Hugging Face's logo

ivanovsdesign
/
rubert-base-ru-huawei-sentiment-fine-up

Text Classification
TensorBoard
Safetensors
bert
Trained with AutoTrain
Model card Files Metrics Community
ivanovsdesign commited on
Commit
fdcd368
•
1 Parent(s): 0be7ef3

Upload folder using huggingface_hub

Browse files
Files changed (19) hide show
  1. README.md +36 -0
  2. checkpoint-7302/config.json +49 -0
  3. checkpoint-7302/model.safetensors +3 -0
  4. checkpoint-7302/optimizer.pt +3 -0
  5. checkpoint-7302/rng_state_0.pth +3 -0
  6. checkpoint-7302/rng_state_1.pth +3 -0
  7. checkpoint-7302/scheduler.pt +3 -0
  8. checkpoint-7302/trainer_state.json +2140 -0
  9. checkpoint-7302/training_args.bin +3 -0
  10. config.json +49 -0
  11. model.safetensors +3 -0
  12. runs/Nov09_23-13-04_6604152ce143/events.out.tfevents.1731193985.6604152ce143.674.0 +2 -2
  13. runs/Nov09_23-13-04_6604152ce143/events.out.tfevents.1731197419.6604152ce143.674.1 +3 -0
  14. special_tokens_map.json +37 -0
  15. tokenizer.json +0 -0
  16. tokenizer_config.json +64 -0
  17. training_args.bin +3 -0
  18. training_params.json +30 -0
  19. vocab.txt +0 -0
README.md ADDED
@@ -0,0 +1,36 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+
2
+ ---
3
+ tags:
4
+ - autotrain
5
+ - text-classification
6
+ base_model: ivanovsdesign/huawei-data-classification
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: 0.9126833081245422
17
+
18
+ f1_macro: 0.47986378330102575
19
+
20
+ f1_micro: 0.6351587383129559
21
+
22
+ f1_weighted: 0.6366218608153963
23
+
24
+ precision_macro: 0.4869491643480056
25
+
26
+ precision_micro: 0.6351587383129559
27
+
28
+ precision_weighted: 0.6494956994259575
29
+
30
+ recall_macro: 0.4963958102076925
31
+
32
+ recall_micro: 0.6351587383129559
33
+
34
+ recall_weighted: 0.6351587383129559
35
+
36
+ accuracy: 0.6351587383129559
checkpoint-7302/config.json ADDED
@@ -0,0 +1,49 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "ivanovsdesign/huawei-data-classification",
3
+ "_num_labels": 5,
4
+ "architectures": [
5
+ "BertForSequenceClassification"
6
+ ],
7
+ "attention_probs_dropout_prob": 0.1,
8
+ "classifier_dropout": null,
9
+ "directionality": "bidi",
10
+ "gradient_checkpointing": false,
11
+ "hidden_act": "gelu",
12
+ "hidden_dropout_prob": 0.1,
13
+ "hidden_size": 768,
14
+ "id2label": {
15
+ "0": "1",
16
+ "1": "2",
17
+ "2": "3",
18
+ "3": "4",
19
+ "4": "5"
20
+ },
21
+ "initializer_range": 0.02,
22
+ "intermediate_size": 3072,
23
+ "label2id": {
24
+ "1": 0,
25
+ "2": 1,
26
+ "3": 2,
27
+ "4": 3,
28
+ "5": 4
29
+ },
30
+ "layer_norm_eps": 1e-12,
31
+ "max_position_embeddings": 512,
32
+ "model_type": "bert",
33
+ "num_attention_heads": 12,
34
+ "num_hidden_layers": 12,
35
+ "output_past": true,
36
+ "pad_token_id": 0,
37
+ "pooler_fc_size": 768,
38
+ "pooler_num_attention_heads": 12,
39
+ "pooler_num_fc_layers": 3,
40
+ "pooler_size_per_head": 128,
41
+ "pooler_type": "first_token_transform",
42
+ "position_embedding_type": "absolute",
43
+ "problem_type": "single_label_classification",
44
+ "torch_dtype": "float32",
45
+ "transformers_version": "4.46.2",
46
+ "type_vocab_size": 2,
47
+ "use_cache": true,
48
+ "vocab_size": 119547
49
+ }
checkpoint-7302/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:1482e954dc7bbef048c49c06bdc65118b9138ec4ca7e91f7933627bf7f2df339
3
+ size 711452684
checkpoint-7302/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d34860e072b3e89ae53915f271347e78ded67fb3cfb4a99a3e2e844495bcacb1
3
+ size 1423026426
checkpoint-7302/rng_state_0.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0854bb32a2482ca3dbd6c9360ad92f55abbc1c6a7252c1425c386b9e1c0ba633
3
+ size 14512
checkpoint-7302/rng_state_1.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:6c8952e15270b4d4ee66d9d14430aadfa68e894b1fcb2fb7df94f6fc7998bf10
3
+ size 14512
checkpoint-7302/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f73a56f8280d093a2719832534cf685e9b73a8df36f27a7e25f92295dcdeb2da
3
+ size 1064
checkpoint-7302/trainer_state.json ADDED
@@ -0,0 +1,2140 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": 0.8904216289520264,
3
+ "best_model_checkpoint": "rubert-base-ru-huawei-sentiment-fine-up/checkpoint-7302",
4
+ "epoch": 3.0,
5
+ "eval_steps": 500,
6
+ "global_step": 7302,
7
+ "is_hyper_param_search": false,
8
+ "is_local_process_zero": true,
9
+ "is_world_process_zero": true,
10
+ "log_history": [
11
+ {
12
+ "epoch": 0.010271158586688579,
13
+ "grad_norm": 5.9403300285339355,
14
+ "learning_rate": 5.705157462345961e-07,
15
+ "loss": 0.7105,
16
+ "step": 25
17
+ },
18
+ {
19
+ "epoch": 0.020542317173377157,
20
+ "grad_norm": 10.47209644317627,
21
+ "learning_rate": 1.1410314924691921e-06,
22
+ "loss": 0.77,
23
+ "step": 50
24
+ },
25
+ {
26
+ "epoch": 0.030813475760065736,
27
+ "grad_norm": 29.545536041259766,
28
+ "learning_rate": 1.7115472387037881e-06,
29
+ "loss": 0.6779,
30
+ "step": 75
31
+ },
32
+ {
33
+ "epoch": 0.041084634346754315,
34
+ "grad_norm": 8.517972946166992,
35
+ "learning_rate": 2.2820629849383843e-06,
36
+ "loss": 0.6978,
37
+ "step": 100
38
+ },
39
+ {
40
+ "epoch": 0.05135579293344289,
41
+ "grad_norm": 5.000863075256348,
42
+ "learning_rate": 2.8525787311729803e-06,
43
+ "loss": 0.7072,
44
+ "step": 125
45
+ },
46
+ {
47
+ "epoch": 0.06162695152013147,
48
+ "grad_norm": 6.92432165145874,
49
+ "learning_rate": 3.4230944774075762e-06,
50
+ "loss": 0.7056,
51
+ "step": 150
52
+ },
53
+ {
54
+ "epoch": 0.07189811010682005,
55
+ "grad_norm": 15.27511978149414,
56
+ "learning_rate": 3.993610223642173e-06,
57
+ "loss": 0.7201,
58
+ "step": 175
59
+ },
60
+ {
61
+ "epoch": 0.08216926869350863,
62
+ "grad_norm": 18.27556037902832,
63
+ "learning_rate": 4.564125969876769e-06,
64
+ "loss": 0.6529,
65
+ "step": 200
66
+ },
67
+ {
68
+ "epoch": 0.0924404272801972,
69
+ "grad_norm": 6.184650897979736,
70
+ "learning_rate": 5.111821086261981e-06,
71
+ "loss": 0.6994,
72
+ "step": 225
73
+ },
74
+ {
75
+ "epoch": 0.10271158586688578,
76
+ "grad_norm": 5.983186721801758,
77
+ "learning_rate": 5.682336832496577e-06,
78
+ "loss": 0.674,
79
+ "step": 250
80
+ },
81
+ {
82
+ "epoch": 0.11298274445357437,
83
+ "grad_norm": 10.52604866027832,
84
+ "learning_rate": 6.252852578731174e-06,
85
+ "loss": 0.6855,
86
+ "step": 275
87
+ },
88
+ {
89
+ "epoch": 0.12325390304026294,
90
+ "grad_norm": 7.054647922515869,
91
+ "learning_rate": 6.823368324965769e-06,
92
+ "loss": 0.6427,
93
+ "step": 300
94
+ },
95
+ {
96
+ "epoch": 0.13352506162695152,
97
+ "grad_norm": 10.486727714538574,
98
+ "learning_rate": 7.393884071200366e-06,
99
+ "loss": 0.7255,
100
+ "step": 325
101
+ },
102
+ {
103
+ "epoch": 0.1437962202136401,
104
+ "grad_norm": 4.579376697540283,
105
+ "learning_rate": 7.964399817434962e-06,
106
+ "loss": 0.5189,
107
+ "step": 350
108
+ },
109
+ {
110
+ "epoch": 0.15406737880032867,
111
+ "grad_norm": 5.1492414474487305,
112
+ "learning_rate": 8.534915563669557e-06,
113
+ "loss": 0.6923,
114
+ "step": 375
115
+ },
116
+ {
117
+ "epoch": 0.16433853738701726,
118
+ "grad_norm": 10.567380905151367,
119
+ "learning_rate": 9.105431309904154e-06,
120
+ "loss": 0.8093,
121
+ "step": 400
122
+ },
123
+ {
124
+ "epoch": 0.17460969597370585,
125
+ "grad_norm": 4.735847473144531,
126
+ "learning_rate": 9.67594705613875e-06,
127
+ "loss": 0.5677,
128
+ "step": 425
129
+ },
130
+ {
131
+ "epoch": 0.1848808545603944,
132
+ "grad_norm": 6.456740856170654,
133
+ "learning_rate": 1.0246462802373347e-05,
134
+ "loss": 0.6611,
135
+ "step": 450
136
+ },
137
+ {
138
+ "epoch": 0.195152013147083,
139
+ "grad_norm": 11.575152397155762,
140
+ "learning_rate": 1.0816978548607942e-05,
141
+ "loss": 0.7,
142
+ "step": 475
143
+ },
144
+ {
145
+ "epoch": 0.20542317173377156,
146
+ "grad_norm": 11.752349853515625,
147
+ "learning_rate": 1.1387494294842537e-05,
148
+ "loss": 0.6638,
149
+ "step": 500
150
+ },
151
+ {
152
+ "epoch": 0.21569433032046015,
153
+ "grad_norm": 21.79088020324707,
154
+ "learning_rate": 1.1958010041077134e-05,
155
+ "loss": 0.596,
156
+ "step": 525
157
+ },
158
+ {
159
+ "epoch": 0.22596548890714874,
160
+ "grad_norm": 6.109403610229492,
161
+ "learning_rate": 1.2528525787311731e-05,
162
+ "loss": 0.6791,
163
+ "step": 550
164
+ },
165
+ {
166
+ "epoch": 0.2362366474938373,
167
+ "grad_norm": 15.077224731445312,
168
+ "learning_rate": 1.3099041533546328e-05,
169
+ "loss": 0.7015,
170
+ "step": 575
171
+ },
172
+ {
173
+ "epoch": 0.2465078060805259,
174
+ "grad_norm": 9.261869430541992,
175
+ "learning_rate": 1.3669557279780923e-05,
176
+ "loss": 0.698,
177
+ "step": 600
178
+ },
179
+ {
180
+ "epoch": 0.25677896466721445,
181
+ "grad_norm": 7.057260036468506,
182
+ "learning_rate": 1.4240073026015518e-05,
183
+ "loss": 0.6274,
184
+ "step": 625
185
+ },
186
+ {
187
+ "epoch": 0.26705012325390304,
188
+ "grad_norm": 7.404267311096191,
189
+ "learning_rate": 1.4810588772250115e-05,
190
+ "loss": 0.6668,
191
+ "step": 650
192
+ },
193
+ {
194
+ "epoch": 0.2773212818405916,
195
+ "grad_norm": 9.93870735168457,
196
+ "learning_rate": 1.538110451848471e-05,
197
+ "loss": 0.7213,
198
+ "step": 675
199
+ },
200
+ {
201
+ "epoch": 0.2875924404272802,
202
+ "grad_norm": 7.443170547485352,
203
+ "learning_rate": 1.5951620264719307e-05,
204
+ "loss": 0.6831,
205
+ "step": 700
206
+ },
207
+ {
208
+ "epoch": 0.29786359901396875,
209
+ "grad_norm": 9.246779441833496,
210
+ "learning_rate": 1.6522136010953902e-05,
211
+ "loss": 0.7139,
212
+ "step": 725
213
+ },
214
+ {
215
+ "epoch": 0.30813475760065734,
216
+ "grad_norm": 6.488850116729736,
217
+ "learning_rate": 1.70926517571885e-05,
218
+ "loss": 0.6312,
219
+ "step": 750
220
+ },
221
+ {
222
+ "epoch": 0.3184059161873459,
223
+ "grad_norm": 6.905552387237549,
224
+ "learning_rate": 1.7663167503423096e-05,
225
+ "loss": 0.6549,
226
+ "step": 775
227
+ },
228
+ {
229
+ "epoch": 0.3286770747740345,
230
+ "grad_norm": 18.033992767333984,
231
+ "learning_rate": 1.823368324965769e-05,
232
+ "loss": 0.5369,
233
+ "step": 800
234
+ },
235
+ {
236
+ "epoch": 0.3389482333607231,
237
+ "grad_norm": 16.312898635864258,
238
+ "learning_rate": 1.880419899589229e-05,
239
+ "loss": 0.6098,
240
+ "step": 825
241
+ },
242
+ {
243
+ "epoch": 0.3492193919474117,
244
+ "grad_norm": 14.354256629943848,
245
+ "learning_rate": 1.9374714742126884e-05,
246
+ "loss": 0.6917,
247
+ "step": 850
248
+ },
249
+ {
250
+ "epoch": 0.35949055053410023,
251
+ "grad_norm": 16.478256225585938,
252
+ "learning_rate": 1.994523048836148e-05,
253
+ "loss": 0.6639,
254
+ "step": 875
255
+ },
256
+ {
257
+ "epoch": 0.3697617091207888,
258
+ "grad_norm": 3.9060003757476807,
259
+ "learning_rate": 2.0515746234596075e-05,
260
+ "loss": 0.6152,
261
+ "step": 900
262
+ },
263
+ {
264
+ "epoch": 0.3800328677074774,
265
+ "grad_norm": 22.826074600219727,
266
+ "learning_rate": 2.108626198083067e-05,
267
+ "loss": 0.7206,
268
+ "step": 925
269
+ },
270
+ {
271
+ "epoch": 0.390304026294166,
272
+ "grad_norm": 8.423962593078613,
273
+ "learning_rate": 2.1656777727065268e-05,
274
+ "loss": 0.6332,
275
+ "step": 950
276
+ },
277
+ {
278
+ "epoch": 0.4005751848808546,
279
+ "grad_norm": 6.902738094329834,
280
+ "learning_rate": 2.2227293473299863e-05,
281
+ "loss": 0.7475,
282
+ "step": 975
283
+ },
284
+ {
285
+ "epoch": 0.4108463434675431,
286
+ "grad_norm": 12.147488594055176,
287
+ "learning_rate": 2.2797809219534462e-05,
288
+ "loss": 0.6305,
289
+ "step": 1000
290
+ },
291
+ {
292
+ "epoch": 0.4211175020542317,
293
+ "grad_norm": 9.553220748901367,
294
+ "learning_rate": 2.3368324965769057e-05,
295
+ "loss": 0.6422,
296
+ "step": 1025
297
+ },
298
+ {
299
+ "epoch": 0.4313886606409203,
300
+ "grad_norm": 10.641242980957031,
301
+ "learning_rate": 2.3938840712003652e-05,
302
+ "loss": 0.7081,
303
+ "step": 1050
304
+ },
305
+ {
306
+ "epoch": 0.4416598192276089,
307
+ "grad_norm": 24.202816009521484,
308
+ "learning_rate": 2.4509356458238247e-05,
309
+ "loss": 0.7302,
310
+ "step": 1075
311
+ },
312
+ {
313
+ "epoch": 0.4519309778142975,
314
+ "grad_norm": 28.98261070251465,
315
+ "learning_rate": 2.5079872204472842e-05,
316
+ "loss": 0.6791,
317
+ "step": 1100
318
+ },
319
+ {
320
+ "epoch": 0.462202136400986,
321
+ "grad_norm": 9.403679847717285,
322
+ "learning_rate": 2.5650387950707437e-05,
323
+ "loss": 0.6186,
324
+ "step": 1125
325
+ },
326
+ {
327
+ "epoch": 0.4724732949876746,
328
+ "grad_norm": 29.80078125,
329
+ "learning_rate": 2.622090369694204e-05,
330
+ "loss": 0.6075,
331
+ "step": 1150
332
+ },
333
+ {
334
+ "epoch": 0.4827444535743632,
335
+ "grad_norm": 6.909419059753418,
336
+ "learning_rate": 2.6791419443176634e-05,
337
+ "loss": 0.6069,
338
+ "step": 1175
339
+ },
340
+ {
341
+ "epoch": 0.4930156121610518,
342
+ "grad_norm": 23.68714141845703,
343
+ "learning_rate": 2.7339114559561846e-05,
344
+ "loss": 0.6602,
345
+ "step": 1200
346
+ },
347
+ {
348
+ "epoch": 0.5032867707477403,
349
+ "grad_norm": 11.304485321044922,
350
+ "learning_rate": 2.790963030579644e-05,
351
+ "loss": 0.689,
352
+ "step": 1225
353
+ },
354
+ {
355
+ "epoch": 0.5135579293344289,
356
+ "grad_norm": 9.599061965942383,
357
+ "learning_rate": 2.8480146052031036e-05,
358
+ "loss": 0.6947,
359
+ "step": 1250
360
+ },
361
+ {
362
+ "epoch": 0.5238290879211175,
363
+ "grad_norm": 17.570533752441406,
364
+ "learning_rate": 2.905066179826563e-05,
365
+ "loss": 0.6341,
366
+ "step": 1275
367
+ },
368
+ {
369
+ "epoch": 0.5341002465078061,
370
+ "grad_norm": 23.402406692504883,
371
+ "learning_rate": 2.962117754450023e-05,
372
+ "loss": 0.6246,
373
+ "step": 1300
374
+ },
375
+ {
376
+ "epoch": 0.5443714050944947,
377
+ "grad_norm": 14.925309181213379,
378
+ "learning_rate": 3.0191693290734825e-05,
379
+ "loss": 0.6841,
380
+ "step": 1325
381
+ },
382
+ {
383
+ "epoch": 0.5546425636811833,
384
+ "grad_norm": 29.768821716308594,
385
+ "learning_rate": 3.076220903696942e-05,
386
+ "loss": 0.6018,
387
+ "step": 1350
388
+ },
389
+ {
390
+ "epoch": 0.5649137222678718,
391
+ "grad_norm": 36.33218765258789,
392
+ "learning_rate": 3.133272478320402e-05,
393
+ "loss": 0.6589,
394
+ "step": 1375
395
+ },
396
+ {
397
+ "epoch": 0.5751848808545604,
398
+ "grad_norm": 15.992273330688477,
399
+ "learning_rate": 3.1903240529438614e-05,
400
+ "loss": 0.6581,
401
+ "step": 1400
402
+ },
403
+ {
404
+ "epoch": 0.585456039441249,
405
+ "grad_norm": 21.353836059570312,
406
+ "learning_rate": 3.247375627567321e-05,
407
+ "loss": 0.6803,
408
+ "step": 1425
409
+ },
410
+ {
411
+ "epoch": 0.5957271980279375,
412
+ "grad_norm": 19.363903045654297,
413
+ "learning_rate": 3.3044272021907804e-05,
414
+ "loss": 0.6636,
415
+ "step": 1450
416
+ },
417
+ {
418
+ "epoch": 0.6059983566146261,
419
+ "grad_norm": 9.495320320129395,
420
+ "learning_rate": 3.3614787768142406e-05,
421
+ "loss": 0.6562,
422
+ "step": 1475
423
+ },
424
+ {
425
+ "epoch": 0.6162695152013147,
426
+ "grad_norm": 18.960094451904297,
427
+ "learning_rate": 3.4185303514377e-05,
428
+ "loss": 0.6603,
429
+ "step": 1500
430
+ },
431
+ {
432
+ "epoch": 0.6265406737880033,
433
+ "grad_norm": 23.797956466674805,
434
+ "learning_rate": 3.4755819260611596e-05,
435
+ "loss": 0.6329,
436
+ "step": 1525
437
+ },
438
+ {
439
+ "epoch": 0.6368118323746919,
440
+ "grad_norm": 13.95300579071045,
441
+ "learning_rate": 3.532633500684619e-05,
442
+ "loss": 0.7285,
443
+ "step": 1550
444
+ },
445
+ {
446
+ "epoch": 0.6470829909613804,
447
+ "grad_norm": 21.96845245361328,
448
+ "learning_rate": 3.5896850753080786e-05,
449
+ "loss": 0.7046,
450
+ "step": 1575
451
+ },
452
+ {
453
+ "epoch": 0.657354149548069,
454
+ "grad_norm": 22.858348846435547,
455
+ "learning_rate": 3.646736649931538e-05,
456
+ "loss": 0.6791,
457
+ "step": 1600
458
+ },
459
+ {
460
+ "epoch": 0.6676253081347576,
461
+ "grad_norm": 19.458847045898438,
462
+ "learning_rate": 3.7037882245549977e-05,
463
+ "loss": 0.7317,
464
+ "step": 1625
465
+ },
466
+ {
467
+ "epoch": 0.6778964667214462,
468
+ "grad_norm": 5.415579319000244,
469
+ "learning_rate": 3.760839799178458e-05,
470
+ "loss": 0.6838,
471
+ "step": 1650
472
+ },
473
+ {
474
+ "epoch": 0.6881676253081348,
475
+ "grad_norm": 16.664749145507812,
476
+ "learning_rate": 3.8178913738019174e-05,
477
+ "loss": 0.5959,
478
+ "step": 1675
479
+ },
480
+ {
481
+ "epoch": 0.6984387838948234,
482
+ "grad_norm": 14.080728530883789,
483
+ "learning_rate": 3.874942948425377e-05,
484
+ "loss": 0.6315,
485
+ "step": 1700
486
+ },
487
+ {
488
+ "epoch": 0.7087099424815119,
489
+ "grad_norm": 10.989696502685547,
490
+ "learning_rate": 3.9319945230488364e-05,
491
+ "loss": 0.7356,
492
+ "step": 1725
493
+ },
494
+ {
495
+ "epoch": 0.7189811010682005,
496
+ "grad_norm": 11.903764724731445,
497
+ "learning_rate": 3.989046097672296e-05,
498
+ "loss": 0.6935,
499
+ "step": 1750
500
+ },
501
+ {
502
+ "epoch": 0.729252259654889,
503
+ "grad_norm": 3.6907122135162354,
504
+ "learning_rate": 4.0460976722957554e-05,
505
+ "loss": 0.6909,
506
+ "step": 1775
507
+ },
508
+ {
509
+ "epoch": 0.7395234182415776,
510
+ "grad_norm": 10.299661636352539,
511
+ "learning_rate": 4.103149246919215e-05,
512
+ "loss": 0.7183,
513
+ "step": 1800
514
+ },
515
+ {
516
+ "epoch": 0.7497945768282662,
517
+ "grad_norm": 18.38177490234375,
518
+ "learning_rate": 4.1602008215426744e-05,
519
+ "loss": 0.6862,
520
+ "step": 1825
521
+ },
522
+ {
523
+ "epoch": 0.7600657354149548,
524
+ "grad_norm": 10.906993865966797,
525
+ "learning_rate": 4.217252396166134e-05,
526
+ "loss": 0.7489,
527
+ "step": 1850
528
+ },
529
+ {
530
+ "epoch": 0.7703368940016434,
531
+ "grad_norm": 11.314064025878906,
532
+ "learning_rate": 4.2743039707895935e-05,
533
+ "loss": 0.7735,
534
+ "step": 1875
535
+ },
536
+ {
537
+ "epoch": 0.780608052588332,
538
+ "grad_norm": 30.717824935913086,
539
+ "learning_rate": 4.3313555454130536e-05,
540
+ "loss": 0.7382,
541
+ "step": 1900
542
+ },
543
+ {
544
+ "epoch": 0.7908792111750206,
545
+ "grad_norm": 26.039527893066406,
546
+ "learning_rate": 4.388407120036513e-05,
547
+ "loss": 0.741,
548
+ "step": 1925
549
+ },
550
+ {
551
+ "epoch": 0.8011503697617092,
552
+ "grad_norm": 21.928796768188477,
553
+ "learning_rate": 4.445458694659973e-05,
554
+ "loss": 0.648,
555
+ "step": 1950
556
+ },
557
+ {
558
+ "epoch": 0.8114215283483976,
559
+ "grad_norm": 15.757039070129395,
560
+ "learning_rate": 4.502510269283433e-05,
561
+ "loss": 0.6965,
562
+ "step": 1975
563
+ },
564
+ {
565
+ "epoch": 0.8216926869350862,
566
+ "grad_norm": 4.059453010559082,
567
+ "learning_rate": 4.5595618439068924e-05,
568
+ "loss": 0.5749,
569
+ "step": 2000
570
+ },
571
+ {
572
+ "epoch": 0.8319638455217748,
573
+ "grad_norm": 25.09820556640625,
574
+ "learning_rate": 4.616613418530352e-05,
575
+ "loss": 0.7363,
576
+ "step": 2025
577
+ },
578
+ {
579
+ "epoch": 0.8422350041084634,
580
+ "grad_norm": 22.875272750854492,
581
+ "learning_rate": 4.6736649931538114e-05,
582
+ "loss": 0.7879,
583
+ "step": 2050
584
+ },
585
+ {
586
+ "epoch": 0.852506162695152,
587
+ "grad_norm": 27.716041564941406,
588
+ "learning_rate": 4.730716567777271e-05,
589
+ "loss": 0.6949,
590
+ "step": 2075
591
+ },
592
+ {
593
+ "epoch": 0.8627773212818406,
594
+ "grad_norm": 7.766871452331543,
595
+ "learning_rate": 4.7877681424007304e-05,
596
+ "loss": 0.7501,
597
+ "step": 2100
598
+ },
599
+ {
600
+ "epoch": 0.8730484798685292,
601
+ "grad_norm": 50.67536163330078,
602
+ "learning_rate": 4.84481971702419e-05,
603
+ "loss": 0.7346,
604
+ "step": 2125
605
+ },
606
+ {
607
+ "epoch": 0.8833196384552178,
608
+ "grad_norm": 30.061460494995117,
609
+ "learning_rate": 4.9018712916476494e-05,
610
+ "loss": 0.702,
611
+ "step": 2150
612
+ },
613
+ {
614
+ "epoch": 0.8935907970419064,
615
+ "grad_norm": 19.098934173583984,
616
+ "learning_rate": 4.9589228662711096e-05,
617
+ "loss": 0.7304,
618
+ "step": 2175
619
+ },
620
+ {
621
+ "epoch": 0.903861955628595,
622
+ "grad_norm": 23.1389217376709,
623
+ "learning_rate": 4.9999984447069524e-05,
624
+ "loss": 0.7753,
625
+ "step": 2200
626
+ },
627
+ {
628
+ "epoch": 0.9141331142152835,
629
+ "grad_norm": 15.105667114257812,
630
+ "learning_rate": 4.9999674976164344e-05,
631
+ "loss": 0.7878,
632
+ "step": 2225
633
+ },
634
+ {
635
+ "epoch": 0.924404272801972,
636
+ "grad_norm": 38.34989929199219,
637
+ "learning_rate": 4.9998968752470534e-05,
638
+ "loss": 0.7374,
639
+ "step": 2250
640
+ },
641
+ {
642
+ "epoch": 0.9346754313886606,
643
+ "grad_norm": 16.31295394897461,
644
+ "learning_rate": 4.9997865787196075e-05,
645
+ "loss": 0.6709,
646
+ "step": 2275
647
+ },
648
+ {
649
+ "epoch": 0.9449465899753492,
650
+ "grad_norm": 9.819225311279297,
651
+ "learning_rate": 4.9996366097845385e-05,
652
+ "loss": 0.7123,
653
+ "step": 2300
654
+ },
655
+ {
656
+ "epoch": 0.9552177485620378,
657
+ "grad_norm": 21.98024559020996,
658
+ "learning_rate": 4.999446970821902e-05,
659
+ "loss": 0.7217,
660
+ "step": 2325
661
+ },
662
+ {
663
+ "epoch": 0.9654889071487264,
664
+ "grad_norm": 13.155325889587402,
665
+ "learning_rate": 4.999217664841326e-05,
666
+ "loss": 0.7577,
667
+ "step": 2350
668
+ },
669
+ {
670
+ "epoch": 0.975760065735415,
671
+ "grad_norm": 7.717748641967773,
672
+ "learning_rate": 4.998948695481972e-05,
673
+ "loss": 0.7836,
674
+ "step": 2375
675
+ },
676
+ {
677
+ "epoch": 0.9860312243221035,
678
+ "grad_norm": 8.747546195983887,
679
+ "learning_rate": 4.998640067012468e-05,
680
+ "loss": 0.8113,
681
+ "step": 2400
682
+ },
683
+ {
684
+ "epoch": 0.9963023829087921,
685
+ "grad_norm": 2.1325223445892334,
686
+ "learning_rate": 4.998291784330846e-05,
687
+ "loss": 0.7189,
688
+ "step": 2425
689
+ },
690
+ {
691
+ "epoch": 1.0,
692
+ "eval_accuracy": 0.6579677386211856,
693
+ "eval_f1_macro": 0.4525953534948945,
694
+ "eval_f1_micro": 0.6579677386211856,
695
+ "eval_f1_weighted": 0.6282433368073111,
696
+ "eval_loss": 0.9721999168395996,
697
+ "eval_precision_macro": 0.49011451139838114,
698
+ "eval_precision_micro": 0.6579677386211856,
699
+ "eval_precision_weighted": 0.6153964777282442,
700
+ "eval_recall_macro": 0.44923284497959803,
701
+ "eval_recall_micro": 0.6579677386211856,
702
+ "eval_recall_weighted": 0.6579677386211856,
703
+ "eval_runtime": 20.3328,
704
+ "eval_samples_per_second": 478.684,
705
+ "eval_steps_per_second": 15.0,
706
+ "step": 2434
707
+ },
708
+ {
709
+ "epoch": 1.0065735414954806,
710
+ "grad_norm": 13.705938339233398,
711
+ "learning_rate": 4.997903852964464e-05,
712
+ "loss": 0.8531,
713
+ "step": 2450
714
+ },
715
+ {
716
+ "epoch": 1.0168447000821692,
717
+ "grad_norm": 13.030699729919434,
718
+ "learning_rate": 4.997476279069914e-05,
719
+ "loss": 0.6717,
720
+ "step": 2475
721
+ },
722
+ {
723
+ "epoch": 1.0271158586688578,
724
+ "grad_norm": 15.674768447875977,
725
+ "learning_rate": 4.9970090694329316e-05,
726
+ "loss": 0.6566,
727
+ "step": 2500
728
+ },
729
+ {
730
+ "epoch": 1.0373870172555464,
731
+ "grad_norm": 11.915541648864746,
732
+ "learning_rate": 4.996502231468281e-05,
733
+ "loss": 0.7618,
734
+ "step": 2525
735
+ },
736
+ {
737
+ "epoch": 1.047658175842235,
738
+ "grad_norm": 7.769870281219482,
739
+ "learning_rate": 4.995955773219642e-05,
740
+ "loss": 0.7373,
741
+ "step": 2550
742
+ },
743
+ {
744
+ "epoch": 1.0579293344289236,
745
+ "grad_norm": 10.591928482055664,
746
+ "learning_rate": 4.9953697033594805e-05,
747
+ "loss": 0.7792,
748
+ "step": 2575
749
+ },
750
+ {
751
+ "epoch": 1.0682004930156122,
752
+ "grad_norm": 7.153607368469238,
753
+ "learning_rate": 4.9947440311889124e-05,
754
+ "loss": 0.7231,
755
+ "step": 2600
756
+ },
757
+ {
758
+ "epoch": 1.0784716516023007,
759
+ "grad_norm": 32.34482955932617,
760
+ "learning_rate": 4.994078766637553e-05,
761
+ "loss": 0.7746,
762
+ "step": 2625
763
+ },
764
+ {
765
+ "epoch": 1.0887428101889893,
766
+ "grad_norm": 14.119093894958496,
767
+ "learning_rate": 4.993373920263363e-05,
768
+ "loss": 0.7627,
769
+ "step": 2650
770
+ },
771
+ {
772
+ "epoch": 1.099013968775678,
773
+ "grad_norm": 6.369531631469727,
774
+ "learning_rate": 4.9926295032524774e-05,
775
+ "loss": 0.6739,
776
+ "step": 2675
777
+ },
778
+ {
779
+ "epoch": 1.1092851273623665,
780
+ "grad_norm": 32.834388732910156,
781
+ "learning_rate": 4.991845527419032e-05,
782
+ "loss": 0.7351,
783
+ "step": 2700
784
+ },
785
+ {
786
+ "epoch": 1.119556285949055,
787
+ "grad_norm": 2.3463103771209717,
788
+ "learning_rate": 4.991022005204972e-05,
789
+ "loss": 0.699,
790
+ "step": 2725
791
+ },
792
+ {
793
+ "epoch": 1.1298274445357437,
794
+ "grad_norm": 19.840364456176758,
795
+ "learning_rate": 4.990158949679856e-05,
796
+ "loss": 0.7338,
797
+ "step": 2750
798
+ },
799
+ {
800
+ "epoch": 1.1400986031224323,
801
+ "grad_norm": 18.111513137817383,
802
+ "learning_rate": 4.9892563745406506e-05,
803
+ "loss": 0.6833,
804
+ "step": 2775
805
+ },
806
+ {
807
+ "epoch": 1.1503697617091209,
808
+ "grad_norm": 8.486504554748535,
809
+ "learning_rate": 4.98831429411151e-05,
810
+ "loss": 0.8016,
811
+ "step": 2800
812
+ },
813
+ {
814
+ "epoch": 1.1606409202958095,
815
+ "grad_norm": 18.1845703125,
816
+ "learning_rate": 4.9873327233435487e-05,
817
+ "loss": 0.7478,
818
+ "step": 2825
819
+ },
820
+ {
821
+ "epoch": 1.170912078882498,
822
+ "grad_norm": 13.720683097839355,
823
+ "learning_rate": 4.9863116778146086e-05,
824
+ "loss": 0.6556,
825
+ "step": 2850
826
+ },
827
+ {
828
+ "epoch": 1.1811832374691864,
829
+ "grad_norm": 0.879672646522522,
830
+ "learning_rate": 4.9852511737290065e-05,
831
+ "loss": 0.773,
832
+ "step": 2875
833
+ },
834
+ {
835
+ "epoch": 1.191454396055875,
836
+ "grad_norm": 5.223727703094482,
837
+ "learning_rate": 4.9841512279172776e-05,
838
+ "loss": 0.6251,
839
+ "step": 2900
840
+ },
841
+ {
842
+ "epoch": 1.2017255546425636,
843
+ "grad_norm": 26.75752830505371,
844
+ "learning_rate": 4.983011857835914e-05,
845
+ "loss": 0.7045,
846
+ "step": 2925
847
+ },
848
+ {
849
+ "epoch": 1.2119967132292522,
850
+ "grad_norm": 18.040035247802734,
851
+ "learning_rate": 4.98183308156708e-05,
852
+ "loss": 0.6991,
853
+ "step": 2950
854
+ },
855
+ {
856
+ "epoch": 1.2222678718159408,
857
+ "grad_norm": 1.8346068859100342,
858
+ "learning_rate": 4.980614917818329e-05,
859
+ "loss": 0.648,
860
+ "step": 2975
861
+ },
862
+ {
863
+ "epoch": 1.2325390304026294,
864
+ "grad_norm": 13.46601390838623,
865
+ "learning_rate": 4.9793573859223076e-05,
866
+ "loss": 0.656,
867
+ "step": 3000
868
+ },
869
+ {
870
+ "epoch": 1.242810188989318,
871
+ "grad_norm": 18.629392623901367,
872
+ "learning_rate": 4.978060505836447e-05,
873
+ "loss": 0.7385,
874
+ "step": 3025
875
+ },
876
+ {
877
+ "epoch": 1.2530813475760065,
878
+ "grad_norm": 8.506892204284668,
879
+ "learning_rate": 4.976724298142646e-05,
880
+ "loss": 0.6356,
881
+ "step": 3050
882
+ },
883
+ {
884
+ "epoch": 1.2633525061626951,
885
+ "grad_norm": 24.73647689819336,
886
+ "learning_rate": 4.975348784046946e-05,
887
+ "loss": 0.7267,
888
+ "step": 3075
889
+ },
890
+ {
891
+ "epoch": 1.2736236647493837,
892
+ "grad_norm": 4.445428848266602,
893
+ "learning_rate": 4.9739339853791925e-05,
894
+ "loss": 0.7148,
895
+ "step": 3100
896
+ },
897
+ {
898
+ "epoch": 1.2838948233360723,
899
+ "grad_norm": 15.720209121704102,
900
+ "learning_rate": 4.9724799245926895e-05,
901
+ "loss": 0.7137,
902
+ "step": 3125
903
+ },
904
+ {
905
+ "epoch": 1.2941659819227609,
906
+ "grad_norm": 21.483966827392578,
907
+ "learning_rate": 4.970986624763845e-05,
908
+ "loss": 0.6836,
909
+ "step": 3150
910
+ },
911
+ {
912
+ "epoch": 1.3044371405094495,
913
+ "grad_norm": 5.221400737762451,
914
+ "learning_rate": 4.969516162830212e-05,
915
+ "loss": 0.7124,
916
+ "step": 3175
917
+ },
918
+ {
919
+ "epoch": 1.314708299096138,
920
+ "grad_norm": 25.75131607055664,
921
+ "learning_rate": 4.967946023802564e-05,
922
+ "loss": 0.7195,
923
+ "step": 3200
924
+ },
925
+ {
926
+ "epoch": 1.3249794576828267,
927
+ "grad_norm": 5.5664167404174805,
928
+ "learning_rate": 4.96633671768702e-05,
929
+ "loss": 0.7364,
930
+ "step": 3225
931
+ },
932
+ {
933
+ "epoch": 1.3352506162695152,
934
+ "grad_norm": 12.063735008239746,
935
+ "learning_rate": 4.9646882700237805e-05,
936
+ "loss": 0.5976,
937
+ "step": 3250
938
+ },
939
+ {
940
+ "epoch": 1.3455217748562038,
941
+ "grad_norm": 30.29339027404785,
942
+ "learning_rate": 4.9630007069742345e-05,
943
+ "loss": 0.714,
944
+ "step": 3275
945
+ },
946
+ {
947
+ "epoch": 1.3557929334428924,
948
+ "grad_norm": 3.4963226318359375,
949
+ "learning_rate": 4.961274055320543e-05,
950
+ "loss": 0.6268,
951
+ "step": 3300
952
+ },
953
+ {
954
+ "epoch": 1.366064092029581,
955
+ "grad_norm": 13.199789047241211,
956
+ "learning_rate": 4.9595083424652164e-05,
957
+ "loss": 0.7015,
958
+ "step": 3325
959
+ },
960
+ {
961
+ "epoch": 1.3763352506162696,
962
+ "grad_norm": 30.219947814941406,
963
+ "learning_rate": 4.957703596430679e-05,
964
+ "loss": 0.6893,
965
+ "step": 3350
966
+ },
967
+ {
968
+ "epoch": 1.3866064092029582,
969
+ "grad_norm": 16.250619888305664,
970
+ "learning_rate": 4.955859845858826e-05,
971
+ "loss": 0.7416,
972
+ "step": 3375
973
+ },
974
+ {
975
+ "epoch": 1.3968775677896468,
976
+ "grad_norm": 11.023963928222656,
977
+ "learning_rate": 4.953977120010563e-05,
978
+ "loss": 0.5854,
979
+ "step": 3400
980
+ },
981
+ {
982
+ "epoch": 1.4071487263763354,
983
+ "grad_norm": 11.984811782836914,
984
+ "learning_rate": 4.952055448765348e-05,
985
+ "loss": 0.7531,
986
+ "step": 3425
987
+ },
988
+ {
989
+ "epoch": 1.417419884963024,
990
+ "grad_norm": 10.170748710632324,
991
+ "learning_rate": 4.950094862620715e-05,
992
+ "loss": 0.684,
993
+ "step": 3450
994
+ },
995
+ {
996
+ "epoch": 1.4276910435497125,
997
+ "grad_norm": 26.633150100708008,
998
+ "learning_rate": 4.9480953926917886e-05,
999
+ "loss": 0.6428,
1000
+ "step": 3475
1001
+ },
1002
+ {
1003
+ "epoch": 1.437962202136401,
1004
+ "grad_norm": 71.91981506347656,
1005
+ "learning_rate": 4.946057070710793e-05,
1006
+ "loss": 0.7043,
1007
+ "step": 3500
1008
+ },
1009
+ {
1010
+ "epoch": 1.4482333607230895,
1011
+ "grad_norm": 55.96223831176758,
1012
+ "learning_rate": 4.943979929026546e-05,
1013
+ "loss": 0.8018,
1014
+ "step": 3525
1015
+ },
1016
+ {
1017
+ "epoch": 1.458504519309778,
1018
+ "grad_norm": 25.171337127685547,
1019
+ "learning_rate": 4.9418640006039464e-05,
1020
+ "loss": 0.6975,
1021
+ "step": 3550
1022
+ },
1023
+ {
1024
+ "epoch": 1.4687756778964667,
1025
+ "grad_norm": 16.321897506713867,
1026
+ "learning_rate": 4.9397093190234495e-05,
1027
+ "loss": 0.7242,
1028
+ "step": 3575
1029
+ },
1030
+ {
1031
+ "epoch": 1.4790468364831553,
1032
+ "grad_norm": 14.492510795593262,
1033
+ "learning_rate": 4.937515918480538e-05,
1034
+ "loss": 0.6604,
1035
+ "step": 3600
1036
+ },
1037
+ {
1038
+ "epoch": 1.4893179950698439,
1039
+ "grad_norm": 3.452258348464966,
1040
+ "learning_rate": 4.935283833785176e-05,
1041
+ "loss": 0.7074,
1042
+ "step": 3625
1043
+ },
1044
+ {
1045
+ "epoch": 1.4995891536565324,
1046
+ "grad_norm": 50.8736457824707,
1047
+ "learning_rate": 4.933013100361257e-05,
1048
+ "loss": 0.8395,
1049
+ "step": 3650
1050
+ },
1051
+ {
1052
+ "epoch": 1.509860312243221,
1053
+ "grad_norm": 17.3315372467041,
1054
+ "learning_rate": 4.930703754246041e-05,
1055
+ "loss": 0.7815,
1056
+ "step": 3675
1057
+ },
1058
+ {
1059
+ "epoch": 1.5201314708299096,
1060
+ "grad_norm": 27.960744857788086,
1061
+ "learning_rate": 4.928355832089587e-05,
1062
+ "loss": 0.8035,
1063
+ "step": 3700
1064
+ },
1065
+ {
1066
+ "epoch": 1.5304026294165982,
1067
+ "grad_norm": 10.72488784790039,
1068
+ "learning_rate": 4.9259693711541645e-05,
1069
+ "loss": 0.7133,
1070
+ "step": 3725
1071
+ },
1072
+ {
1073
+ "epoch": 1.5406737880032868,
1074
+ "grad_norm": 32.7005729675293,
1075
+ "learning_rate": 4.923544409313668e-05,
1076
+ "loss": 0.6515,
1077
+ "step": 3750
1078
+ },
1079
+ {
1080
+ "epoch": 1.5509449465899754,
1081
+ "grad_norm": 23.12322998046875,
1082
+ "learning_rate": 4.921080985053012e-05,
1083
+ "loss": 0.762,
1084
+ "step": 3775
1085
+ },
1086
+ {
1087
+ "epoch": 1.561216105176664,
1088
+ "grad_norm": 14.074947357177734,
1089
+ "learning_rate": 4.918579137467523e-05,
1090
+ "loss": 0.7669,
1091
+ "step": 3800
1092
+ },
1093
+ {
1094
+ "epoch": 1.5714872637633523,
1095
+ "grad_norm": 23.544923782348633,
1096
+ "learning_rate": 4.9160389062623166e-05,
1097
+ "loss": 0.6909,
1098
+ "step": 3825
1099
+ },
1100
+ {
1101
+ "epoch": 1.581758422350041,
1102
+ "grad_norm": 12.691238403320312,
1103
+ "learning_rate": 4.9134603317516714e-05,
1104
+ "loss": 0.6231,
1105
+ "step": 3850
1106
+ },
1107
+ {
1108
+ "epoch": 1.5920295809367295,
1109
+ "grad_norm": 48.76417541503906,
1110
+ "learning_rate": 4.910843454858383e-05,
1111
+ "loss": 0.7732,
1112
+ "step": 3875
1113
+ },
1114
+ {
1115
+ "epoch": 1.6023007395234181,
1116
+ "grad_norm": 12.748458862304688,
1117
+ "learning_rate": 4.908188317113119e-05,
1118
+ "loss": 0.6368,
1119
+ "step": 3900
1120
+ },
1121
+ {
1122
+ "epoch": 1.6125718981101067,
1123
+ "grad_norm": 52.46083068847656,
1124
+ "learning_rate": 4.905494960653762e-05,
1125
+ "loss": 0.7586,
1126
+ "step": 3925
1127
+ },
1128
+ {
1129
+ "epoch": 1.6228430566967953,
1130
+ "grad_norm": 31.678646087646484,
1131
+ "learning_rate": 4.902763428224733e-05,
1132
+ "loss": 0.8258,
1133
+ "step": 3950
1134
+ },
1135
+ {
1136
+ "epoch": 1.6331142152834839,
1137
+ "grad_norm": 9.948538780212402,
1138
+ "learning_rate": 4.8999937631763196e-05,
1139
+ "loss": 0.7389,
1140
+ "step": 3975
1141
+ },
1142
+ {
1143
+ "epoch": 1.6433853738701725,
1144
+ "grad_norm": 19.353836059570312,
1145
+ "learning_rate": 4.8971860094639874e-05,
1146
+ "loss": 0.6712,
1147
+ "step": 4000
1148
+ },
1149
+ {
1150
+ "epoch": 1.653656532456861,
1151
+ "grad_norm": 12.594712257385254,
1152
+ "learning_rate": 4.894340211647681e-05,
1153
+ "loss": 0.7761,
1154
+ "step": 4025
1155
+ },
1156
+ {
1157
+ "epoch": 1.6639276910435497,
1158
+ "grad_norm": 25.41390609741211,
1159
+ "learning_rate": 4.891456414891116e-05,
1160
+ "loss": 0.7182,
1161
+ "step": 4050
1162
+ },
1163
+ {
1164
+ "epoch": 1.6741988496302382,
1165
+ "grad_norm": 19.755008697509766,
1166
+ "learning_rate": 4.888534664961062e-05,
1167
+ "loss": 0.6639,
1168
+ "step": 4075
1169
+ },
1170
+ {
1171
+ "epoch": 1.6844700082169268,
1172
+ "grad_norm": 20.488435745239258,
1173
+ "learning_rate": 4.8855750082266216e-05,
1174
+ "loss": 0.7489,
1175
+ "step": 4100
1176
+ },
1177
+ {
1178
+ "epoch": 1.6947411668036154,
1179
+ "grad_norm": 9.432578086853027,
1180
+ "learning_rate": 4.882577491658486e-05,
1181
+ "loss": 0.7505,
1182
+ "step": 4125
1183
+ },
1184
+ {
1185
+ "epoch": 1.705012325390304,
1186
+ "grad_norm": 24.474281311035156,
1187
+ "learning_rate": 4.879542162828197e-05,
1188
+ "loss": 0.6442,
1189
+ "step": 4150
1190
+ },
1191
+ {
1192
+ "epoch": 1.7152834839769926,
1193
+ "grad_norm": 22.1693058013916,
1194
+ "learning_rate": 4.876469069907388e-05,
1195
+ "loss": 0.8103,
1196
+ "step": 4175
1197
+ },
1198
+ {
1199
+ "epoch": 1.7255546425636812,
1200
+ "grad_norm": 33.30137634277344,
1201
+ "learning_rate": 4.8733582616670195e-05,
1202
+ "loss": 0.7918,
1203
+ "step": 4200
1204
+ },
1205
+ {
1206
+ "epoch": 1.7358258011503698,
1207
+ "grad_norm": 16.692672729492188,
1208
+ "learning_rate": 4.870209787476608e-05,
1209
+ "loss": 0.69,
1210
+ "step": 4225
1211
+ },
1212
+ {
1213
+ "epoch": 1.7460969597370584,
1214
+ "grad_norm": 9.02397632598877,
1215
+ "learning_rate": 4.86702369730344e-05,
1216
+ "loss": 0.6181,
1217
+ "step": 4250
1218
+ },
1219
+ {
1220
+ "epoch": 1.756368118323747,
1221
+ "grad_norm": 11.770980834960938,
1222
+ "learning_rate": 4.863800041711779e-05,
1223
+ "loss": 0.7505,
1224
+ "step": 4275
1225
+ },
1226
+ {
1227
+ "epoch": 1.7666392769104355,
1228
+ "grad_norm": 30.661998748779297,
1229
+ "learning_rate": 4.860538871862064e-05,
1230
+ "loss": 0.6744,
1231
+ "step": 4300
1232
+ },
1233
+ {
1234
+ "epoch": 1.7769104354971241,
1235
+ "grad_norm": 14.791523933410645,
1236
+ "learning_rate": 4.8572402395100966e-05,
1237
+ "loss": 0.7865,
1238
+ "step": 4325
1239
+ },
1240
+ {
1241
+ "epoch": 1.7871815940838127,
1242
+ "grad_norm": 11.105860710144043,
1243
+ "learning_rate": 4.8539041970062216e-05,
1244
+ "loss": 0.7299,
1245
+ "step": 4350
1246
+ },
1247
+ {
1248
+ "epoch": 1.7974527526705013,
1249
+ "grad_norm": 1.4959219694137573,
1250
+ "learning_rate": 4.8505307972944934e-05,
1251
+ "loss": 0.6649,
1252
+ "step": 4375
1253
+ },
1254
+ {
1255
+ "epoch": 1.80772391125719,
1256
+ "grad_norm": 38.041778564453125,
1257
+ "learning_rate": 4.8471200939118376e-05,
1258
+ "loss": 0.853,
1259
+ "step": 4400
1260
+ },
1261
+ {
1262
+ "epoch": 1.8179950698438785,
1263
+ "grad_norm": 13.44230842590332,
1264
+ "learning_rate": 4.843672140987202e-05,
1265
+ "loss": 0.6944,
1266
+ "step": 4425
1267
+ },
1268
+ {
1269
+ "epoch": 1.828266228430567,
1270
+ "grad_norm": 3.37754487991333,
1271
+ "learning_rate": 4.8401869932406946e-05,
1272
+ "loss": 0.6653,
1273
+ "step": 4450
1274
+ },
1275
+ {
1276
+ "epoch": 1.8385373870172557,
1277
+ "grad_norm": 10.962355613708496,
1278
+ "learning_rate": 4.8366647059827185e-05,
1279
+ "loss": 0.6699,
1280
+ "step": 4475
1281
+ },
1282
+ {
1283
+ "epoch": 1.8488085456039443,
1284
+ "grad_norm": 23.833568572998047,
1285
+ "learning_rate": 4.8331053351130915e-05,
1286
+ "loss": 0.7714,
1287
+ "step": 4500
1288
+ },
1289
+ {
1290
+ "epoch": 1.8590797041906328,
1291
+ "grad_norm": 20.1926212310791,
1292
+ "learning_rate": 4.829508937120162e-05,
1293
+ "loss": 0.6746,
1294
+ "step": 4525
1295
+ },
1296
+ {
1297
+ "epoch": 1.8693508627773214,
1298
+ "grad_norm": 4.717026710510254,
1299
+ "learning_rate": 4.825875569079907e-05,
1300
+ "loss": 0.764,
1301
+ "step": 4550
1302
+ },
1303
+ {
1304
+ "epoch": 1.87962202136401,
1305
+ "grad_norm": 51.555458068847656,
1306
+ "learning_rate": 4.822205288655035e-05,
1307
+ "loss": 0.5943,
1308
+ "step": 4575
1309
+ },
1310
+ {
1311
+ "epoch": 1.8898931799506984,
1312
+ "grad_norm": 17.627649307250977,
1313
+ "learning_rate": 4.818498154094062e-05,
1314
+ "loss": 0.6794,
1315
+ "step": 4600
1316
+ },
1317
+ {
1318
+ "epoch": 1.900164338537387,
1319
+ "grad_norm": 15.7213716506958,
1320
+ "learning_rate": 4.814754224230391e-05,
1321
+ "loss": 0.6973,
1322
+ "step": 4625
1323
+ },
1324
+ {
1325
+ "epoch": 1.9104354971240756,
1326
+ "grad_norm": 9.14035415649414,
1327
+ "learning_rate": 4.81097355848138e-05,
1328
+ "loss": 0.9402,
1329
+ "step": 4650
1330
+ },
1331
+ {
1332
+ "epoch": 1.9207066557107642,
1333
+ "grad_norm": 15.820075988769531,
1334
+ "learning_rate": 4.807156216847396e-05,
1335
+ "loss": 0.8003,
1336
+ "step": 4675
1337
+ },
1338
+ {
1339
+ "epoch": 1.9309778142974527,
1340
+ "grad_norm": 1.9751650094985962,
1341
+ "learning_rate": 4.803302259910863e-05,
1342
+ "loss": 0.746,
1343
+ "step": 4700
1344
+ },
1345
+ {
1346
+ "epoch": 1.9412489728841413,
1347
+ "grad_norm": 34.23078155517578,
1348
+ "learning_rate": 4.799411748835305e-05,
1349
+ "loss": 0.774,
1350
+ "step": 4725
1351
+ },
1352
+ {
1353
+ "epoch": 1.95152013147083,
1354
+ "grad_norm": 33.61397933959961,
1355
+ "learning_rate": 4.7954847453643686e-05,
1356
+ "loss": 0.8032,
1357
+ "step": 4750
1358
+ },
1359
+ {
1360
+ "epoch": 1.9617912900575185,
1361
+ "grad_norm": 22.3249454498291,
1362
+ "learning_rate": 4.7915213118208456e-05,
1363
+ "loss": 0.7198,
1364
+ "step": 4775
1365
+ },
1366
+ {
1367
+ "epoch": 1.972062448644207,
1368
+ "grad_norm": 5.324288368225098,
1369
+ "learning_rate": 4.787521511105689e-05,
1370
+ "loss": 0.7335,
1371
+ "step": 4800
1372
+ },
1373
+ {
1374
+ "epoch": 1.9823336072308957,
1375
+ "grad_norm": 7.5206990242004395,
1376
+ "learning_rate": 4.783485406697005e-05,
1377
+ "loss": 0.7861,
1378
+ "step": 4825
1379
+ },
1380
+ {
1381
+ "epoch": 1.9926047658175843,
1382
+ "grad_norm": 18.279809951782227,
1383
+ "learning_rate": 4.7794130626490544e-05,
1384
+ "loss": 0.659,
1385
+ "step": 4850
1386
+ },
1387
+ {
1388
+ "epoch": 2.0,
1389
+ "eval_accuracy": 0.661871981917189,
1390
+ "eval_f1_macro": 0.458297977909026,
1391
+ "eval_f1_micro": 0.661871981917189,
1392
+ "eval_f1_weighted": 0.6408854719266703,
1393
+ "eval_loss": 0.9106447696685791,
1394
+ "eval_precision_macro": 0.44316873127833,
1395
+ "eval_precision_micro": 0.661871981917189,
1396
+ "eval_precision_weighted": 0.6229543768525556,
1397
+ "eval_recall_macro": 0.4775160837702286,
1398
+ "eval_recall_micro": 0.661871981917189,
1399
+ "eval_recall_weighted": 0.661871981917189,
1400
+ "eval_runtime": 20.2196,
1401
+ "eval_samples_per_second": 481.365,
1402
+ "eval_steps_per_second": 15.084,
1403
+ "step": 4868
1404
+ },
1405
+ {
1406
+ "epoch": 2.0028759244042726,
1407
+ "grad_norm": 27.522146224975586,
1408
+ "learning_rate": 4.775304543591232e-05,
1409
+ "loss": 0.7247,
1410
+ "step": 4875
1411
+ },
1412
+ {
1413
+ "epoch": 2.0131470829909612,
1414
+ "grad_norm": 2.502640962600708,
1415
+ "learning_rate": 4.7711599147270416e-05,
1416
+ "loss": 0.7228,
1417
+ "step": 4900
1418
+ },
1419
+ {
1420
+ "epoch": 2.02341824157765,
1421
+ "grad_norm": 10.220170974731445,
1422
+ "learning_rate": 4.766979241833059e-05,
1423
+ "loss": 0.6397,
1424
+ "step": 4925
1425
+ },
1426
+ {
1427
+ "epoch": 2.0336894001643384,
1428
+ "grad_norm": 23.93318748474121,
1429
+ "learning_rate": 4.7627625912578945e-05,
1430
+ "loss": 0.6842,
1431
+ "step": 4950
1432
+ },
1433
+ {
1434
+ "epoch": 2.043960558751027,
1435
+ "grad_norm": 10.872408866882324,
1436
+ "learning_rate": 4.758510029921132e-05,
1437
+ "loss": 0.7494,
1438
+ "step": 4975
1439
+ },
1440
+ {
1441
+ "epoch": 2.0542317173377156,
1442
+ "grad_norm": 29.06859588623047,
1443
+ "learning_rate": 4.754221625312271e-05,
1444
+ "loss": 0.7152,
1445
+ "step": 5000
1446
+ },
1447
+ {
1448
+ "epoch": 2.064502875924404,
1449
+ "grad_norm": 13.37592887878418,
1450
+ "learning_rate": 4.749897445489656e-05,
1451
+ "loss": 0.6585,
1452
+ "step": 5025
1453
+ },
1454
+ {
1455
+ "epoch": 2.0747740345110928,
1456
+ "grad_norm": 7.18634033203125,
1457
+ "learning_rate": 4.745537559079396e-05,
1458
+ "loss": 0.6494,
1459
+ "step": 5050
1460
+ },
1461
+ {
1462
+ "epoch": 2.0850451930977814,
1463
+ "grad_norm": 5.377128601074219,
1464
+ "learning_rate": 4.741142035274274e-05,
1465
+ "loss": 0.7026,
1466
+ "step": 5075
1467
+ },
1468
+ {
1469
+ "epoch": 2.09531635168447,
1470
+ "grad_norm": 5.074624538421631,
1471
+ "learning_rate": 4.736710943832649e-05,
1472
+ "loss": 0.732,
1473
+ "step": 5100
1474
+ },
1475
+ {
1476
+ "epoch": 2.1055875102711585,
1477
+ "grad_norm": 3.903517246246338,
1478
+ "learning_rate": 4.7322443550773535e-05,
1479
+ "loss": 0.6671,
1480
+ "step": 5125
1481
+ },
1482
+ {
1483
+ "epoch": 2.115858668857847,
1484
+ "grad_norm": 21.475658416748047,
1485
+ "learning_rate": 4.727742339894567e-05,
1486
+ "loss": 0.7383,
1487
+ "step": 5150
1488
+ },
1489
+ {
1490
+ "epoch": 2.1261298274445357,
1491
+ "grad_norm": 17.559850692749023,
1492
+ "learning_rate": 4.723204969732704e-05,
1493
+ "loss": 0.7126,
1494
+ "step": 5175
1495
+ },
1496
+ {
1497
+ "epoch": 2.1364009860312243,
1498
+ "grad_norm": 3.733065366744995,
1499
+ "learning_rate": 4.7186323166012714e-05,
1500
+ "loss": 0.7779,
1501
+ "step": 5200
1502
+ },
1503
+ {
1504
+ "epoch": 2.146672144617913,
1505
+ "grad_norm": 50.46138381958008,
1506
+ "learning_rate": 4.714024453069727e-05,
1507
+ "loss": 0.6233,
1508
+ "step": 5225
1509
+ },
1510
+ {
1511
+ "epoch": 2.1569433032046015,
1512
+ "grad_norm": 56.56818771362305,
1513
+ "learning_rate": 4.7093814522663304e-05,
1514
+ "loss": 0.8527,
1515
+ "step": 5250
1516
+ },
1517
+ {
1518
+ "epoch": 2.16721446179129,
1519
+ "grad_norm": 3.0601882934570312,
1520
+ "learning_rate": 4.70470338787698e-05,
1521
+ "loss": 0.7016,
1522
+ "step": 5275
1523
+ },
1524
+ {
1525
+ "epoch": 2.1774856203779787,
1526
+ "grad_norm": 13.10061264038086,
1527
+ "learning_rate": 4.700179527152305e-05,
1528
+ "loss": 0.7397,
1529
+ "step": 5300
1530
+ },
1531
+ {
1532
+ "epoch": 2.1877567789646672,
1533
+ "grad_norm": 39.44890594482422,
1534
+ "learning_rate": 4.695432954012213e-05,
1535
+ "loss": 0.6456,
1536
+ "step": 5325
1537
+ },
1538
+ {
1539
+ "epoch": 2.198027937551356,
1540
+ "grad_norm": 22.93960952758789,
1541
+ "learning_rate": 4.690651538653268e-05,
1542
+ "loss": 0.7832,
1543
+ "step": 5350
1544
+ },
1545
+ {
1546
+ "epoch": 2.2082990961380444,
1547
+ "grad_norm": 22.193588256835938,
1548
+ "learning_rate": 4.685835356958051e-05,
1549
+ "loss": 0.6787,
1550
+ "step": 5375
1551
+ },
1552
+ {
1553
+ "epoch": 2.218570254724733,
1554
+ "grad_norm": 31.868240356445312,
1555
+ "learning_rate": 4.680984485360899e-05,
1556
+ "loss": 0.6281,
1557
+ "step": 5400
1558
+ },
1559
+ {
1560
+ "epoch": 2.2288414133114216,
1561
+ "grad_norm": 25.054807662963867,
1562
+ "learning_rate": 4.676099000846684e-05,
1563
+ "loss": 0.817,
1564
+ "step": 5425
1565
+ },
1566
+ {
1567
+ "epoch": 2.23911257189811,
1568
+ "grad_norm": 10.673225402832031,
1569
+ "learning_rate": 4.671178980949599e-05,
1570
+ "loss": 0.7448,
1571
+ "step": 5450
1572
+ },
1573
+ {
1574
+ "epoch": 2.249383730484799,
1575
+ "grad_norm": 4.617043972015381,
1576
+ "learning_rate": 4.666224503751923e-05,
1577
+ "loss": 0.7868,
1578
+ "step": 5475
1579
+ },
1580
+ {
1581
+ "epoch": 2.2596548890714874,
1582
+ "grad_norm": 16.736486434936523,
1583
+ "learning_rate": 4.661235647882785e-05,
1584
+ "loss": 0.7392,
1585
+ "step": 5500
1586
+ },
1587
+ {
1588
+ "epoch": 2.269926047658176,
1589
+ "grad_norm": 143.8787078857422,
1590
+ "learning_rate": 4.6562124925169126e-05,
1591
+ "loss": 0.7608,
1592
+ "step": 5525
1593
+ },
1594
+ {
1595
+ "epoch": 2.2801972062448645,
1596
+ "grad_norm": 12.933761596679688,
1597
+ "learning_rate": 4.651155117373378e-05,
1598
+ "loss": 0.7637,
1599
+ "step": 5550
1600
+ },
1601
+ {
1602
+ "epoch": 2.290468364831553,
1603
+ "grad_norm": 19.505279541015625,
1604
+ "learning_rate": 4.646063602714334e-05,
1605
+ "loss": 0.72,
1606
+ "step": 5575
1607
+ },
1608
+ {
1609
+ "epoch": 2.3007395234182417,
1610
+ "grad_norm": 18.15628433227539,
1611
+ "learning_rate": 4.6409380293437355e-05,
1612
+ "loss": 0.7434,
1613
+ "step": 5600
1614
+ },
1615
+ {
1616
+ "epoch": 2.3110106820049303,
1617
+ "grad_norm": 184.67726135253906,
1618
+ "learning_rate": 4.635778478606061e-05,
1619
+ "loss": 0.6824,
1620
+ "step": 5625
1621
+ },
1622
+ {
1623
+ "epoch": 2.321281840591619,
1624
+ "grad_norm": 15.403681755065918,
1625
+ "learning_rate": 4.6305850323850216e-05,
1626
+ "loss": 0.7953,
1627
+ "step": 5650
1628
+ },
1629
+ {
1630
+ "epoch": 2.3315529991783075,
1631
+ "grad_norm": 22.72369956970215,
1632
+ "learning_rate": 4.625357773102259e-05,
1633
+ "loss": 0.7348,
1634
+ "step": 5675
1635
+ },
1636
+ {
1637
+ "epoch": 2.341824157764996,
1638
+ "grad_norm": 9.090083122253418,
1639
+ "learning_rate": 4.62009678371604e-05,
1640
+ "loss": 0.6719,
1641
+ "step": 5700
1642
+ },
1643
+ {
1644
+ "epoch": 2.3520953163516847,
1645
+ "grad_norm": 13.648232460021973,
1646
+ "learning_rate": 4.6148021477199386e-05,
1647
+ "loss": 0.7489,
1648
+ "step": 5725
1649
+ },
1650
+ {
1651
+ "epoch": 2.362366474938373,
1652
+ "grad_norm": 24.75493621826172,
1653
+ "learning_rate": 4.60947394914151e-05,
1654
+ "loss": 0.7936,
1655
+ "step": 5750
1656
+ },
1657
+ {
1658
+ "epoch": 2.372637633525062,
1659
+ "grad_norm": 15.41668701171875,
1660
+ "learning_rate": 4.604112272540961e-05,
1661
+ "loss": 0.7393,
1662
+ "step": 5775
1663
+ },
1664
+ {
1665
+ "epoch": 2.38290879211175,
1666
+ "grad_norm": 20.62226104736328,
1667
+ "learning_rate": 4.5987172030098024e-05,
1668
+ "loss": 0.6496,
1669
+ "step": 5800
1670
+ },
1671
+ {
1672
+ "epoch": 2.393179950698439,
1673
+ "grad_norm": 29.890432357788086,
1674
+ "learning_rate": 4.5932888261695026e-05,
1675
+ "loss": 0.7946,
1676
+ "step": 5825
1677
+ },
1678
+ {
1679
+ "epoch": 2.403451109285127,
1680
+ "grad_norm": 45.65409851074219,
1681
+ "learning_rate": 4.587827228170129e-05,
1682
+ "loss": 0.668,
1683
+ "step": 5850
1684
+ },
1685
+ {
1686
+ "epoch": 2.4137222678718158,
1687
+ "grad_norm": 23.496747970581055,
1688
+ "learning_rate": 4.5823324956889764e-05,
1689
+ "loss": 0.7671,
1690
+ "step": 5875
1691
+ },
1692
+ {
1693
+ "epoch": 2.4239934264585044,
1694
+ "grad_norm": 107.06421661376953,
1695
+ "learning_rate": 4.576804715929196e-05,
1696
+ "loss": 0.7784,
1697
+ "step": 5900
1698
+ },
1699
+ {
1700
+ "epoch": 2.434264585045193,
1701
+ "grad_norm": 24.115577697753906,
1702
+ "learning_rate": 4.571243976618411e-05,
1703
+ "loss": 0.6693,
1704
+ "step": 5925
1705
+ },
1706
+ {
1707
+ "epoch": 2.4445357436318815,
1708
+ "grad_norm": 7.726546287536621,
1709
+ "learning_rate": 4.56565036600732e-05,
1710
+ "loss": 0.7062,
1711
+ "step": 5950
1712
+ },
1713
+ {
1714
+ "epoch": 2.45480690221857,
1715
+ "grad_norm": 8.461091041564941,
1716
+ "learning_rate": 4.5600239728683035e-05,
1717
+ "loss": 0.7455,
1718
+ "step": 5975
1719
+ },
1720
+ {
1721
+ "epoch": 2.4650780608052587,
1722
+ "grad_norm": 19.741405487060547,
1723
+ "learning_rate": 4.554364886494008e-05,
1724
+ "loss": 0.7246,
1725
+ "step": 6000
1726
+ },
1727
+ {
1728
+ "epoch": 2.4753492193919473,
1729
+ "grad_norm": 15.383966445922852,
1730
+ "learning_rate": 4.548673196695934e-05,
1731
+ "loss": 0.6792,
1732
+ "step": 6025
1733
+ },
1734
+ {
1735
+ "epoch": 2.485620377978636,
1736
+ "grad_norm": 31.002105712890625,
1737
+ "learning_rate": 4.5429489938030055e-05,
1738
+ "loss": 0.6649,
1739
+ "step": 6050
1740
+ },
1741
+ {
1742
+ "epoch": 2.4958915365653245,
1743
+ "grad_norm": 20.87677001953125,
1744
+ "learning_rate": 4.5371923686601446e-05,
1745
+ "loss": 0.8137,
1746
+ "step": 6075
1747
+ },
1748
+ {
1749
+ "epoch": 2.506162695152013,
1750
+ "grad_norm": 6.823126792907715,
1751
+ "learning_rate": 4.531403412626821e-05,
1752
+ "loss": 0.7345,
1753
+ "step": 6100
1754
+ },
1755
+ {
1756
+ "epoch": 2.5164338537387017,
1757
+ "grad_norm": 10.97050952911377,
1758
+ "learning_rate": 4.525582217575607e-05,
1759
+ "loss": 0.662,
1760
+ "step": 6125
1761
+ },
1762
+ {
1763
+ "epoch": 2.5267050123253902,
1764
+ "grad_norm": 23.480966567993164,
1765
+ "learning_rate": 4.5197288758907194e-05,
1766
+ "loss": 0.7744,
1767
+ "step": 6150
1768
+ },
1769
+ {
1770
+ "epoch": 2.536976170912079,
1771
+ "grad_norm": 71.35833740234375,
1772
+ "learning_rate": 4.5138434804665505e-05,
1773
+ "loss": 0.6405,
1774
+ "step": 6175
1775
+ },
1776
+ {
1777
+ "epoch": 2.5472473294987674,
1778
+ "grad_norm": 2.59698486328125,
1779
+ "learning_rate": 4.5079261247061974e-05,
1780
+ "loss": 0.7056,
1781
+ "step": 6200
1782
+ },
1783
+ {
1784
+ "epoch": 2.557518488085456,
1785
+ "grad_norm": 13.829898834228516,
1786
+ "learning_rate": 4.501976902519975e-05,
1787
+ "loss": 0.7639,
1788
+ "step": 6225
1789
+ },
1790
+ {
1791
+ "epoch": 2.5677896466721446,
1792
+ "grad_norm": 32.521240234375,
1793
+ "learning_rate": 4.4959959083239335e-05,
1794
+ "loss": 0.711,
1795
+ "step": 6250
1796
+ },
1797
+ {
1798
+ "epoch": 2.578060805258833,
1799
+ "grad_norm": 9.438847541809082,
1800
+ "learning_rate": 4.489983237038349e-05,
1801
+ "loss": 0.7104,
1802
+ "step": 6275
1803
+ },
1804
+ {
1805
+ "epoch": 2.5883319638455218,
1806
+ "grad_norm": 9.817651748657227,
1807
+ "learning_rate": 4.483938984086228e-05,
1808
+ "loss": 0.7139,
1809
+ "step": 6300
1810
+ },
1811
+ {
1812
+ "epoch": 2.5986031224322104,
1813
+ "grad_norm": 8.208524703979492,
1814
+ "learning_rate": 4.4778632453917854e-05,
1815
+ "loss": 0.7199,
1816
+ "step": 6325
1817
+ },
1818
+ {
1819
+ "epoch": 2.608874281018899,
1820
+ "grad_norm": 20.00709342956543,
1821
+ "learning_rate": 4.471756117378926e-05,
1822
+ "loss": 0.7504,
1823
+ "step": 6350
1824
+ },
1825
+ {
1826
+ "epoch": 2.6191454396055875,
1827
+ "grad_norm": 8.180582046508789,
1828
+ "learning_rate": 4.4656176969697123e-05,
1829
+ "loss": 0.7396,
1830
+ "step": 6375
1831
+ },
1832
+ {
1833
+ "epoch": 2.629416598192276,
1834
+ "grad_norm": 21.006549835205078,
1835
+ "learning_rate": 4.4594480815828274e-05,
1836
+ "loss": 0.767,
1837
+ "step": 6400
1838
+ },
1839
+ {
1840
+ "epoch": 2.6396877567789647,
1841
+ "grad_norm": 9.309197425842285,
1842
+ "learning_rate": 4.45324736913203e-05,
1843
+ "loss": 0.7447,
1844
+ "step": 6425
1845
+ },
1846
+ {
1847
+ "epoch": 2.6499589153656533,
1848
+ "grad_norm": 47.43417739868164,
1849
+ "learning_rate": 4.447015658024595e-05,
1850
+ "loss": 0.7457,
1851
+ "step": 6450
1852
+ },
1853
+ {
1854
+ "epoch": 2.660230073952342,
1855
+ "grad_norm": 14.902249336242676,
1856
+ "learning_rate": 4.4407530471597614e-05,
1857
+ "loss": 0.652,
1858
+ "step": 6475
1859
+ },
1860
+ {
1861
+ "epoch": 2.6705012325390305,
1862
+ "grad_norm": 16.622949600219727,
1863
+ "learning_rate": 4.434459635927152e-05,
1864
+ "loss": 0.6759,
1865
+ "step": 6500
1866
+ },
1867
+ {
1868
+ "epoch": 2.680772391125719,
1869
+ "grad_norm": 45.0763053894043,
1870
+ "learning_rate": 4.428135524205206e-05,
1871
+ "loss": 0.6605,
1872
+ "step": 6525
1873
+ },
1874
+ {
1875
+ "epoch": 2.6910435497124077,
1876
+ "grad_norm": 160.1131134033203,
1877
+ "learning_rate": 4.4217808123595846e-05,
1878
+ "loss": 0.7286,
1879
+ "step": 6550
1880
+ },
1881
+ {
1882
+ "epoch": 2.7013147082990963,
1883
+ "grad_norm": 20.303207397460938,
1884
+ "learning_rate": 4.415395601241586e-05,
1885
+ "loss": 0.7143,
1886
+ "step": 6575
1887
+ },
1888
+ {
1889
+ "epoch": 2.711585866885785,
1890
+ "grad_norm": 8.981250762939453,
1891
+ "learning_rate": 4.408979992186539e-05,
1892
+ "loss": 0.7694,
1893
+ "step": 6600
1894
+ },
1895
+ {
1896
+ "epoch": 2.7218570254724734,
1897
+ "grad_norm": 8.836969375610352,
1898
+ "learning_rate": 4.402534087012201e-05,
1899
+ "loss": 0.7134,
1900
+ "step": 6625
1901
+ },
1902
+ {
1903
+ "epoch": 2.732128184059162,
1904
+ "grad_norm": 1.7732627391815186,
1905
+ "learning_rate": 4.3960579880171346e-05,
1906
+ "loss": 0.6537,
1907
+ "step": 6650
1908
+ },
1909
+ {
1910
+ "epoch": 2.7423993426458506,
1911
+ "grad_norm": 22.986120223999023,
1912
+ "learning_rate": 4.389551797979089e-05,
1913
+ "loss": 0.7382,
1914
+ "step": 6675
1915
+ },
1916
+ {
1917
+ "epoch": 2.752670501232539,
1918
+ "grad_norm": 7.212994575500488,
1919
+ "learning_rate": 4.383015620153369e-05,
1920
+ "loss": 0.6945,
1921
+ "step": 6700
1922
+ },
1923
+ {
1924
+ "epoch": 2.7629416598192273,
1925
+ "grad_norm": 9.178274154663086,
1926
+ "learning_rate": 4.376449558271194e-05,
1927
+ "loss": 0.8691,
1928
+ "step": 6725
1929
+ },
1930
+ {
1931
+ "epoch": 2.7732128184059164,
1932
+ "grad_norm": 22.552732467651367,
1933
+ "learning_rate": 4.369853716538053e-05,
1934
+ "loss": 0.7417,
1935
+ "step": 6750
1936
+ },
1937
+ {
1938
+ "epoch": 2.7834839769926045,
1939
+ "grad_norm": 18.054012298583984,
1940
+ "learning_rate": 4.3632281996320526e-05,
1941
+ "loss": 0.7231,
1942
+ "step": 6775
1943
+ },
1944
+ {
1945
+ "epoch": 2.7937551355792936,
1946
+ "grad_norm": 20.407691955566406,
1947
+ "learning_rate": 4.3565731127022504e-05,
1948
+ "loss": 0.6774,
1949
+ "step": 6800
1950
+ },
1951
+ {
1952
+ "epoch": 2.8040262941659817,
1953
+ "grad_norm": 12.27515983581543,
1954
+ "learning_rate": 4.349888561366991e-05,
1955
+ "loss": 0.7521,
1956
+ "step": 6825
1957
+ },
1958
+ {
1959
+ "epoch": 2.8142974527526707,
1960
+ "grad_norm": 32.21310806274414,
1961
+ "learning_rate": 4.343174651712232e-05,
1962
+ "loss": 0.6581,
1963
+ "step": 6850
1964
+ },
1965
+ {
1966
+ "epoch": 2.824568611339359,
1967
+ "grad_norm": 19.447738647460938,
1968
+ "learning_rate": 4.33643149028985e-05,
1969
+ "loss": 0.73,
1970
+ "step": 6875
1971
+ },
1972
+ {
1973
+ "epoch": 2.834839769926048,
1974
+ "grad_norm": 6.433927536010742,
1975
+ "learning_rate": 4.329659184115963e-05,
1976
+ "loss": 0.8247,
1977
+ "step": 6900
1978
+ },
1979
+ {
1980
+ "epoch": 2.845110928512736,
1981
+ "grad_norm": 21.21660804748535,
1982
+ "learning_rate": 4.322857840669222e-05,
1983
+ "loss": 0.7508,
1984
+ "step": 6925
1985
+ },
1986
+ {
1987
+ "epoch": 2.855382087099425,
1988
+ "grad_norm": 4.219921588897705,
1989
+ "learning_rate": 4.3160275678891085e-05,
1990
+ "loss": 0.7135,
1991
+ "step": 6950
1992
+ },
1993
+ {
1994
+ "epoch": 2.8656532456861132,
1995
+ "grad_norm": 12.678546905517578,
1996
+ "learning_rate": 4.309168474174224e-05,
1997
+ "loss": 0.7757,
1998
+ "step": 6975
1999
+ },
2000
+ {
2001
+ "epoch": 2.875924404272802,
2002
+ "grad_norm": 10.19421100616455,
2003
+ "learning_rate": 4.3022806683805653e-05,
2004
+ "loss": 0.6503,
2005
+ "step": 7000
2006
+ },
2007
+ {
2008
+ "epoch": 2.8861955628594904,
2009
+ "grad_norm": 6.024997234344482,
2010
+ "learning_rate": 4.2953642598198e-05,
2011
+ "loss": 0.7013,
2012
+ "step": 7025
2013
+ },
2014
+ {
2015
+ "epoch": 2.896466721446179,
2016
+ "grad_norm": 45.83235549926758,
2017
+ "learning_rate": 4.288419358257531e-05,
2018
+ "loss": 0.7507,
2019
+ "step": 7050
2020
+ },
2021
+ {
2022
+ "epoch": 2.9067378800328676,
2023
+ "grad_norm": 13.147794723510742,
2024
+ "learning_rate": 4.281446073911553e-05,
2025
+ "loss": 0.6307,
2026
+ "step": 7075
2027
+ },
2028
+ {
2029
+ "epoch": 2.917009038619556,
2030
+ "grad_norm": 26.889511108398438,
2031
+ "learning_rate": 4.2744445174501045e-05,
2032
+ "loss": 0.667,
2033
+ "step": 7100
2034
+ },
2035
+ {
2036
+ "epoch": 2.9272801972062448,
2037
+ "grad_norm": 11.249592781066895,
2038
+ "learning_rate": 4.2674147999901144e-05,
2039
+ "loss": 0.7892,
2040
+ "step": 7125
2041
+ },
2042
+ {
2043
+ "epoch": 2.9375513557929334,
2044
+ "grad_norm": 8.001336097717285,
2045
+ "learning_rate": 4.260357033095431e-05,
2046
+ "loss": 0.6859,
2047
+ "step": 7150
2048
+ },
2049
+ {
2050
+ "epoch": 2.947822514379622,
2051
+ "grad_norm": 21.009946823120117,
2052
+ "learning_rate": 4.25327132877506e-05,
2053
+ "loss": 0.7299,
2054
+ "step": 7175
2055
+ },
2056
+ {
2057
+ "epoch": 2.9580936729663105,
2058
+ "grad_norm": 14.191951751708984,
2059
+ "learning_rate": 4.2461577994813814e-05,
2060
+ "loss": 0.6725,
2061
+ "step": 7200
2062
+ },
2063
+ {
2064
+ "epoch": 2.968364831552999,
2065
+ "grad_norm": 6.6921916007995605,
2066
+ "learning_rate": 4.2390165581083654e-05,
2067
+ "loss": 0.6824,
2068
+ "step": 7225
2069
+ },
2070
+ {
2071
+ "epoch": 2.9786359901396877,
2072
+ "grad_norm": 39.17061233520508,
2073
+ "learning_rate": 4.231847717989782e-05,
2074
+ "loss": 0.7409,
2075
+ "step": 7250
2076
+ },
2077
+ {
2078
+ "epoch": 2.9889071487263763,
2079
+ "grad_norm": 31.245113372802734,
2080
+ "learning_rate": 4.224651392897404e-05,
2081
+ "loss": 0.7009,
2082
+ "step": 7275
2083
+ },
2084
+ {
2085
+ "epoch": 2.999178307313065,
2086
+ "grad_norm": 17.3699893951416,
2087
+ "learning_rate": 4.217427697039198e-05,
2088
+ "loss": 0.7931,
2089
+ "step": 7300
2090
+ },
2091
+ {
2092
+ "epoch": 3.0,
2093
+ "eval_accuracy": 0.6470769546902291,
2094
+ "eval_f1_macro": 0.48909871800127924,
2095
+ "eval_f1_micro": 0.6470769546902291,
2096
+ "eval_f1_weighted": 0.6468481218149891,
2097
+ "eval_loss": 0.8904216289520264,
2098
+ "eval_precision_macro": 0.5045882669843967,
2099
+ "eval_precision_micro": 0.6470769546902291,
2100
+ "eval_precision_weighted": 0.6554944836574212,
2101
+ "eval_recall_macro": 0.49646946445535234,
2102
+ "eval_recall_micro": 0.6470769546902291,
2103
+ "eval_recall_weighted": 0.6470769546902291,
2104
+ "eval_runtime": 20.0256,
2105
+ "eval_samples_per_second": 486.028,
2106
+ "eval_steps_per_second": 15.231,
2107
+ "step": 7302
2108
+ }
2109
+ ],
2110
+ "logging_steps": 25,
2111
+ "max_steps": 21906,
2112
+ "num_input_tokens_seen": 0,
2113
+ "num_train_epochs": 9,
2114
+ "save_steps": 500,
2115
+ "stateful_callbacks": {
2116
+ "EarlyStoppingCallback": {
2117
+ "args": {
2118
+ "early_stopping_patience": 5,
2119
+ "early_stopping_threshold": 0.01
2120
+ },
2121
+ "attributes": {
2122
+ "early_stopping_patience_counter": 0
2123
+ }
2124
+ },
2125
+ "TrainerControl": {
2126
+ "args": {
2127
+ "should_epoch_stop": false,
2128
+ "should_evaluate": false,
2129
+ "should_log": false,
2130
+ "should_save": true,
2131
+ "should_training_stop": false
2132
+ },
2133
+ "attributes": {}
2134
+ }
2135
+ },
2136
+ "total_flos": 1.5370309738954752e+16,
2137
+ "train_batch_size": 8,
2138
+ "trial_name": null,
2139
+ "trial_params": null
2140
+ }
checkpoint-7302/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7b3846c407f453ea89f497122124bfe218afe31e6685d5602147a195a385fd0f
3
+ size 5368
config.json ADDED
@@ -0,0 +1,49 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "ivanovsdesign/huawei-data-classification",
3
+ "_num_labels": 5,
4
+ "architectures": [
5
+ "BertForSequenceClassification"
6
+ ],
7
+ "attention_probs_dropout_prob": 0.1,
8
+ "classifier_dropout": null,
9
+ "directionality": "bidi",
10
+ "gradient_checkpointing": false,
11
+ "hidden_act": "gelu",
12
+ "hidden_dropout_prob": 0.1,
13
+ "hidden_size": 768,
14
+ "id2label": {
15
+ "0": "1",
16
+ "1": "2",
17
+ "2": "3",
18
+ "3": "4",
19
+ "4": "5"
20
+ },
21
+ "initializer_range": 0.02,
22
+ "intermediate_size": 3072,
23
+ "label2id": {
24
+ "1": 0,
25
+ "2": 1,
26
+ "3": 2,
27
+ "4": 3,
28
+ "5": 4
29
+ },
30
+ "layer_norm_eps": 1e-12,
31
+ "max_position_embeddings": 512,
32
+ "model_type": "bert",
33
+ "num_attention_heads": 12,
34
+ "num_hidden_layers": 12,
35
+ "output_past": true,
36
+ "pad_token_id": 0,
37
+ "pooler_fc_size": 768,
38
+ "pooler_num_attention_heads": 12,
39
+ "pooler_num_fc_layers": 3,
40
+ "pooler_size_per_head": 128,
41
+ "pooler_type": "first_token_transform",
42
+ "position_embedding_type": "absolute",
43
+ "problem_type": "single_label_classification",
44
+ "torch_dtype": "float32",
45
+ "transformers_version": "4.46.2",
46
+ "type_vocab_size": 2,
47
+ "use_cache": true,
48
+ "vocab_size": 119547
49
+ }
model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:1482e954dc7bbef048c49c06bdc65118b9138ec4ca7e91f7933627bf7f2df339
3
+ size 711452684
runs/Nov09_23-13-04_6604152ce143/events.out.tfevents.1731193985.6604152ce143.674.0 CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:b4f554538e5db7e9ca455e001ed0b47b55283ef14116b5599db2bc05caaefc38
3
- size 156759
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b3c1d8a27a5a46e496f1d77ba07a257ae833f32c1b63e295e178a2f7102f67ee
3
+ size 177317
runs/Nov09_23-13-04_6604152ce143/events.out.tfevents.1731197419.6604152ce143.674.1 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e92665de491602d132478b126e2841223ced61698002bbafebadd32a25da6538
3
+ size 936
special_tokens_map.json ADDED
@@ -0,0 +1,37 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "cls_token": {
3
+ "content": "[CLS]",
4
+ "lstrip": false,
5
+ "normalized": false,
6
+ "rstrip": false,
7
+ "single_word": false
8
+ },
9
+ "mask_token": {
10
+ "content": "[MASK]",
11
+ "lstrip": false,
12
+ "normalized": false,
13
+ "rstrip": false,
14
+ "single_word": false
15
+ },
16
+ "pad_token": {
17
+ "content": "[PAD]",
18
+ "lstrip": false,
19
+ "normalized": false,
20
+ "rstrip": false,
21
+ "single_word": false
22
+ },
23
+ "sep_token": {
24
+ "content": "[SEP]",
25
+ "lstrip": false,
26
+ "normalized": false,
27
+ "rstrip": false,
28
+ "single_word": false
29
+ },
30
+ "unk_token": {
31
+ "content": "[UNK]",
32
+ "lstrip": false,
33
+ "normalized": false,
34
+ "rstrip": false,
35
+ "single_word": false
36
+ }
37
+ }
tokenizer.json ADDED
The diff for this file is too large to render. See raw diff
 
tokenizer_config.json ADDED
@@ -0,0 +1,64 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "added_tokens_decoder": {
3
+ "0": {
4
+ "content": "[PAD]",
5
+ "lstrip": false,
6
+ "normalized": false,
7
+ "rstrip": false,
8
+ "single_word": false,
9
+ "special": true
10
+ },
11
+ "100": {
12
+ "content": "[UNK]",
13
+ "lstrip": false,
14
+ "normalized": false,
15
+ "rstrip": false,
16
+ "single_word": false,
17
+ "special": true
18
+ },
19
+ "101": {
20
+ "content": "[CLS]",
21
+ "lstrip": false,
22
+ "normalized": false,
23
+ "rstrip": false,
24
+ "single_word": false,
25
+ "special": true
26
+ },
27
+ "102": {
28
+ "content": "[SEP]",
29
+ "lstrip": false,
30
+ "normalized": false,
31
+ "rstrip": false,
32
+ "single_word": false,
33
+ "special": true
34
+ },
35
+ "103": {
36
+ "content": "[MASK]",
37
+ "lstrip": false,
38
+ "normalized": false,
39
+ "rstrip": false,
40
+ "single_word": false,
41
+ "special": true
42
+ }
43
+ },
44
+ "clean_up_tokenization_spaces": true,
45
+ "cls_token": "[CLS]",
46
+ "do_basic_tokenize": true,
47
+ "do_lower_case": false,
48
+ "mask_token": "[MASK]",
49
+ "max_length": 256,
50
+ "model_max_length": 1000000000000000019884624838656,
51
+ "never_split": null,
52
+ "pad_to_multiple_of": null,
53
+ "pad_token": "[PAD]",
54
+ "pad_token_type_id": 0,
55
+ "padding_side": "right",
56
+ "sep_token": "[SEP]",
57
+ "stride": 0,
58
+ "strip_accents": null,
59
+ "tokenize_chinese_chars": true,
60
+ "tokenizer_class": "BertTokenizer",
61
+ "truncation_side": "right",
62
+ "truncation_strategy": "longest_first",
63
+ "unk_token": "[UNK]"
64
+ }
training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7b3846c407f453ea89f497122124bfe218afe31e6685d5602147a195a385fd0f
3
+ size 5368
training_params.json ADDED
@@ -0,0 +1,30 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "data_path": "rubert-base-ru-huawei-sentiment-fine-up/autotrain-data",
3
+ "model": "ivanovsdesign/huawei-data-classification",
4
+ "lr": 5e-05,
5
+ "epochs": 9,
6
+ "max_seq_length": 256,
7
+ "batch_size": 8,
8
+ "warmup_ratio": 0.1,
9
+ "gradient_accumulation": 1,
10
+ "optimizer": "adamw_torch",
11
+ "scheduler": "cosine_with_restarts",
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": "rubert-base-ru-huawei-sentiment-fine-up",
21
+ "auto_find_batch_size": false,
22
+ "mixed_precision": "fp16",
23
+ "save_total_limit": 1,
24
+ "push_to_hub": true,
25
+ "eval_strategy": "epoch",
26
+ "username": "ivanovsdesign",
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