Hugging Face's logo

dhananjay2912
/
deberta_aci_bench_medical_section_classifier

Text Classification
Transformers
TensorBoard
Safetensors
deberta-v2
Trained with AutoTrain
text-embeddings-inference
Model card Files
xet
 Metrics Community
dhananjay2912 commited on
Commit
2f889a9
·
verified ·
1 Parent(s): e6ad748

Upload folder using huggingface_hub

Browse files
Files changed (19) hide show
  1. README.md +37 -0
  2. added_tokens.json +3 -0
  3. checkpoint-453/config.json +80 -0
  4. checkpoint-453/model.safetensors +3 -0
  5. checkpoint-453/optimizer.pt +3 -0
  6. checkpoint-453/rng_state.pth +3 -0
  7. checkpoint-453/scheduler.pt +3 -0
  8. checkpoint-453/trainer_state.json +1657 -0
  9. checkpoint-453/training_args.bin +3 -0
  10. config.json +80 -0
  11. model.safetensors +3 -0
  12. runs/Apr28_03-13-55_r-dhananjay2912-deberta-section-classifier-el2khqmv-8c8a0-vi3xw/events.out.tfevents.1714274035.r-dhananjay2912-deberta-section-classifier-el2khqmv-8c8a0-vi3xw.61.0 +2 -2
  13. runs/Apr28_03-13-55_r-dhananjay2912-deberta-section-classifier-el2khqmv-8c8a0-vi3xw/events.out.tfevents.1714274127.r-dhananjay2912-deberta-section-classifier-el2khqmv-8c8a0-vi3xw.61.1 +3 -0
  14. special_tokens_map.json +15 -0
  15. spm.model +3 -0
  16. tokenizer.json +0 -0
  17. tokenizer_config.json +58 -0
  18. training_args.bin +3 -0
  19. training_params.json +28 -0
README.md ADDED
@@ -0,0 +1,37 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+
2
+ ---
3
+ tags:
4
+ - autotrain
5
+ - text-classification
6
+ widget:
7
+ - text: "I love AutoTrain"
8
+ datasets:
9
+ - autotrain-9c20u-twasm/autotrain-data
10
+ ---
11
+
12
+ # Model Trained Using AutoTrain
13
+
14
+ - Problem type: Text Classification
15
+
16
+ ## Validation Metrics
17
+ loss: 1.4913766384124756
18
+
19
+ f1_macro: 0.28164367547346275
20
+
21
+ f1_micro: 0.64
22
+
23
+ f1_weighted: 0.5917376665887304
24
+
25
+ precision_macro: 0.2705775014459225
26
+
27
+ precision_micro: 0.64
28
+
29
+ precision_weighted: 0.5802396761133604
30
+
31
+ recall_macro: 0.3324350649350649
32
+
33
+ recall_micro: 0.64
34
+
35
+ recall_weighted: 0.64
36
+
37
+ accuracy: 0.64
added_tokens.json ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ {
2
+ "[MASK]": 128000
3
+ }
checkpoint-453/config.json ADDED
@@ -0,0 +1,80 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "microsoft/deberta-v3-base",
3
+ "_num_labels": 20,
4
+ "architectures": [
5
+ "DebertaV2ForSequenceClassification"
6
+ ],
7
+ "attention_probs_dropout_prob": 0.1,
8
+ "hidden_act": "gelu",
9
+ "hidden_dropout_prob": 0.1,
10
+ "hidden_size": 768,
11
+ "id2label": {
12
+ "0": "ALLERGY",
13
+ "1": "ASSESSMENT",
14
+ "2": "CC",
15
+ "3": "DIAGNOSIS",
16
+ "4": "DISPOSITION",
17
+ "5": "EDCOURSE",
18
+ "6": "EXAM",
19
+ "7": "FAM/SOCHX",
20
+ "8": "GENHX",
21
+ "9": "GYNHX",
22
+ "10": "IMAGING",
23
+ "11": "IMMUNIZATIONS",
24
+ "12": "LABS",
25
+ "13": "MEDICATIONS",
26
+ "14": "OTHER_HISTORY",
27
+ "15": "PASTMEDICALHX",
28
+ "16": "PASTSURGICAL",
29
+ "17": "PLAN",
30
+ "18": "PROCEDURES",
31
+ "19": "ROS"
32
+ },
33
+ "initializer_range": 0.02,
34
+ "intermediate_size": 3072,
35
+ "label2id": {
36
+ "ALLERGY": 0,
37
+ "ASSESSMENT": 1,
38
+ "CC": 2,
39
+ "DIAGNOSIS": 3,
40
+ "DISPOSITION": 4,
41
+ "EDCOURSE": 5,
42
+ "EXAM": 6,
43
+ "FAM/SOCHX": 7,
44
+ "GENHX": 8,
45
+ "GYNHX": 9,
46
+ "IMAGING": 10,
47
+ "IMMUNIZATIONS": 11,
48
+ "LABS": 12,
49
+ "MEDICATIONS": 13,
50
+ "OTHER_HISTORY": 14,
51
+ "PASTMEDICALHX": 15,
52
+ "PASTSURGICAL": 16,
53
+ "PLAN": 17,
54
+ "PROCEDURES": 18,
55
+ "ROS": 19
56
+ },
57
+ "layer_norm_eps": 1e-07,
58
+ "max_position_embeddings": 512,
59
+ "max_relative_positions": -1,
60
+ "model_type": "deberta-v2",
61
+ "norm_rel_ebd": "layer_norm",
62
+ "num_attention_heads": 12,
63
+ "num_hidden_layers": 12,
64
+ "pad_token_id": 0,
65
+ "pooler_dropout": 0,
66
+ "pooler_hidden_act": "gelu",
67
+ "pooler_hidden_size": 768,
68
+ "pos_att_type": [
69
+ "p2c",
70
+ "c2p"
71
+ ],
72
+ "position_biased_input": false,
73
+ "position_buckets": 256,
74
+ "relative_attention": true,
75
+ "share_att_key": true,
76
+ "torch_dtype": "float32",
77
+ "transformers_version": "4.40.1",
78
+ "type_vocab_size": 0,
79
+ "vocab_size": 128100
80
+ }
checkpoint-453/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9978b903ae04e38477cfbe4c26b677f26202043973af526dd217734af02f25a8
3
+ size 737774648
checkpoint-453/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:1b68088eab00b4814fe3f91902478e18eec2c80f743e1c211a3e8d1b1f9724b2
3
+ size 1475669114
checkpoint-453/rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3634414f112d61a01d7992b90ef33013ff7b8fc254dc6fa0804b27050b7420e7
3
+ size 14244
checkpoint-453/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ba9abac2a9ca38a6ee08487e65596abd4f2e71ca0d947051295b5380a7e2cca2
3
+ size 1064
checkpoint-453/trainer_state.json ADDED
@@ -0,0 +1,1657 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": 1.4913766384124756,
3
+ "best_model_checkpoint": "autotrain-9c20u-twasm/checkpoint-453",
4
+ "epoch": 3.0,
5
+ "eval_steps": 500,
6
+ "global_step": 453,
7
+ "is_hyper_param_search": false,
8
+ "is_local_process_zero": true,
9
+ "is_world_process_zero": true,
10
+ "log_history": [
11
+ {
12
+ "epoch": 0.013245033112582781,
13
+ "grad_norm": 3.3913655281066895,
14
+ "learning_rate": 2.173913043478261e-06,
15
+ "loss": 2.9946,
16
+ "step": 2
17
+ },
18
+ {
19
+ "epoch": 0.026490066225165563,
20
+ "grad_norm": 3.8051860332489014,
21
+ "learning_rate": 4.347826086956522e-06,
22
+ "loss": 3.0636,
23
+ "step": 4
24
+ },
25
+ {
26
+ "epoch": 0.039735099337748346,
27
+ "grad_norm": 4.160909652709961,
28
+ "learning_rate": 6.521739130434783e-06,
29
+ "loss": 3.0633,
30
+ "step": 6
31
+ },
32
+ {
33
+ "epoch": 0.052980132450331126,
34
+ "grad_norm": 3.9198436737060547,
35
+ "learning_rate": 8.695652173913044e-06,
36
+ "loss": 3.0431,
37
+ "step": 8
38
+ },
39
+ {
40
+ "epoch": 0.06622516556291391,
41
+ "grad_norm": 3.415198564529419,
42
+ "learning_rate": 1.0869565217391305e-05,
43
+ "loss": 3.0373,
44
+ "step": 10
45
+ },
46
+ {
47
+ "epoch": 0.07947019867549669,
48
+ "grad_norm": 4.432325839996338,
49
+ "learning_rate": 1.3043478260869566e-05,
50
+ "loss": 3.0171,
51
+ "step": 12
52
+ },
53
+ {
54
+ "epoch": 0.09271523178807947,
55
+ "grad_norm": 3.669222354888916,
56
+ "learning_rate": 1.5217391304347828e-05,
57
+ "loss": 3.0287,
58
+ "step": 14
59
+ },
60
+ {
61
+ "epoch": 0.10596026490066225,
62
+ "grad_norm": 4.237015724182129,
63
+ "learning_rate": 1.739130434782609e-05,
64
+ "loss": 2.9171,
65
+ "step": 16
66
+ },
67
+ {
68
+ "epoch": 0.11920529801324503,
69
+ "grad_norm": 4.314387798309326,
70
+ "learning_rate": 1.956521739130435e-05,
71
+ "loss": 2.9663,
72
+ "step": 18
73
+ },
74
+ {
75
+ "epoch": 0.13245033112582782,
76
+ "grad_norm": 4.890565872192383,
77
+ "learning_rate": 2.173913043478261e-05,
78
+ "loss": 2.885,
79
+ "step": 20
80
+ },
81
+ {
82
+ "epoch": 0.1456953642384106,
83
+ "grad_norm": 4.127910137176514,
84
+ "learning_rate": 2.391304347826087e-05,
85
+ "loss": 2.9042,
86
+ "step": 22
87
+ },
88
+ {
89
+ "epoch": 0.15894039735099338,
90
+ "grad_norm": 5.797154903411865,
91
+ "learning_rate": 2.608695652173913e-05,
92
+ "loss": 2.743,
93
+ "step": 24
94
+ },
95
+ {
96
+ "epoch": 0.17218543046357615,
97
+ "grad_norm": 4.610489845275879,
98
+ "learning_rate": 2.826086956521739e-05,
99
+ "loss": 2.727,
100
+ "step": 26
101
+ },
102
+ {
103
+ "epoch": 0.18543046357615894,
104
+ "grad_norm": 4.88078498840332,
105
+ "learning_rate": 3.0434782608695656e-05,
106
+ "loss": 2.6978,
107
+ "step": 28
108
+ },
109
+ {
110
+ "epoch": 0.1986754966887417,
111
+ "grad_norm": 4.7736334800720215,
112
+ "learning_rate": 3.260869565217392e-05,
113
+ "loss": 2.6417,
114
+ "step": 30
115
+ },
116
+ {
117
+ "epoch": 0.2119205298013245,
118
+ "grad_norm": 8.030829429626465,
119
+ "learning_rate": 3.478260869565218e-05,
120
+ "loss": 2.6245,
121
+ "step": 32
122
+ },
123
+ {
124
+ "epoch": 0.2251655629139073,
125
+ "grad_norm": 7.201488971710205,
126
+ "learning_rate": 3.695652173913043e-05,
127
+ "loss": 2.1068,
128
+ "step": 34
129
+ },
130
+ {
131
+ "epoch": 0.23841059602649006,
132
+ "grad_norm": 6.687085151672363,
133
+ "learning_rate": 3.804347826086957e-05,
134
+ "loss": 2.065,
135
+ "step": 36
136
+ },
137
+ {
138
+ "epoch": 0.25165562913907286,
139
+ "grad_norm": 7.2274675369262695,
140
+ "learning_rate": 4.021739130434783e-05,
141
+ "loss": 2.2102,
142
+ "step": 38
143
+ },
144
+ {
145
+ "epoch": 0.26490066225165565,
146
+ "grad_norm": 8.127673149108887,
147
+ "learning_rate": 4.239130434782609e-05,
148
+ "loss": 2.1585,
149
+ "step": 40
150
+ },
151
+ {
152
+ "epoch": 0.2781456953642384,
153
+ "grad_norm": 8.082962989807129,
154
+ "learning_rate": 4.456521739130435e-05,
155
+ "loss": 1.651,
156
+ "step": 42
157
+ },
158
+ {
159
+ "epoch": 0.2913907284768212,
160
+ "grad_norm": 9.228837013244629,
161
+ "learning_rate": 4.673913043478261e-05,
162
+ "loss": 2.5421,
163
+ "step": 44
164
+ },
165
+ {
166
+ "epoch": 0.304635761589404,
167
+ "grad_norm": 8.194953918457031,
168
+ "learning_rate": 4.891304347826087e-05,
169
+ "loss": 1.7752,
170
+ "step": 46
171
+ },
172
+ {
173
+ "epoch": 0.31788079470198677,
174
+ "grad_norm": 11.126862525939941,
175
+ "learning_rate": 4.987714987714988e-05,
176
+ "loss": 2.6941,
177
+ "step": 48
178
+ },
179
+ {
180
+ "epoch": 0.33112582781456956,
181
+ "grad_norm": 7.496665954589844,
182
+ "learning_rate": 4.963144963144963e-05,
183
+ "loss": 2.156,
184
+ "step": 50
185
+ },
186
+ {
187
+ "epoch": 0.3443708609271523,
188
+ "grad_norm": 6.483855724334717,
189
+ "learning_rate": 4.9385749385749387e-05,
190
+ "loss": 1.7352,
191
+ "step": 52
192
+ },
193
+ {
194
+ "epoch": 0.3576158940397351,
195
+ "grad_norm": 7.87312650680542,
196
+ "learning_rate": 4.914004914004915e-05,
197
+ "loss": 1.8533,
198
+ "step": 54
199
+ },
200
+ {
201
+ "epoch": 0.3708609271523179,
202
+ "grad_norm": 14.155879020690918,
203
+ "learning_rate": 4.8894348894348894e-05,
204
+ "loss": 1.9163,
205
+ "step": 56
206
+ },
207
+ {
208
+ "epoch": 0.3841059602649007,
209
+ "grad_norm": 6.899637699127197,
210
+ "learning_rate": 4.8648648648648654e-05,
211
+ "loss": 2.0497,
212
+ "step": 58
213
+ },
214
+ {
215
+ "epoch": 0.3973509933774834,
216
+ "grad_norm": 9.851097106933594,
217
+ "learning_rate": 4.840294840294841e-05,
218
+ "loss": 2.0549,
219
+ "step": 60
220
+ },
221
+ {
222
+ "epoch": 0.4105960264900662,
223
+ "grad_norm": 7.761284828186035,
224
+ "learning_rate": 4.8157248157248155e-05,
225
+ "loss": 2.4832,
226
+ "step": 62
227
+ },
228
+ {
229
+ "epoch": 0.423841059602649,
230
+ "grad_norm": 7.7907819747924805,
231
+ "learning_rate": 4.7911547911547915e-05,
232
+ "loss": 2.0239,
233
+ "step": 64
234
+ },
235
+ {
236
+ "epoch": 0.4370860927152318,
237
+ "grad_norm": 6.249843597412109,
238
+ "learning_rate": 4.766584766584767e-05,
239
+ "loss": 1.5607,
240
+ "step": 66
241
+ },
242
+ {
243
+ "epoch": 0.4503311258278146,
244
+ "grad_norm": 5.760603904724121,
245
+ "learning_rate": 4.742014742014742e-05,
246
+ "loss": 1.7551,
247
+ "step": 68
248
+ },
249
+ {
250
+ "epoch": 0.46357615894039733,
251
+ "grad_norm": 8.171866416931152,
252
+ "learning_rate": 4.7174447174447176e-05,
253
+ "loss": 1.7013,
254
+ "step": 70
255
+ },
256
+ {
257
+ "epoch": 0.4768211920529801,
258
+ "grad_norm": 7.343658924102783,
259
+ "learning_rate": 4.692874692874693e-05,
260
+ "loss": 2.1362,
261
+ "step": 72
262
+ },
263
+ {
264
+ "epoch": 0.4900662251655629,
265
+ "grad_norm": 11.477375030517578,
266
+ "learning_rate": 4.6683046683046684e-05,
267
+ "loss": 2.0896,
268
+ "step": 74
269
+ },
270
+ {
271
+ "epoch": 0.5033112582781457,
272
+ "grad_norm": 6.7531256675720215,
273
+ "learning_rate": 4.6437346437346444e-05,
274
+ "loss": 1.5969,
275
+ "step": 76
276
+ },
277
+ {
278
+ "epoch": 0.5165562913907285,
279
+ "grad_norm": 9.825312614440918,
280
+ "learning_rate": 4.619164619164619e-05,
281
+ "loss": 2.1228,
282
+ "step": 78
283
+ },
284
+ {
285
+ "epoch": 0.5298013245033113,
286
+ "grad_norm": 7.430478096008301,
287
+ "learning_rate": 4.594594594594595e-05,
288
+ "loss": 2.1018,
289
+ "step": 80
290
+ },
291
+ {
292
+ "epoch": 0.543046357615894,
293
+ "grad_norm": 5.134402275085449,
294
+ "learning_rate": 4.5700245700245705e-05,
295
+ "loss": 1.5301,
296
+ "step": 82
297
+ },
298
+ {
299
+ "epoch": 0.5562913907284768,
300
+ "grad_norm": 11.335017204284668,
301
+ "learning_rate": 4.545454545454546e-05,
302
+ "loss": 1.5022,
303
+ "step": 84
304
+ },
305
+ {
306
+ "epoch": 0.5695364238410596,
307
+ "grad_norm": 8.13528823852539,
308
+ "learning_rate": 4.520884520884521e-05,
309
+ "loss": 2.2219,
310
+ "step": 86
311
+ },
312
+ {
313
+ "epoch": 0.5827814569536424,
314
+ "grad_norm": 5.774362564086914,
315
+ "learning_rate": 4.4963144963144966e-05,
316
+ "loss": 1.5469,
317
+ "step": 88
318
+ },
319
+ {
320
+ "epoch": 0.5960264900662252,
321
+ "grad_norm": 10.505985260009766,
322
+ "learning_rate": 4.471744471744472e-05,
323
+ "loss": 2.1449,
324
+ "step": 90
325
+ },
326
+ {
327
+ "epoch": 0.609271523178808,
328
+ "grad_norm": 9.9073486328125,
329
+ "learning_rate": 4.447174447174447e-05,
330
+ "loss": 1.9911,
331
+ "step": 92
332
+ },
333
+ {
334
+ "epoch": 0.6225165562913907,
335
+ "grad_norm": 7.591387748718262,
336
+ "learning_rate": 4.422604422604423e-05,
337
+ "loss": 1.2767,
338
+ "step": 94
339
+ },
340
+ {
341
+ "epoch": 0.6357615894039735,
342
+ "grad_norm": 12.243816375732422,
343
+ "learning_rate": 4.398034398034398e-05,
344
+ "loss": 2.4575,
345
+ "step": 96
346
+ },
347
+ {
348
+ "epoch": 0.6490066225165563,
349
+ "grad_norm": 12.278913497924805,
350
+ "learning_rate": 4.373464373464374e-05,
351
+ "loss": 2.6426,
352
+ "step": 98
353
+ },
354
+ {
355
+ "epoch": 0.6622516556291391,
356
+ "grad_norm": 8.50022029876709,
357
+ "learning_rate": 4.348894348894349e-05,
358
+ "loss": 1.8184,
359
+ "step": 100
360
+ },
361
+ {
362
+ "epoch": 0.6754966887417219,
363
+ "grad_norm": 15.808201789855957,
364
+ "learning_rate": 4.324324324324325e-05,
365
+ "loss": 1.8794,
366
+ "step": 102
367
+ },
368
+ {
369
+ "epoch": 0.6887417218543046,
370
+ "grad_norm": 11.01276969909668,
371
+ "learning_rate": 4.2997542997543e-05,
372
+ "loss": 1.7045,
373
+ "step": 104
374
+ },
375
+ {
376
+ "epoch": 0.7019867549668874,
377
+ "grad_norm": 6.432559490203857,
378
+ "learning_rate": 4.2751842751842756e-05,
379
+ "loss": 1.3662,
380
+ "step": 106
381
+ },
382
+ {
383
+ "epoch": 0.7152317880794702,
384
+ "grad_norm": 12.68566608428955,
385
+ "learning_rate": 4.250614250614251e-05,
386
+ "loss": 1.0612,
387
+ "step": 108
388
+ },
389
+ {
390
+ "epoch": 0.7284768211920529,
391
+ "grad_norm": 7.320284843444824,
392
+ "learning_rate": 4.226044226044226e-05,
393
+ "loss": 1.6157,
394
+ "step": 110
395
+ },
396
+ {
397
+ "epoch": 0.7417218543046358,
398
+ "grad_norm": 9.558548927307129,
399
+ "learning_rate": 4.2014742014742017e-05,
400
+ "loss": 1.6096,
401
+ "step": 112
402
+ },
403
+ {
404
+ "epoch": 0.7549668874172185,
405
+ "grad_norm": 12.876482009887695,
406
+ "learning_rate": 4.176904176904177e-05,
407
+ "loss": 2.0925,
408
+ "step": 114
409
+ },
410
+ {
411
+ "epoch": 0.7682119205298014,
412
+ "grad_norm": 16.353004455566406,
413
+ "learning_rate": 4.1523341523341524e-05,
414
+ "loss": 2.4218,
415
+ "step": 116
416
+ },
417
+ {
418
+ "epoch": 0.7814569536423841,
419
+ "grad_norm": 8.213098526000977,
420
+ "learning_rate": 4.127764127764128e-05,
421
+ "loss": 1.4581,
422
+ "step": 118
423
+ },
424
+ {
425
+ "epoch": 0.7947019867549668,
426
+ "grad_norm": 7.289383888244629,
427
+ "learning_rate": 4.103194103194104e-05,
428
+ "loss": 1.3969,
429
+ "step": 120
430
+ },
431
+ {
432
+ "epoch": 0.8079470198675497,
433
+ "grad_norm": 12.726005554199219,
434
+ "learning_rate": 4.0786240786240785e-05,
435
+ "loss": 1.6086,
436
+ "step": 122
437
+ },
438
+ {
439
+ "epoch": 0.8211920529801324,
440
+ "grad_norm": 6.612705230712891,
441
+ "learning_rate": 4.0540540540540545e-05,
442
+ "loss": 1.228,
443
+ "step": 124
444
+ },
445
+ {
446
+ "epoch": 0.8344370860927153,
447
+ "grad_norm": 8.954344749450684,
448
+ "learning_rate": 4.02948402948403e-05,
449
+ "loss": 1.9795,
450
+ "step": 126
451
+ },
452
+ {
453
+ "epoch": 0.847682119205298,
454
+ "grad_norm": 14.46446418762207,
455
+ "learning_rate": 4.004914004914005e-05,
456
+ "loss": 1.4446,
457
+ "step": 128
458
+ },
459
+ {
460
+ "epoch": 0.8609271523178808,
461
+ "grad_norm": 6.220669746398926,
462
+ "learning_rate": 3.9803439803439806e-05,
463
+ "loss": 2.0856,
464
+ "step": 130
465
+ },
466
+ {
467
+ "epoch": 0.8741721854304636,
468
+ "grad_norm": 6.742675304412842,
469
+ "learning_rate": 3.955773955773956e-05,
470
+ "loss": 1.1761,
471
+ "step": 132
472
+ },
473
+ {
474
+ "epoch": 0.8874172185430463,
475
+ "grad_norm": 6.69525146484375,
476
+ "learning_rate": 3.9312039312039314e-05,
477
+ "loss": 1.2581,
478
+ "step": 134
479
+ },
480
+ {
481
+ "epoch": 0.9006622516556292,
482
+ "grad_norm": 5.910060405731201,
483
+ "learning_rate": 3.906633906633907e-05,
484
+ "loss": 1.1815,
485
+ "step": 136
486
+ },
487
+ {
488
+ "epoch": 0.9139072847682119,
489
+ "grad_norm": 5.171950817108154,
490
+ "learning_rate": 3.882063882063882e-05,
491
+ "loss": 0.9752,
492
+ "step": 138
493
+ },
494
+ {
495
+ "epoch": 0.9271523178807947,
496
+ "grad_norm": 7.012332916259766,
497
+ "learning_rate": 3.857493857493858e-05,
498
+ "loss": 1.7936,
499
+ "step": 140
500
+ },
501
+ {
502
+ "epoch": 0.9403973509933775,
503
+ "grad_norm": 7.54193639755249,
504
+ "learning_rate": 3.8329238329238335e-05,
505
+ "loss": 1.4808,
506
+ "step": 142
507
+ },
508
+ {
509
+ "epoch": 0.9536423841059603,
510
+ "grad_norm": 7.4099531173706055,
511
+ "learning_rate": 3.808353808353808e-05,
512
+ "loss": 1.2554,
513
+ "step": 144
514
+ },
515
+ {
516
+ "epoch": 0.9668874172185431,
517
+ "grad_norm": 10.860753059387207,
518
+ "learning_rate": 3.783783783783784e-05,
519
+ "loss": 2.0704,
520
+ "step": 146
521
+ },
522
+ {
523
+ "epoch": 0.9801324503311258,
524
+ "grad_norm": 6.87387752532959,
525
+ "learning_rate": 3.7592137592137596e-05,
526
+ "loss": 1.6014,
527
+ "step": 148
528
+ },
529
+ {
530
+ "epoch": 0.9933774834437086,
531
+ "grad_norm": 6.1860175132751465,
532
+ "learning_rate": 3.734643734643735e-05,
533
+ "loss": 0.9634,
534
+ "step": 150
535
+ },
536
+ {
537
+ "epoch": 1.0,
538
+ "eval_accuracy": 0.43,
539
+ "eval_f1_macro": 0.10414486290241251,
540
+ "eval_f1_micro": 0.43,
541
+ "eval_f1_weighted": 0.3778887980009637,
542
+ "eval_loss": 2.039196252822876,
543
+ "eval_precision_macro": 0.12514957264957266,
544
+ "eval_precision_micro": 0.43,
545
+ "eval_precision_weighted": 0.4138803418803419,
546
+ "eval_recall_macro": 0.1286363636363636,
547
+ "eval_recall_micro": 0.43,
548
+ "eval_recall_weighted": 0.43,
549
+ "eval_runtime": 0.4069,
550
+ "eval_samples_per_second": 245.77,
551
+ "eval_steps_per_second": 17.204,
552
+ "step": 151
553
+ },
554
+ {
555
+ "epoch": 1.0066225165562914,
556
+ "grad_norm": 13.422128677368164,
557
+ "learning_rate": 3.71007371007371e-05,
558
+ "loss": 1.1134,
559
+ "step": 152
560
+ },
561
+ {
562
+ "epoch": 1.0198675496688743,
563
+ "grad_norm": 7.478858947753906,
564
+ "learning_rate": 3.685503685503686e-05,
565
+ "loss": 1.177,
566
+ "step": 154
567
+ },
568
+ {
569
+ "epoch": 1.033112582781457,
570
+ "grad_norm": 9.430647850036621,
571
+ "learning_rate": 3.660933660933661e-05,
572
+ "loss": 2.2911,
573
+ "step": 156
574
+ },
575
+ {
576
+ "epoch": 1.0463576158940397,
577
+ "grad_norm": 5.825343608856201,
578
+ "learning_rate": 3.6363636363636364e-05,
579
+ "loss": 1.2509,
580
+ "step": 158
581
+ },
582
+ {
583
+ "epoch": 1.0596026490066226,
584
+ "grad_norm": 9.491595268249512,
585
+ "learning_rate": 3.611793611793612e-05,
586
+ "loss": 1.3177,
587
+ "step": 160
588
+ },
589
+ {
590
+ "epoch": 1.0728476821192052,
591
+ "grad_norm": 5.076850891113281,
592
+ "learning_rate": 3.587223587223588e-05,
593
+ "loss": 0.9434,
594
+ "step": 162
595
+ },
596
+ {
597
+ "epoch": 1.086092715231788,
598
+ "grad_norm": 5.8091206550598145,
599
+ "learning_rate": 3.562653562653563e-05,
600
+ "loss": 1.4945,
601
+ "step": 164
602
+ },
603
+ {
604
+ "epoch": 1.099337748344371,
605
+ "grad_norm": 5.286921501159668,
606
+ "learning_rate": 3.538083538083538e-05,
607
+ "loss": 0.9087,
608
+ "step": 166
609
+ },
610
+ {
611
+ "epoch": 1.1125827814569536,
612
+ "grad_norm": 14.105121612548828,
613
+ "learning_rate": 3.513513513513514e-05,
614
+ "loss": 1.8953,
615
+ "step": 168
616
+ },
617
+ {
618
+ "epoch": 1.1258278145695364,
619
+ "grad_norm": 7.049529552459717,
620
+ "learning_rate": 3.488943488943489e-05,
621
+ "loss": 1.2879,
622
+ "step": 170
623
+ },
624
+ {
625
+ "epoch": 1.1390728476821192,
626
+ "grad_norm": 5.330937385559082,
627
+ "learning_rate": 3.4643734643734647e-05,
628
+ "loss": 1.0429,
629
+ "step": 172
630
+ },
631
+ {
632
+ "epoch": 1.152317880794702,
633
+ "grad_norm": 14.564863204956055,
634
+ "learning_rate": 3.43980343980344e-05,
635
+ "loss": 1.2875,
636
+ "step": 174
637
+ },
638
+ {
639
+ "epoch": 1.1655629139072847,
640
+ "grad_norm": 13.017091751098633,
641
+ "learning_rate": 3.4152334152334154e-05,
642
+ "loss": 1.0357,
643
+ "step": 176
644
+ },
645
+ {
646
+ "epoch": 1.1788079470198676,
647
+ "grad_norm": 13.703240394592285,
648
+ "learning_rate": 3.390663390663391e-05,
649
+ "loss": 1.5905,
650
+ "step": 178
651
+ },
652
+ {
653
+ "epoch": 1.1920529801324504,
654
+ "grad_norm": 7.464919567108154,
655
+ "learning_rate": 3.366093366093366e-05,
656
+ "loss": 1.5741,
657
+ "step": 180
658
+ },
659
+ {
660
+ "epoch": 1.205298013245033,
661
+ "grad_norm": 7.947140216827393,
662
+ "learning_rate": 3.3415233415233415e-05,
663
+ "loss": 1.845,
664
+ "step": 182
665
+ },
666
+ {
667
+ "epoch": 1.218543046357616,
668
+ "grad_norm": 6.652373790740967,
669
+ "learning_rate": 3.3169533169533175e-05,
670
+ "loss": 1.4041,
671
+ "step": 184
672
+ },
673
+ {
674
+ "epoch": 1.2317880794701987,
675
+ "grad_norm": 5.824616432189941,
676
+ "learning_rate": 3.292383292383293e-05,
677
+ "loss": 1.5873,
678
+ "step": 186
679
+ },
680
+ {
681
+ "epoch": 1.2450331125827814,
682
+ "grad_norm": 10.095503807067871,
683
+ "learning_rate": 3.2678132678132676e-05,
684
+ "loss": 1.4017,
685
+ "step": 188
686
+ },
687
+ {
688
+ "epoch": 1.2582781456953642,
689
+ "grad_norm": 5.2991766929626465,
690
+ "learning_rate": 3.2432432432432436e-05,
691
+ "loss": 1.395,
692
+ "step": 190
693
+ },
694
+ {
695
+ "epoch": 1.271523178807947,
696
+ "grad_norm": 4.981668472290039,
697
+ "learning_rate": 3.218673218673219e-05,
698
+ "loss": 1.4415,
699
+ "step": 192
700
+ },
701
+ {
702
+ "epoch": 1.2847682119205297,
703
+ "grad_norm": 8.634035110473633,
704
+ "learning_rate": 3.1941031941031943e-05,
705
+ "loss": 1.3291,
706
+ "step": 194
707
+ },
708
+ {
709
+ "epoch": 1.2980132450331126,
710
+ "grad_norm": 8.355801582336426,
711
+ "learning_rate": 3.16953316953317e-05,
712
+ "loss": 0.9761,
713
+ "step": 196
714
+ },
715
+ {
716
+ "epoch": 1.3112582781456954,
717
+ "grad_norm": NaN,
718
+ "learning_rate": 3.1572481572481574e-05,
719
+ "loss": 1.1352,
720
+ "step": 198
721
+ },
722
+ {
723
+ "epoch": 1.3245033112582782,
724
+ "grad_norm": 5.318390846252441,
725
+ "learning_rate": 3.132678132678133e-05,
726
+ "loss": 1.1314,
727
+ "step": 200
728
+ },
729
+ {
730
+ "epoch": 1.3377483443708609,
731
+ "grad_norm": 7.791065692901611,
732
+ "learning_rate": 3.108108108108108e-05,
733
+ "loss": 1.5313,
734
+ "step": 202
735
+ },
736
+ {
737
+ "epoch": 1.3509933774834437,
738
+ "grad_norm": 21.741308212280273,
739
+ "learning_rate": 3.083538083538084e-05,
740
+ "loss": 1.7144,
741
+ "step": 204
742
+ },
743
+ {
744
+ "epoch": 1.3642384105960264,
745
+ "grad_norm": 14.097997665405273,
746
+ "learning_rate": 3.058968058968059e-05,
747
+ "loss": 1.749,
748
+ "step": 206
749
+ },
750
+ {
751
+ "epoch": 1.3774834437086092,
752
+ "grad_norm": 6.286694526672363,
753
+ "learning_rate": 3.0343980343980342e-05,
754
+ "loss": 1.3443,
755
+ "step": 208
756
+ },
757
+ {
758
+ "epoch": 1.390728476821192,
759
+ "grad_norm": 7.453071117401123,
760
+ "learning_rate": 3.0098280098280103e-05,
761
+ "loss": 0.85,
762
+ "step": 210
763
+ },
764
+ {
765
+ "epoch": 1.403973509933775,
766
+ "grad_norm": 10.761088371276855,
767
+ "learning_rate": 2.9852579852579853e-05,
768
+ "loss": 0.8886,
769
+ "step": 212
770
+ },
771
+ {
772
+ "epoch": 1.4172185430463577,
773
+ "grad_norm": 5.939634799957275,
774
+ "learning_rate": 2.9606879606879607e-05,
775
+ "loss": 1.2804,
776
+ "step": 214
777
+ },
778
+ {
779
+ "epoch": 1.4304635761589404,
780
+ "grad_norm": 9.565838813781738,
781
+ "learning_rate": 2.9361179361179364e-05,
782
+ "loss": 1.5924,
783
+ "step": 216
784
+ },
785
+ {
786
+ "epoch": 1.4437086092715232,
787
+ "grad_norm": 10.187567710876465,
788
+ "learning_rate": 2.9115479115479117e-05,
789
+ "loss": 1.3317,
790
+ "step": 218
791
+ },
792
+ {
793
+ "epoch": 1.4569536423841059,
794
+ "grad_norm": 8.318490028381348,
795
+ "learning_rate": 2.8869778869778868e-05,
796
+ "loss": 1.617,
797
+ "step": 220
798
+ },
799
+ {
800
+ "epoch": 1.4701986754966887,
801
+ "grad_norm": 7.841922283172607,
802
+ "learning_rate": 2.8624078624078625e-05,
803
+ "loss": 1.463,
804
+ "step": 222
805
+ },
806
+ {
807
+ "epoch": 1.4834437086092715,
808
+ "grad_norm": 6.505579471588135,
809
+ "learning_rate": 2.8378378378378378e-05,
810
+ "loss": 1.123,
811
+ "step": 224
812
+ },
813
+ {
814
+ "epoch": 1.4966887417218544,
815
+ "grad_norm": 5.509284496307373,
816
+ "learning_rate": 2.8132678132678135e-05,
817
+ "loss": 1.0619,
818
+ "step": 226
819
+ },
820
+ {
821
+ "epoch": 1.5099337748344372,
822
+ "grad_norm": 4.9719367027282715,
823
+ "learning_rate": 2.788697788697789e-05,
824
+ "loss": 0.6785,
825
+ "step": 228
826
+ },
827
+ {
828
+ "epoch": 1.5231788079470199,
829
+ "grad_norm": 7.88023567199707,
830
+ "learning_rate": 2.764127764127764e-05,
831
+ "loss": 1.1481,
832
+ "step": 230
833
+ },
834
+ {
835
+ "epoch": 1.5364238410596025,
836
+ "grad_norm": 5.841632843017578,
837
+ "learning_rate": 2.73955773955774e-05,
838
+ "loss": 0.7702,
839
+ "step": 232
840
+ },
841
+ {
842
+ "epoch": 1.5496688741721854,
843
+ "grad_norm": 8.680448532104492,
844
+ "learning_rate": 2.714987714987715e-05,
845
+ "loss": 1.4025,
846
+ "step": 234
847
+ },
848
+ {
849
+ "epoch": 1.5629139072847682,
850
+ "grad_norm": 5.259287357330322,
851
+ "learning_rate": 2.6904176904176904e-05,
852
+ "loss": 0.8707,
853
+ "step": 236
854
+ },
855
+ {
856
+ "epoch": 1.576158940397351,
857
+ "grad_norm": 5.991212368011475,
858
+ "learning_rate": 2.665847665847666e-05,
859
+ "loss": 0.5955,
860
+ "step": 238
861
+ },
862
+ {
863
+ "epoch": 1.589403973509934,
864
+ "grad_norm": 11.001547813415527,
865
+ "learning_rate": 2.6412776412776414e-05,
866
+ "loss": 1.5866,
867
+ "step": 240
868
+ },
869
+ {
870
+ "epoch": 1.6026490066225165,
871
+ "grad_norm": 5.934292316436768,
872
+ "learning_rate": 2.616707616707617e-05,
873
+ "loss": 1.3472,
874
+ "step": 242
875
+ },
876
+ {
877
+ "epoch": 1.6158940397350994,
878
+ "grad_norm": 16.11164665222168,
879
+ "learning_rate": 2.5921375921375925e-05,
880
+ "loss": 1.2663,
881
+ "step": 244
882
+ },
883
+ {
884
+ "epoch": 1.629139072847682,
885
+ "grad_norm": 10.690239906311035,
886
+ "learning_rate": 2.5675675675675675e-05,
887
+ "loss": 1.5754,
888
+ "step": 246
889
+ },
890
+ {
891
+ "epoch": 1.6423841059602649,
892
+ "grad_norm": 7.545533657073975,
893
+ "learning_rate": 2.5429975429975432e-05,
894
+ "loss": 0.9722,
895
+ "step": 248
896
+ },
897
+ {
898
+ "epoch": 1.6556291390728477,
899
+ "grad_norm": 7.82634162902832,
900
+ "learning_rate": 2.5184275184275186e-05,
901
+ "loss": 1.8525,
902
+ "step": 250
903
+ },
904
+ {
905
+ "epoch": 1.6688741721854305,
906
+ "grad_norm": 11.477102279663086,
907
+ "learning_rate": 2.493857493857494e-05,
908
+ "loss": 0.9904,
909
+ "step": 252
910
+ },
911
+ {
912
+ "epoch": 1.6821192052980134,
913
+ "grad_norm": 8.056424140930176,
914
+ "learning_rate": 2.4692874692874693e-05,
915
+ "loss": 1.6071,
916
+ "step": 254
917
+ },
918
+ {
919
+ "epoch": 1.695364238410596,
920
+ "grad_norm": 8.872392654418945,
921
+ "learning_rate": 2.4447174447174447e-05,
922
+ "loss": 1.8732,
923
+ "step": 256
924
+ },
925
+ {
926
+ "epoch": 1.7086092715231787,
927
+ "grad_norm": 8.307268142700195,
928
+ "learning_rate": 2.4201474201474204e-05,
929
+ "loss": 1.2843,
930
+ "step": 258
931
+ },
932
+ {
933
+ "epoch": 1.7218543046357615,
934
+ "grad_norm": 8.889561653137207,
935
+ "learning_rate": 2.3955773955773958e-05,
936
+ "loss": 1.8855,
937
+ "step": 260
938
+ },
939
+ {
940
+ "epoch": 1.7350993377483444,
941
+ "grad_norm": 7.262237071990967,
942
+ "learning_rate": 2.371007371007371e-05,
943
+ "loss": 0.6363,
944
+ "step": 262
945
+ },
946
+ {
947
+ "epoch": 1.7483443708609272,
948
+ "grad_norm": 9.860976219177246,
949
+ "learning_rate": 2.3464373464373465e-05,
950
+ "loss": 0.9588,
951
+ "step": 264
952
+ },
953
+ {
954
+ "epoch": 1.76158940397351,
955
+ "grad_norm": 8.475176811218262,
956
+ "learning_rate": 2.3218673218673222e-05,
957
+ "loss": 1.4638,
958
+ "step": 266
959
+ },
960
+ {
961
+ "epoch": 1.7748344370860927,
962
+ "grad_norm": 6.681599140167236,
963
+ "learning_rate": 2.2972972972972976e-05,
964
+ "loss": 0.9692,
965
+ "step": 268
966
+ },
967
+ {
968
+ "epoch": 1.7880794701986755,
969
+ "grad_norm": 9.975168228149414,
970
+ "learning_rate": 2.272727272727273e-05,
971
+ "loss": 2.0131,
972
+ "step": 270
973
+ },
974
+ {
975
+ "epoch": 1.8013245033112582,
976
+ "grad_norm": 9.302199363708496,
977
+ "learning_rate": 2.2481572481572483e-05,
978
+ "loss": 1.4159,
979
+ "step": 272
980
+ },
981
+ {
982
+ "epoch": 1.814569536423841,
983
+ "grad_norm": 10.039788246154785,
984
+ "learning_rate": 2.2235872235872237e-05,
985
+ "loss": 1.5486,
986
+ "step": 274
987
+ },
988
+ {
989
+ "epoch": 1.8278145695364238,
990
+ "grad_norm": 7.697207927703857,
991
+ "learning_rate": 2.199017199017199e-05,
992
+ "loss": 1.0797,
993
+ "step": 276
994
+ },
995
+ {
996
+ "epoch": 1.8410596026490067,
997
+ "grad_norm": 7.76730489730835,
998
+ "learning_rate": 2.1744471744471744e-05,
999
+ "loss": 1.1921,
1000
+ "step": 278
1001
+ },
1002
+ {
1003
+ "epoch": 1.8543046357615895,
1004
+ "grad_norm": 7.227729320526123,
1005
+ "learning_rate": 2.14987714987715e-05,
1006
+ "loss": 1.6573,
1007
+ "step": 280
1008
+ },
1009
+ {
1010
+ "epoch": 1.8675496688741722,
1011
+ "grad_norm": 6.351704120635986,
1012
+ "learning_rate": 2.1253071253071255e-05,
1013
+ "loss": 0.6416,
1014
+ "step": 282
1015
+ },
1016
+ {
1017
+ "epoch": 1.8807947019867548,
1018
+ "grad_norm": 7.4418511390686035,
1019
+ "learning_rate": 2.1007371007371008e-05,
1020
+ "loss": 1.7128,
1021
+ "step": 284
1022
+ },
1023
+ {
1024
+ "epoch": 1.8940397350993377,
1025
+ "grad_norm": 9.884121894836426,
1026
+ "learning_rate": 2.0761670761670762e-05,
1027
+ "loss": 2.3348,
1028
+ "step": 286
1029
+ },
1030
+ {
1031
+ "epoch": 1.9072847682119205,
1032
+ "grad_norm": 11.193192481994629,
1033
+ "learning_rate": 2.051597051597052e-05,
1034
+ "loss": 0.9744,
1035
+ "step": 288
1036
+ },
1037
+ {
1038
+ "epoch": 1.9205298013245033,
1039
+ "grad_norm": 6.8167829513549805,
1040
+ "learning_rate": 2.0270270270270273e-05,
1041
+ "loss": 1.2415,
1042
+ "step": 290
1043
+ },
1044
+ {
1045
+ "epoch": 1.9337748344370862,
1046
+ "grad_norm": 6.478491306304932,
1047
+ "learning_rate": 2.0024570024570026e-05,
1048
+ "loss": 1.1914,
1049
+ "step": 292
1050
+ },
1051
+ {
1052
+ "epoch": 1.9470198675496688,
1053
+ "grad_norm": 7.969542026519775,
1054
+ "learning_rate": 1.977886977886978e-05,
1055
+ "loss": 1.5221,
1056
+ "step": 294
1057
+ },
1058
+ {
1059
+ "epoch": 1.9602649006622517,
1060
+ "grad_norm": 7.827632427215576,
1061
+ "learning_rate": 1.9533169533169534e-05,
1062
+ "loss": 0.7076,
1063
+ "step": 296
1064
+ },
1065
+ {
1066
+ "epoch": 1.9735099337748343,
1067
+ "grad_norm": 5.793867111206055,
1068
+ "learning_rate": 1.928746928746929e-05,
1069
+ "loss": 1.1292,
1070
+ "step": 298
1071
+ },
1072
+ {
1073
+ "epoch": 1.9867549668874172,
1074
+ "grad_norm": 7.288516998291016,
1075
+ "learning_rate": 1.904176904176904e-05,
1076
+ "loss": 0.8474,
1077
+ "step": 300
1078
+ },
1079
+ {
1080
+ "epoch": 2.0,
1081
+ "grad_norm": 19.54204559326172,
1082
+ "learning_rate": 1.8796068796068798e-05,
1083
+ "loss": 0.8084,
1084
+ "step": 302
1085
+ },
1086
+ {
1087
+ "epoch": 2.0,
1088
+ "eval_accuracy": 0.52,
1089
+ "eval_f1_macro": 0.16279761904761905,
1090
+ "eval_f1_micro": 0.52,
1091
+ "eval_f1_weighted": 0.4582380952380952,
1092
+ "eval_loss": 1.6385834217071533,
1093
+ "eval_precision_macro": 0.17940422322775265,
1094
+ "eval_precision_micro": 0.52,
1095
+ "eval_precision_weighted": 0.47573152337858227,
1096
+ "eval_recall_macro": 0.2141883116883117,
1097
+ "eval_recall_micro": 0.52,
1098
+ "eval_recall_weighted": 0.52,
1099
+ "eval_runtime": 0.3788,
1100
+ "eval_samples_per_second": 263.985,
1101
+ "eval_steps_per_second": 18.479,
1102
+ "step": 302
1103
+ },
1104
+ {
1105
+ "epoch": 2.013245033112583,
1106
+ "grad_norm": 3.1326494216918945,
1107
+ "learning_rate": 1.855036855036855e-05,
1108
+ "loss": 0.4869,
1109
+ "step": 304
1110
+ },
1111
+ {
1112
+ "epoch": 2.0264900662251657,
1113
+ "grad_norm": 7.844410419464111,
1114
+ "learning_rate": 1.8304668304668305e-05,
1115
+ "loss": 1.0237,
1116
+ "step": 306
1117
+ },
1118
+ {
1119
+ "epoch": 2.0397350993377485,
1120
+ "grad_norm": 8.277088165283203,
1121
+ "learning_rate": 1.805896805896806e-05,
1122
+ "loss": 0.8911,
1123
+ "step": 308
1124
+ },
1125
+ {
1126
+ "epoch": 2.052980132450331,
1127
+ "grad_norm": 4.819505214691162,
1128
+ "learning_rate": 1.7813267813267816e-05,
1129
+ "loss": 0.6859,
1130
+ "step": 310
1131
+ },
1132
+ {
1133
+ "epoch": 2.066225165562914,
1134
+ "grad_norm": 7.4773359298706055,
1135
+ "learning_rate": 1.756756756756757e-05,
1136
+ "loss": 1.0567,
1137
+ "step": 312
1138
+ },
1139
+ {
1140
+ "epoch": 2.0794701986754967,
1141
+ "grad_norm": 8.229156494140625,
1142
+ "learning_rate": 1.7321867321867323e-05,
1143
+ "loss": 1.2833,
1144
+ "step": 314
1145
+ },
1146
+ {
1147
+ "epoch": 2.0927152317880795,
1148
+ "grad_norm": 11.932225227355957,
1149
+ "learning_rate": 1.7076167076167077e-05,
1150
+ "loss": 1.2839,
1151
+ "step": 316
1152
+ },
1153
+ {
1154
+ "epoch": 2.1059602649006623,
1155
+ "grad_norm": 4.1798095703125,
1156
+ "learning_rate": 1.683046683046683e-05,
1157
+ "loss": 0.335,
1158
+ "step": 318
1159
+ },
1160
+ {
1161
+ "epoch": 2.119205298013245,
1162
+ "grad_norm": 4.319066524505615,
1163
+ "learning_rate": 1.6584766584766588e-05,
1164
+ "loss": 0.3695,
1165
+ "step": 320
1166
+ },
1167
+ {
1168
+ "epoch": 2.1324503311258276,
1169
+ "grad_norm": 4.868326187133789,
1170
+ "learning_rate": 1.6339066339066338e-05,
1171
+ "loss": 0.5056,
1172
+ "step": 322
1173
+ },
1174
+ {
1175
+ "epoch": 2.1456953642384105,
1176
+ "grad_norm": 11.324278831481934,
1177
+ "learning_rate": 1.6093366093366095e-05,
1178
+ "loss": 1.6924,
1179
+ "step": 324
1180
+ },
1181
+ {
1182
+ "epoch": 2.1589403973509933,
1183
+ "grad_norm": 7.622723579406738,
1184
+ "learning_rate": 1.584766584766585e-05,
1185
+ "loss": 1.1639,
1186
+ "step": 326
1187
+ },
1188
+ {
1189
+ "epoch": 2.172185430463576,
1190
+ "grad_norm": 9.650656700134277,
1191
+ "learning_rate": 1.5601965601965606e-05,
1192
+ "loss": 1.449,
1193
+ "step": 328
1194
+ },
1195
+ {
1196
+ "epoch": 2.185430463576159,
1197
+ "grad_norm": 8.743961334228516,
1198
+ "learning_rate": 1.5356265356265356e-05,
1199
+ "loss": 1.319,
1200
+ "step": 330
1201
+ },
1202
+ {
1203
+ "epoch": 2.198675496688742,
1204
+ "grad_norm": 8.732221603393555,
1205
+ "learning_rate": 1.5110565110565111e-05,
1206
+ "loss": 1.2402,
1207
+ "step": 332
1208
+ },
1209
+ {
1210
+ "epoch": 2.2119205298013247,
1211
+ "grad_norm": 7.377650737762451,
1212
+ "learning_rate": 1.4864864864864867e-05,
1213
+ "loss": 1.0142,
1214
+ "step": 334
1215
+ },
1216
+ {
1217
+ "epoch": 2.225165562913907,
1218
+ "grad_norm": 7.734024524688721,
1219
+ "learning_rate": 1.4619164619164619e-05,
1220
+ "loss": 1.0699,
1221
+ "step": 336
1222
+ },
1223
+ {
1224
+ "epoch": 2.23841059602649,
1225
+ "grad_norm": 6.139077663421631,
1226
+ "learning_rate": 1.4373464373464374e-05,
1227
+ "loss": 0.5206,
1228
+ "step": 338
1229
+ },
1230
+ {
1231
+ "epoch": 2.251655629139073,
1232
+ "grad_norm": 7.571850299835205,
1233
+ "learning_rate": 1.412776412776413e-05,
1234
+ "loss": 0.8841,
1235
+ "step": 340
1236
+ },
1237
+ {
1238
+ "epoch": 2.2649006622516556,
1239
+ "grad_norm": 9.102130889892578,
1240
+ "learning_rate": 1.3882063882063885e-05,
1241
+ "loss": 0.6576,
1242
+ "step": 342
1243
+ },
1244
+ {
1245
+ "epoch": 2.2781456953642385,
1246
+ "grad_norm": 8.144769668579102,
1247
+ "learning_rate": 1.3636363636363637e-05,
1248
+ "loss": 0.7361,
1249
+ "step": 344
1250
+ },
1251
+ {
1252
+ "epoch": 2.2913907284768213,
1253
+ "grad_norm": 9.98493480682373,
1254
+ "learning_rate": 1.339066339066339e-05,
1255
+ "loss": 1.1676,
1256
+ "step": 346
1257
+ },
1258
+ {
1259
+ "epoch": 2.304635761589404,
1260
+ "grad_norm": 10.540783882141113,
1261
+ "learning_rate": 1.3144963144963146e-05,
1262
+ "loss": 1.3709,
1263
+ "step": 348
1264
+ },
1265
+ {
1266
+ "epoch": 2.3178807947019866,
1267
+ "grad_norm": 10.3671293258667,
1268
+ "learning_rate": 1.2899262899262901e-05,
1269
+ "loss": 0.8397,
1270
+ "step": 350
1271
+ },
1272
+ {
1273
+ "epoch": 2.3311258278145695,
1274
+ "grad_norm": 4.67352819442749,
1275
+ "learning_rate": 1.2653562653562653e-05,
1276
+ "loss": 0.539,
1277
+ "step": 352
1278
+ },
1279
+ {
1280
+ "epoch": 2.3443708609271523,
1281
+ "grad_norm": 9.957860946655273,
1282
+ "learning_rate": 1.2407862407862408e-05,
1283
+ "loss": 1.4744,
1284
+ "step": 354
1285
+ },
1286
+ {
1287
+ "epoch": 2.357615894039735,
1288
+ "grad_norm": 8.535019874572754,
1289
+ "learning_rate": 1.2162162162162164e-05,
1290
+ "loss": 1.3841,
1291
+ "step": 356
1292
+ },
1293
+ {
1294
+ "epoch": 2.370860927152318,
1295
+ "grad_norm": 8.11552906036377,
1296
+ "learning_rate": 1.1916461916461917e-05,
1297
+ "loss": 1.1301,
1298
+ "step": 358
1299
+ },
1300
+ {
1301
+ "epoch": 2.384105960264901,
1302
+ "grad_norm": 7.197366237640381,
1303
+ "learning_rate": 1.1670761670761671e-05,
1304
+ "loss": 0.6156,
1305
+ "step": 360
1306
+ },
1307
+ {
1308
+ "epoch": 2.3973509933774833,
1309
+ "grad_norm": 4.6600022315979,
1310
+ "learning_rate": 1.1425061425061426e-05,
1311
+ "loss": 0.7303,
1312
+ "step": 362
1313
+ },
1314
+ {
1315
+ "epoch": 2.410596026490066,
1316
+ "grad_norm": 6.252689838409424,
1317
+ "learning_rate": 1.117936117936118e-05,
1318
+ "loss": 0.9769,
1319
+ "step": 364
1320
+ },
1321
+ {
1322
+ "epoch": 2.423841059602649,
1323
+ "grad_norm": 9.90804672241211,
1324
+ "learning_rate": 1.0933660933660935e-05,
1325
+ "loss": 1.093,
1326
+ "step": 366
1327
+ },
1328
+ {
1329
+ "epoch": 2.437086092715232,
1330
+ "grad_norm": 9.742344856262207,
1331
+ "learning_rate": 1.0687960687960689e-05,
1332
+ "loss": 1.0279,
1333
+ "step": 368
1334
+ },
1335
+ {
1336
+ "epoch": 2.4503311258278146,
1337
+ "grad_norm": 9.368986129760742,
1338
+ "learning_rate": 1.0442260442260443e-05,
1339
+ "loss": 1.2061,
1340
+ "step": 370
1341
+ },
1342
+ {
1343
+ "epoch": 2.4635761589403975,
1344
+ "grad_norm": 6.378131866455078,
1345
+ "learning_rate": 1.0196560196560196e-05,
1346
+ "loss": 1.0066,
1347
+ "step": 372
1348
+ },
1349
+ {
1350
+ "epoch": 2.47682119205298,
1351
+ "grad_norm": 7.304665565490723,
1352
+ "learning_rate": 9.950859950859952e-06,
1353
+ "loss": 2.2284,
1354
+ "step": 374
1355
+ },
1356
+ {
1357
+ "epoch": 2.4900662251655628,
1358
+ "grad_norm": 5.73162317276001,
1359
+ "learning_rate": 9.705159705159705e-06,
1360
+ "loss": 0.4927,
1361
+ "step": 376
1362
+ },
1363
+ {
1364
+ "epoch": 2.5033112582781456,
1365
+ "grad_norm": 9.693008422851562,
1366
+ "learning_rate": 9.45945945945946e-06,
1367
+ "loss": 0.9368,
1368
+ "step": 378
1369
+ },
1370
+ {
1371
+ "epoch": 2.5165562913907285,
1372
+ "grad_norm": 7.258613586425781,
1373
+ "learning_rate": 9.213759213759214e-06,
1374
+ "loss": 0.5097,
1375
+ "step": 380
1376
+ },
1377
+ {
1378
+ "epoch": 2.5298013245033113,
1379
+ "grad_norm": 6.431227684020996,
1380
+ "learning_rate": 8.96805896805897e-06,
1381
+ "loss": 0.9039,
1382
+ "step": 382
1383
+ },
1384
+ {
1385
+ "epoch": 2.543046357615894,
1386
+ "grad_norm": 8.578726768493652,
1387
+ "learning_rate": 8.722358722358723e-06,
1388
+ "loss": 0.8253,
1389
+ "step": 384
1390
+ },
1391
+ {
1392
+ "epoch": 2.556291390728477,
1393
+ "grad_norm": 5.389899730682373,
1394
+ "learning_rate": 8.476658476658477e-06,
1395
+ "loss": 0.4513,
1396
+ "step": 386
1397
+ },
1398
+ {
1399
+ "epoch": 2.5695364238410594,
1400
+ "grad_norm": 10.525187492370605,
1401
+ "learning_rate": 8.230958230958232e-06,
1402
+ "loss": 1.1571,
1403
+ "step": 388
1404
+ },
1405
+ {
1406
+ "epoch": 2.5827814569536423,
1407
+ "grad_norm": 5.4934186935424805,
1408
+ "learning_rate": 7.985257985257986e-06,
1409
+ "loss": 0.7821,
1410
+ "step": 390
1411
+ },
1412
+ {
1413
+ "epoch": 2.596026490066225,
1414
+ "grad_norm": 9.986252784729004,
1415
+ "learning_rate": 7.73955773955774e-06,
1416
+ "loss": 0.8019,
1417
+ "step": 392
1418
+ },
1419
+ {
1420
+ "epoch": 2.609271523178808,
1421
+ "grad_norm": 9.1427001953125,
1422
+ "learning_rate": 7.493857493857494e-06,
1423
+ "loss": 1.349,
1424
+ "step": 394
1425
+ },
1426
+ {
1427
+ "epoch": 2.622516556291391,
1428
+ "grad_norm": 8.039468765258789,
1429
+ "learning_rate": 7.2481572481572485e-06,
1430
+ "loss": 1.0867,
1431
+ "step": 396
1432
+ },
1433
+ {
1434
+ "epoch": 2.6357615894039736,
1435
+ "grad_norm": 9.093656539916992,
1436
+ "learning_rate": 7.002457002457002e-06,
1437
+ "loss": 1.5422,
1438
+ "step": 398
1439
+ },
1440
+ {
1441
+ "epoch": 2.6490066225165565,
1442
+ "grad_norm": 13.18973445892334,
1443
+ "learning_rate": 6.7567567567567575e-06,
1444
+ "loss": 0.4601,
1445
+ "step": 400
1446
+ },
1447
+ {
1448
+ "epoch": 2.662251655629139,
1449
+ "grad_norm": 9.986296653747559,
1450
+ "learning_rate": 6.511056511056511e-06,
1451
+ "loss": 1.0286,
1452
+ "step": 402
1453
+ },
1454
+ {
1455
+ "epoch": 2.6754966887417218,
1456
+ "grad_norm": 9.373101234436035,
1457
+ "learning_rate": 6.2653562653562665e-06,
1458
+ "loss": 0.8767,
1459
+ "step": 404
1460
+ },
1461
+ {
1462
+ "epoch": 2.6887417218543046,
1463
+ "grad_norm": 9.866012573242188,
1464
+ "learning_rate": 6.019656019656019e-06,
1465
+ "loss": 1.1937,
1466
+ "step": 406
1467
+ },
1468
+ {
1469
+ "epoch": 2.7019867549668874,
1470
+ "grad_norm": 6.277444839477539,
1471
+ "learning_rate": 5.773955773955774e-06,
1472
+ "loss": 1.1616,
1473
+ "step": 408
1474
+ },
1475
+ {
1476
+ "epoch": 2.7152317880794703,
1477
+ "grad_norm": 5.679011344909668,
1478
+ "learning_rate": 5.528255528255528e-06,
1479
+ "loss": 0.7223,
1480
+ "step": 410
1481
+ },
1482
+ {
1483
+ "epoch": 2.7284768211920527,
1484
+ "grad_norm": 4.74644660949707,
1485
+ "learning_rate": 5.282555282555283e-06,
1486
+ "loss": 0.5082,
1487
+ "step": 412
1488
+ },
1489
+ {
1490
+ "epoch": 2.741721854304636,
1491
+ "grad_norm": 13.121922492980957,
1492
+ "learning_rate": 5.036855036855037e-06,
1493
+ "loss": 1.5552,
1494
+ "step": 414
1495
+ },
1496
+ {
1497
+ "epoch": 2.7549668874172184,
1498
+ "grad_norm": 8.272148132324219,
1499
+ "learning_rate": 4.791154791154792e-06,
1500
+ "loss": 0.9687,
1501
+ "step": 416
1502
+ },
1503
+ {
1504
+ "epoch": 2.7682119205298013,
1505
+ "grad_norm": 7.133453369140625,
1506
+ "learning_rate": 4.5454545454545455e-06,
1507
+ "loss": 0.5568,
1508
+ "step": 418
1509
+ },
1510
+ {
1511
+ "epoch": 2.781456953642384,
1512
+ "grad_norm": 5.385309219360352,
1513
+ "learning_rate": 4.2997542997543e-06,
1514
+ "loss": 0.5379,
1515
+ "step": 420
1516
+ },
1517
+ {
1518
+ "epoch": 2.794701986754967,
1519
+ "grad_norm": 9.613791465759277,
1520
+ "learning_rate": 4.0540540540540545e-06,
1521
+ "loss": 1.3592,
1522
+ "step": 422
1523
+ },
1524
+ {
1525
+ "epoch": 2.80794701986755,
1526
+ "grad_norm": 8.493631362915039,
1527
+ "learning_rate": 3.8083538083538086e-06,
1528
+ "loss": 1.1874,
1529
+ "step": 424
1530
+ },
1531
+ {
1532
+ "epoch": 2.821192052980132,
1533
+ "grad_norm": 5.972334861755371,
1534
+ "learning_rate": 3.562653562653563e-06,
1535
+ "loss": 0.8684,
1536
+ "step": 426
1537
+ },
1538
+ {
1539
+ "epoch": 2.8344370860927155,
1540
+ "grad_norm": 10.33562183380127,
1541
+ "learning_rate": 3.3169533169533168e-06,
1542
+ "loss": 1.0385,
1543
+ "step": 428
1544
+ },
1545
+ {
1546
+ "epoch": 2.847682119205298,
1547
+ "grad_norm": 4.6230363845825195,
1548
+ "learning_rate": 3.0712530712530717e-06,
1549
+ "loss": 0.8805,
1550
+ "step": 430
1551
+ },
1552
+ {
1553
+ "epoch": 2.8609271523178808,
1554
+ "grad_norm": 16.691070556640625,
1555
+ "learning_rate": 2.9484029484029485e-06,
1556
+ "loss": 1.494,
1557
+ "step": 432
1558
+ },
1559
+ {
1560
+ "epoch": 2.8741721854304636,
1561
+ "grad_norm": 6.581092834472656,
1562
+ "learning_rate": 2.702702702702703e-06,
1563
+ "loss": 0.4563,
1564
+ "step": 434
1565
+ },
1566
+ {
1567
+ "epoch": 2.8874172185430464,
1568
+ "grad_norm": 8.806626319885254,
1569
+ "learning_rate": 2.457002457002457e-06,
1570
+ "loss": 1.3625,
1571
+ "step": 436
1572
+ },
1573
+ {
1574
+ "epoch": 2.9006622516556293,
1575
+ "grad_norm": 5.90773344039917,
1576
+ "learning_rate": 2.211302211302211e-06,
1577
+ "loss": 1.1598,
1578
+ "step": 438
1579
+ },
1580
+ {
1581
+ "epoch": 2.9139072847682117,
1582
+ "grad_norm": 7.406730651855469,
1583
+ "learning_rate": 1.9656019656019657e-06,
1584
+ "loss": 0.7975,
1585
+ "step": 440
1586
+ },
1587
+ {
1588
+ "epoch": 2.9271523178807946,
1589
+ "grad_norm": 5.467130184173584,
1590
+ "learning_rate": 1.71990171990172e-06,
1591
+ "loss": 0.4307,
1592
+ "step": 442
1593
+ },
1594
+ {
1595
+ "epoch": 2.9403973509933774,
1596
+ "grad_norm": 7.527857780456543,
1597
+ "learning_rate": 1.4742014742014743e-06,
1598
+ "loss": 0.6769,
1599
+ "step": 444
1600
+ },
1601
+ {
1602
+ "epoch": 2.9536423841059603,
1603
+ "grad_norm": 11.855218887329102,
1604
+ "learning_rate": 1.2285012285012285e-06,
1605
+ "loss": 1.4558,
1606
+ "step": 446
1607
+ },
1608
+ {
1609
+ "epoch": 2.966887417218543,
1610
+ "grad_norm": 9.057221412658691,
1611
+ "learning_rate": 9.828009828009828e-07,
1612
+ "loss": 1.6682,
1613
+ "step": 448
1614
+ },
1615
+ {
1616
+ "epoch": 2.980132450331126,
1617
+ "grad_norm": 8.127178192138672,
1618
+ "learning_rate": 7.371007371007371e-07,
1619
+ "loss": 0.9139,
1620
+ "step": 450
1621
+ },
1622
+ {
1623
+ "epoch": 2.993377483443709,
1624
+ "grad_norm": 7.9965128898620605,
1625
+ "learning_rate": 4.914004914004914e-07,
1626
+ "loss": 0.7603,
1627
+ "step": 452
1628
+ },
1629
+ {
1630
+ "epoch": 3.0,
1631
+ "eval_accuracy": 0.64,
1632
+ "eval_f1_macro": 0.28164367547346275,
1633
+ "eval_f1_micro": 0.64,
1634
+ "eval_f1_weighted": 0.5917376665887304,
1635
+ "eval_loss": 1.4913766384124756,
1636
+ "eval_precision_macro": 0.2705775014459225,
1637
+ "eval_precision_micro": 0.64,
1638
+ "eval_precision_weighted": 0.5802396761133604,
1639
+ "eval_recall_macro": 0.3324350649350649,
1640
+ "eval_recall_micro": 0.64,
1641
+ "eval_recall_weighted": 0.64,
1642
+ "eval_runtime": 0.4059,
1643
+ "eval_samples_per_second": 246.346,
1644
+ "eval_steps_per_second": 17.244,
1645
+ "step": 453
1646
+ }
1647
+ ],
1648
+ "logging_steps": 2,
1649
+ "max_steps": 453,
1650
+ "num_input_tokens_seen": 0,
1651
+ "num_train_epochs": 3,
1652
+ "save_steps": 500,
1653
+ "total_flos": 237039835640832.0,
1654
+ "train_batch_size": 8,
1655
+ "trial_name": null,
1656
+ "trial_params": null
1657
+ }
checkpoint-453/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:28b30d78c7419f944218ae57eb5e536b3505745c5786856999a5e4567e999875
3
+ size 5048
config.json ADDED
@@ -0,0 +1,80 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "microsoft/deberta-v3-base",
3
+ "_num_labels": 20,
4
+ "architectures": [
5
+ "DebertaV2ForSequenceClassification"
6
+ ],
7
+ "attention_probs_dropout_prob": 0.1,
8
+ "hidden_act": "gelu",
9
+ "hidden_dropout_prob": 0.1,
10
+ "hidden_size": 768,
11
+ "id2label": {
12
+ "0": "ALLERGY",
13
+ "1": "ASSESSMENT",
14
+ "2": "CC",
15
+ "3": "DIAGNOSIS",
16
+ "4": "DISPOSITION",
17
+ "5": "EDCOURSE",
18
+ "6": "EXAM",
19
+ "7": "FAM/SOCHX",
20
+ "8": "GENHX",
21
+ "9": "GYNHX",
22
+ "10": "IMAGING",
23
+ "11": "IMMUNIZATIONS",
24
+ "12": "LABS",
25
+ "13": "MEDICATIONS",
26
+ "14": "OTHER_HISTORY",
27
+ "15": "PASTMEDICALHX",
28
+ "16": "PASTSURGICAL",
29
+ "17": "PLAN",
30
+ "18": "PROCEDURES",
31
+ "19": "ROS"
32
+ },
33
+ "initializer_range": 0.02,
34
+ "intermediate_size": 3072,
35
+ "label2id": {
36
+ "ALLERGY": 0,
37
+ "ASSESSMENT": 1,
38
+ "CC": 2,
39
+ "DIAGNOSIS": 3,
40
+ "DISPOSITION": 4,
41
+ "EDCOURSE": 5,
42
+ "EXAM": 6,
43
+ "FAM/SOCHX": 7,
44
+ "GENHX": 8,
45
+ "GYNHX": 9,
46
+ "IMAGING": 10,
47
+ "IMMUNIZATIONS": 11,
48
+ "LABS": 12,
49
+ "MEDICATIONS": 13,
50
+ "OTHER_HISTORY": 14,
51
+ "PASTMEDICALHX": 15,
52
+ "PASTSURGICAL": 16,
53
+ "PLAN": 17,
54
+ "PROCEDURES": 18,
55
+ "ROS": 19
56
+ },
57
+ "layer_norm_eps": 1e-07,
58
+ "max_position_embeddings": 512,
59
+ "max_relative_positions": -1,
60
+ "model_type": "deberta-v2",
61
+ "norm_rel_ebd": "layer_norm",
62
+ "num_attention_heads": 12,
63
+ "num_hidden_layers": 12,
64
+ "pad_token_id": 0,
65
+ "pooler_dropout": 0,
66
+ "pooler_hidden_act": "gelu",
67
+ "pooler_hidden_size": 768,
68
+ "pos_att_type": [
69
+ "p2c",
70
+ "c2p"
71
+ ],
72
+ "position_biased_input": false,
73
+ "position_buckets": 256,
74
+ "relative_attention": true,
75
+ "share_att_key": true,
76
+ "torch_dtype": "float32",
77
+ "transformers_version": "4.40.1",
78
+ "type_vocab_size": 0,
79
+ "vocab_size": 128100
80
+ }
model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9978b903ae04e38477cfbe4c26b677f26202043973af526dd217734af02f25a8
3
+ size 737774648
runs/Apr28_03-13-55_r-dhananjay2912-deberta-section-classifier-el2khqmv-8c8a0-vi3xw/events.out.tfevents.1714274035.r-dhananjay2912-deberta-section-classifier-el2khqmv-8c8a0-vi3xw.61.0 CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:a7f6c3e746d4d1d25aa051472672474e43074ad0e12a02848c7a5dbd90824eb4
3
- size 5755
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:6933b6b4a27dedeb2b8eb01af87227854753cfbf93cea74d8422f6ee9a72244d
3
+ size 56042
runs/Apr28_03-13-55_r-dhananjay2912-deberta-section-classifier-el2khqmv-8c8a0-vi3xw/events.out.tfevents.1714274127.r-dhananjay2912-deberta-section-classifier-el2khqmv-8c8a0-vi3xw.61.1 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:df63b97ee7735f56a342cf76e5e141e50bae4ae88f07528e38841cc018bb28b8
3
+ size 921
special_tokens_map.json ADDED
@@ -0,0 +1,15 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "bos_token": "[CLS]",
3
+ "cls_token": "[CLS]",
4
+ "eos_token": "[SEP]",
5
+ "mask_token": "[MASK]",
6
+ "pad_token": "[PAD]",
7
+ "sep_token": "[SEP]",
8
+ "unk_token": {
9
+ "content": "[UNK]",
10
+ "lstrip": false,
11
+ "normalized": true,
12
+ "rstrip": false,
13
+ "single_word": false
14
+ }
15
+ }
spm.model ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c679fbf93643d19aab7ee10c0b99e460bdbc02fedf34b92b05af343b4af586fd
3
+ size 2464616
tokenizer.json ADDED
The diff for this file is too large to render. See raw diff
 
tokenizer_config.json ADDED
@@ -0,0 +1,58 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
+ "1": {
12
+ "content": "[CLS]",
13
+ "lstrip": false,
14
+ "normalized": false,
15
+ "rstrip": false,
16
+ "single_word": false,
17
+ "special": true
18
+ },
19
+ "2": {
20
+ "content": "[SEP]",
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
+ "128000": {
36
+ "content": "[MASK]",
37
+ "lstrip": false,
38
+ "normalized": false,
39
+ "rstrip": false,
40
+ "single_word": false,
41
+ "special": true
42
+ }
43
+ },
44
+ "bos_token": "[CLS]",
45
+ "clean_up_tokenization_spaces": true,
46
+ "cls_token": "[CLS]",
47
+ "do_lower_case": false,
48
+ "eos_token": "[SEP]",
49
+ "mask_token": "[MASK]",
50
+ "model_max_length": 1000000000000000019884624838656,
51
+ "pad_token": "[PAD]",
52
+ "sep_token": "[SEP]",
53
+ "sp_model_kwargs": {},
54
+ "split_by_punct": false,
55
+ "tokenizer_class": "DebertaV2Tokenizer",
56
+ "unk_token": "[UNK]",
57
+ "vocab_type": "spm"
58
+ }
training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:28b30d78c7419f944218ae57eb5e536b3505745c5786856999a5e4567e999875
3
+ size 5048
training_params.json ADDED
@@ -0,0 +1,28 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "data_path": "autotrain-9c20u-twasm/autotrain-data",
3
+ "model": "microsoft/deberta-v3-base",
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": "autotrain-9c20u-twasm",
21
+ "auto_find_batch_size": false,
22
+ "mixed_precision": "fp16",
23
+ "save_total_limit": 1,
24
+ "push_to_hub": true,
25
+ "evaluation_strategy": "epoch",
26
+ "username": "dhananjay2912",
27
+ "log": "tensorboard"
28
+ }