstrip": false,
21
+ "single_word": false
22
+ },
23
+ "mask_token": {
24
+ "content": "<mask>",
25
+ "lstrip": true,
26
+ "normalized": false,
27
+ "rstrip": false,
28
+ "single_word": false
29
+ },
30
+ "pad_token": {
31
+ "content": "<pad>",
32
+ "lstrip": false,
33
+ "normalized": false,
34
+ "rstrip": false,
35
+ "single_word": false
36
+ },
37
+ "sep_token": {
38
+ "content": "</s>",
39
+ "lstrip": false,
40
+ "normalized": false,
41
+ "rstrip": false,
42
+ "single_word": false
43
+ },
44
+ "unk_token": {
45
+ "content": "[UNK]",
46
+ "lstrip": false,
47
+ "normalized": false,
48
+ "rstrip": false,
49
+ "single_word": false
50
+ }
51
+ }
tokenizer.json ADDED
The diff for this file is too large to render. See raw diff
 
tokenizer_config.json ADDED
@@ -0,0 +1,72 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "added_tokens_decoder": {
3
+ "0": {
4
+ "content": "<s>",
5
+ "lstrip": false,
6
+ "normalized": false,
7
+ "rstrip": false,
8
+ "single_word": false,
9
+ "special": true
10
+ },
11
+ "1": {
12
+ "content": "<pad>",
13
+ "lstrip": false,
14
+ "normalized": false,
15
+ "rstrip": false,
16
+ "single_word": false,
17
+ "special": true
18
+ },
19
+ "2": {
20
+ "content": "</s>",
21
+ "lstrip": false,
22
+ "normalized": false,
23
+ "rstrip": false,
24
+ "single_word": false,
25
+ "special": true
26
+ },
27
+ "3": {
28
+ "content": "<unk>",
29
+ "lstrip": false,
30
+ "normalized": true,
31
+ "rstrip": false,
32
+ "single_word": false,
33
+ "special": true
34
+ },
35
+ "104": {
36
+ "content": "[UNK]",
37
+ "lstrip": false,
38
+ "normalized": false,
39
+ "rstrip": false,
40
+ "single_word": false,
41
+ "special": true
42
+ },
43
+ "30526": {
44
+ "content": "<mask>",
45
+ "lstrip": true,
46
+ "normalized": false,
47
+ "rstrip": false,
48
+ "single_word": false,
49
+ "special": true
50
+ }
51
+ },
52
+ "bos_token": "<s>",
53
+ "clean_up_tokenization_spaces": false,
54
+ "cls_token": "<s>",
55
+ "do_lower_case": true,
56
+ "eos_token": "</s>",
57
+ "mask_token": "<mask>",
58
+ "max_length": 128,
59
+ "model_max_length": 512,
60
+ "pad_to_multiple_of": null,
61
+ "pad_token": "<pad>",
62
+ "pad_token_type_id": 0,
63
+ "padding_side": "right",
64
+ "sep_token": "</s>",
65
+ "stride": 0,
66
+ "strip_accents": null,
67
+ "tokenize_chinese_chars": true,
68
+ "tokenizer_class": "MPNetTokenizer",
69
+ "truncation_side": "right",
70
+ "truncation_strategy": "longest_first",
71
+ "unk_token": "[UNK]"
72
+ }
training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:aa8e0b2fe76e94cceaa03b7f4e4ef2c0f0fff2c9b1dcb64d4179cfaeca5eed8d
3
+ size 5304
training_params.json ADDED
@@ -0,0 +1,30 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "data_path": "scanning-model-bert/autotrain-data",
3
+ "model": "sentence-transformers/all-mpnet-base-v2",
4
+ "lr": 5e-05,
5
+ "epochs": 3,
6
+ "max_seq_length": 128,
7
+ "batch_size": 8,
8
+ "warmup_ratio": 0.1,
9
+ "gradient_accumulation": 1,
10
+ "optimizer": "adamw_torch",
11
+ "scheduler": "linear",
12
+ "weight_decay": 0.0,
13
+ "max_grad_norm": 1.0,
14
+ "seed": 42,
15
+ "train_split": "train",
16
+ "valid_split": "validation",
17
+ "text_column": "autotrain_text",
18
+ "target_column": "autotrain_label",
19
+ "logging_steps": -1,
20
+ "project_name": "scanning-model-bert",
21
+ "auto_find_batch_size": true,
22
+ "mixed_precision": "fp16",
23
+ "save_total_limit": 1,
24
+ "push_to_hub": true,
25
+ "eval_strategy": "epoch",
26
+ "username": "hoangcung165",
27
+ "log": "tensorboard",
28
+ "early_stopping_patience": 5,
29
+ "early_stopping_threshold": 0.01
30
+ }
vocab.txt ADDED
The diff for this file is too large to render. See raw diff