Hugging Face's logo

dhananjay2912
/
clinicalbert_aci_bench_section_classifier

Text Classification
Transformers
TensorBoard
Safetensors
distilbert
Trained with AutoTrain
text-embeddings-inference
Model card Files
xet
Metrics Community
dhananjay2912 commited on
Commit
1bc010d
·
verified ·
1 Parent(s): 439352d

Upload folder using huggingface_hub

Browse files
Files changed (18) hide show
  1. README.md +37 -0
  2. checkpoint-453/config.json +71 -0
  3. checkpoint-453/model.safetensors +3 -0
  4. checkpoint-453/optimizer.pt +3 -0
  5. checkpoint-453/rng_state.pth +3 -0
  6. checkpoint-453/scheduler.pt +3 -0
  7. checkpoint-453/trainer_state.json +1657 -0
  8. checkpoint-453/training_args.bin +3 -0
  9. config.json +71 -0
  10. model.safetensors +3 -0
  11. runs/Apr28_03-26-11_r-dhananjay2912-deberta-section-classifier-d65r8w9r-def20-gpx5z/events.out.tfevents.1714274772.r-dhananjay2912-deberta-section-classifier-d65r8w9r-def20-gpx5z.63.0 +2 -2
  12. runs/Apr28_03-26-11_r-dhananjay2912-deberta-section-classifier-d65r8w9r-def20-gpx5z/events.out.tfevents.1714274854.r-dhananjay2912-deberta-section-classifier-d65r8w9r-def20-gpx5z.63.1 +3 -0
  13. special_tokens_map.json +7 -0
  14. tokenizer.json +0 -0
  15. tokenizer_config.json +58 -0
  16. training_args.bin +3 -0
  17. training_params.json +28 -0
  18. vocab.txt +0 -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-qs45q-p9syi/autotrain-data
10
+ ---
11
+
12
+ # Model Trained Using AutoTrain
13
+
14
+ - Problem type: Text Classification
15
+
16
+ ## Validation Metrics
17
+ loss: 1.0647395849227905
18
+
19
+ f1_macro: 0.4558641367469575
20
+
21
+ f1_micro: 0.74
22
+
23
+ f1_weighted: 0.7165269403625714
24
+
25
+ precision_macro: 0.47941017316017315
26
+
27
+ precision_micro: 0.74
28
+
29
+ precision_weighted: 0.7261709956709957
30
+
31
+ recall_macro: 0.4931601731601732
32
+
33
+ recall_micro: 0.74
34
+
35
+ recall_weighted: 0.74
36
+
37
+ accuracy: 0.74
checkpoint-453/config.json ADDED
@@ -0,0 +1,71 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "medicalai/ClinicalBERT",
3
+ "_num_labels": 20,
4
+ "activation": "gelu",
5
+ "architectures": [
6
+ "DistilBertForSequenceClassification"
7
+ ],
8
+ "attention_dropout": 0.1,
9
+ "dim": 768,
10
+ "dropout": 0.1,
11
+ "hidden_dim": 3072,
12
+ "id2label": {
13
+ "0": "ALLERGY",
14
+ "1": "ASSESSMENT",
15
+ "2": "CC",
16
+ "3": "DIAGNOSIS",
17
+ "4": "DISPOSITION",
18
+ "5": "EDCOURSE",
19
+ "6": "EXAM",
20
+ "7": "FAM/SOCHX",
21
+ "8": "GENHX",
22
+ "9": "GYNHX",
23
+ "10": "IMAGING",
24
+ "11": "IMMUNIZATIONS",
25
+ "12": "LABS",
26
+ "13": "MEDICATIONS",
27
+ "14": "OTHER_HISTORY",
28
+ "15": "PASTMEDICALHX",
29
+ "16": "PASTSURGICAL",
30
+ "17": "PLAN",
31
+ "18": "PROCEDURES",
32
+ "19": "ROS"
33
+ },
34
+ "initializer_range": 0.02,
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
+ "max_position_embeddings": 512,
58
+ "model_type": "distilbert",
59
+ "n_heads": 12,
60
+ "n_layers": 6,
61
+ "output_past": true,
62
+ "pad_token_id": 0,
63
+ "problem_type": "single_label_classification",
64
+ "qa_dropout": 0.1,
65
+ "seq_classif_dropout": 0.2,
66
+ "sinusoidal_pos_embds": false,
67
+ "tie_weights_": true,
68
+ "torch_dtype": "float32",
69
+ "transformers_version": "4.40.1",
70
+ "vocab_size": 119547
71
+ }
checkpoint-453/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:fad10e6d7c07618a95f6302513c82383b7479673f4b1b5770ab408d882e4de0a
3
+ size 541372744
checkpoint-453/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a8aa067891fbfde160629bd8bd892180e2d00854df60666f24a20ca7fd45a48d
3
+ size 1082807610
checkpoint-453/rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:1f1d8eb6a33f2183646af87f8fafa1f927368a2f02f288b1780b367de67212df
3
+ size 14244
checkpoint-453/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:97fdaefb041ac12ceaa625476076c5e2b9c22b8f8a001b6268056e786ae69ea9
3
+ size 1064
checkpoint-453/trainer_state.json ADDED
@@ -0,0 +1,1657 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": 1.0647395849227905,
3
+ "best_model_checkpoint": "autotrain-qs45q-p9syi/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": 3.4723896980285645,
14
+ "learning_rate": 1.3157894736842106e-06,
15
+ "loss": 3.0146,
16
+ "step": 2
17
+ },
18
+ {
19
+ "epoch": 0.026490066225165563,
20
+ "grad_norm": 3.5154001712799072,
21
+ "learning_rate": 2.631578947368421e-06,
22
+ "loss": 3.0404,
23
+ "step": 4
24
+ },
25
+ {
26
+ "epoch": 0.039735099337748346,
27
+ "grad_norm": 3.4036054611206055,
28
+ "learning_rate": 3.9473684210526315e-06,
29
+ "loss": 3.0648,
30
+ "step": 6
31
+ },
32
+ {
33
+ "epoch": 0.052980132450331126,
34
+ "grad_norm": 3.3698582649230957,
35
+ "learning_rate": 5.263157894736842e-06,
36
+ "loss": 3.0555,
37
+ "step": 8
38
+ },
39
+ {
40
+ "epoch": 0.06622516556291391,
41
+ "grad_norm": 3.074321985244751,
42
+ "learning_rate": 6.578947368421053e-06,
43
+ "loss": 3.0161,
44
+ "step": 10
45
+ },
46
+ {
47
+ "epoch": 0.07947019867549669,
48
+ "grad_norm": 3.69840407371521,
49
+ "learning_rate": 7.894736842105263e-06,
50
+ "loss": 3.0226,
51
+ "step": 12
52
+ },
53
+ {
54
+ "epoch": 0.09271523178807947,
55
+ "grad_norm": 3.0229387283325195,
56
+ "learning_rate": 9.210526315789474e-06,
57
+ "loss": 3.0117,
58
+ "step": 14
59
+ },
60
+ {
61
+ "epoch": 0.10596026490066225,
62
+ "grad_norm": 3.633251190185547,
63
+ "learning_rate": 1.0526315789473684e-05,
64
+ "loss": 2.9922,
65
+ "step": 16
66
+ },
67
+ {
68
+ "epoch": 0.11920529801324503,
69
+ "grad_norm": 3.404634952545166,
70
+ "learning_rate": 1.1842105263157895e-05,
71
+ "loss": 2.9323,
72
+ "step": 18
73
+ },
74
+ {
75
+ "epoch": 0.13245033112582782,
76
+ "grad_norm": 4.926875114440918,
77
+ "learning_rate": 1.3157894736842106e-05,
78
+ "loss": 2.9597,
79
+ "step": 20
80
+ },
81
+ {
82
+ "epoch": 0.1456953642384106,
83
+ "grad_norm": 3.341071367263794,
84
+ "learning_rate": 1.4473684210526317e-05,
85
+ "loss": 2.927,
86
+ "step": 22
87
+ },
88
+ {
89
+ "epoch": 0.15894039735099338,
90
+ "grad_norm": 3.596402645111084,
91
+ "learning_rate": 1.5789473684210526e-05,
92
+ "loss": 2.8622,
93
+ "step": 24
94
+ },
95
+ {
96
+ "epoch": 0.17218543046357615,
97
+ "grad_norm": 3.554511547088623,
98
+ "learning_rate": 1.7105263157894737e-05,
99
+ "loss": 2.8552,
100
+ "step": 26
101
+ },
102
+ {
103
+ "epoch": 0.18543046357615894,
104
+ "grad_norm": 3.871518135070801,
105
+ "learning_rate": 1.8421052631578947e-05,
106
+ "loss": 2.8545,
107
+ "step": 28
108
+ },
109
+ {
110
+ "epoch": 0.1986754966887417,
111
+ "grad_norm": 4.008641719818115,
112
+ "learning_rate": 1.9736842105263158e-05,
113
+ "loss": 2.8275,
114
+ "step": 30
115
+ },
116
+ {
117
+ "epoch": 0.2119205298013245,
118
+ "grad_norm": 4.181613445281982,
119
+ "learning_rate": 2.105263157894737e-05,
120
+ "loss": 2.7814,
121
+ "step": 32
122
+ },
123
+ {
124
+ "epoch": 0.2251655629139073,
125
+ "grad_norm": 4.396944999694824,
126
+ "learning_rate": 2.236842105263158e-05,
127
+ "loss": 2.6354,
128
+ "step": 34
129
+ },
130
+ {
131
+ "epoch": 0.23841059602649006,
132
+ "grad_norm": 4.029453754425049,
133
+ "learning_rate": 2.368421052631579e-05,
134
+ "loss": 2.5842,
135
+ "step": 36
136
+ },
137
+ {
138
+ "epoch": 0.25165562913907286,
139
+ "grad_norm": 3.9177920818328857,
140
+ "learning_rate": 2.5e-05,
141
+ "loss": 2.6199,
142
+ "step": 38
143
+ },
144
+ {
145
+ "epoch": 0.26490066225165565,
146
+ "grad_norm": 5.803523063659668,
147
+ "learning_rate": 2.6315789473684212e-05,
148
+ "loss": 2.5609,
149
+ "step": 40
150
+ },
151
+ {
152
+ "epoch": 0.2781456953642384,
153
+ "grad_norm": 6.542681694030762,
154
+ "learning_rate": 2.7631578947368426e-05,
155
+ "loss": 2.2103,
156
+ "step": 42
157
+ },
158
+ {
159
+ "epoch": 0.2913907284768212,
160
+ "grad_norm": 4.970069408416748,
161
+ "learning_rate": 2.8947368421052634e-05,
162
+ "loss": 2.5223,
163
+ "step": 44
164
+ },
165
+ {
166
+ "epoch": 0.304635761589404,
167
+ "grad_norm": 7.797888278961182,
168
+ "learning_rate": 3.0263157894736844e-05,
169
+ "loss": 2.2236,
170
+ "step": 46
171
+ },
172
+ {
173
+ "epoch": 0.31788079470198677,
174
+ "grad_norm": 5.971277713775635,
175
+ "learning_rate": 3.157894736842105e-05,
176
+ "loss": 2.6381,
177
+ "step": 48
178
+ },
179
+ {
180
+ "epoch": 0.33112582781456956,
181
+ "grad_norm": 6.939202785491943,
182
+ "learning_rate": 3.289473684210527e-05,
183
+ "loss": 2.3483,
184
+ "step": 50
185
+ },
186
+ {
187
+ "epoch": 0.3443708609271523,
188
+ "grad_norm": 5.877003192901611,
189
+ "learning_rate": 3.421052631578947e-05,
190
+ "loss": 1.9744,
191
+ "step": 52
192
+ },
193
+ {
194
+ "epoch": 0.3576158940397351,
195
+ "grad_norm": 5.517603397369385,
196
+ "learning_rate": 3.5526315789473684e-05,
197
+ "loss": 1.9791,
198
+ "step": 54
199
+ },
200
+ {
201
+ "epoch": 0.3708609271523179,
202
+ "grad_norm": 7.6599440574646,
203
+ "learning_rate": 3.6842105263157895e-05,
204
+ "loss": 2.0475,
205
+ "step": 56
206
+ },
207
+ {
208
+ "epoch": 0.3841059602649007,
209
+ "grad_norm": 6.77654504776001,
210
+ "learning_rate": 3.815789473684211e-05,
211
+ "loss": 1.923,
212
+ "step": 58
213
+ },
214
+ {
215
+ "epoch": 0.3973509933774834,
216
+ "grad_norm": 6.017172336578369,
217
+ "learning_rate": 3.9473684210526316e-05,
218
+ "loss": 1.9688,
219
+ "step": 60
220
+ },
221
+ {
222
+ "epoch": 0.4105960264900662,
223
+ "grad_norm": 5.819124221801758,
224
+ "learning_rate": 4.078947368421053e-05,
225
+ "loss": 2.2777,
226
+ "step": 62
227
+ },
228
+ {
229
+ "epoch": 0.423841059602649,
230
+ "grad_norm": 5.240574359893799,
231
+ "learning_rate": 4.210526315789474e-05,
232
+ "loss": 1.7219,
233
+ "step": 64
234
+ },
235
+ {
236
+ "epoch": 0.4370860927152318,
237
+ "grad_norm": 7.376441478729248,
238
+ "learning_rate": 4.342105263157895e-05,
239
+ "loss": 1.5784,
240
+ "step": 66
241
+ },
242
+ {
243
+ "epoch": 0.4503311258278146,
244
+ "grad_norm": 5.029400825500488,
245
+ "learning_rate": 4.473684210526316e-05,
246
+ "loss": 1.6895,
247
+ "step": 68
248
+ },
249
+ {
250
+ "epoch": 0.46357615894039733,
251
+ "grad_norm": 10.123241424560547,
252
+ "learning_rate": 4.605263157894737e-05,
253
+ "loss": 1.6955,
254
+ "step": 70
255
+ },
256
+ {
257
+ "epoch": 0.4768211920529801,
258
+ "grad_norm": 6.67560338973999,
259
+ "learning_rate": 4.736842105263158e-05,
260
+ "loss": 1.9591,
261
+ "step": 72
262
+ },
263
+ {
264
+ "epoch": 0.4900662251655629,
265
+ "grad_norm": 8.467432975769043,
266
+ "learning_rate": 4.868421052631579e-05,
267
+ "loss": 2.0633,
268
+ "step": 74
269
+ },
270
+ {
271
+ "epoch": 0.5033112582781457,
272
+ "grad_norm": 5.2431535720825195,
273
+ "learning_rate": 5e-05,
274
+ "loss": 1.3436,
275
+ "step": 76
276
+ },
277
+ {
278
+ "epoch": 0.5165562913907285,
279
+ "grad_norm": 7.770199298858643,
280
+ "learning_rate": 4.985272459499264e-05,
281
+ "loss": 1.9379,
282
+ "step": 78
283
+ },
284
+ {
285
+ "epoch": 0.5298013245033113,
286
+ "grad_norm": 5.734556198120117,
287
+ "learning_rate": 4.9705449189985276e-05,
288
+ "loss": 1.7279,
289
+ "step": 80
290
+ },
291
+ {
292
+ "epoch": 0.543046357615894,
293
+ "grad_norm": 6.759193420410156,
294
+ "learning_rate": 4.955817378497791e-05,
295
+ "loss": 1.4606,
296
+ "step": 82
297
+ },
298
+ {
299
+ "epoch": 0.5562913907284768,
300
+ "grad_norm": 10.67810344696045,
301
+ "learning_rate": 4.941089837997055e-05,
302
+ "loss": 1.3069,
303
+ "step": 84
304
+ },
305
+ {
306
+ "epoch": 0.5695364238410596,
307
+ "grad_norm": 7.0206098556518555,
308
+ "learning_rate": 4.926362297496319e-05,
309
+ "loss": 1.8698,
310
+ "step": 86
311
+ },
312
+ {
313
+ "epoch": 0.5827814569536424,
314
+ "grad_norm": 4.365483283996582,
315
+ "learning_rate": 4.911634756995582e-05,
316
+ "loss": 1.5244,
317
+ "step": 88
318
+ },
319
+ {
320
+ "epoch": 0.5960264900662252,
321
+ "grad_norm": 8.567715644836426,
322
+ "learning_rate": 4.8969072164948454e-05,
323
+ "loss": 1.7503,
324
+ "step": 90
325
+ },
326
+ {
327
+ "epoch": 0.609271523178808,
328
+ "grad_norm": 7.213935852050781,
329
+ "learning_rate": 4.882179675994109e-05,
330
+ "loss": 1.6514,
331
+ "step": 92
332
+ },
333
+ {
334
+ "epoch": 0.6225165562913907,
335
+ "grad_norm": 5.607529163360596,
336
+ "learning_rate": 4.867452135493373e-05,
337
+ "loss": 1.0184,
338
+ "step": 94
339
+ },
340
+ {
341
+ "epoch": 0.6357615894039735,
342
+ "grad_norm": 9.907747268676758,
343
+ "learning_rate": 4.8527245949926364e-05,
344
+ "loss": 1.7225,
345
+ "step": 96
346
+ },
347
+ {
348
+ "epoch": 0.6490066225165563,
349
+ "grad_norm": 13.226700782775879,
350
+ "learning_rate": 4.8379970544919e-05,
351
+ "loss": 1.9432,
352
+ "step": 98
353
+ },
354
+ {
355
+ "epoch": 0.6622516556291391,
356
+ "grad_norm": 11.118176460266113,
357
+ "learning_rate": 4.823269513991164e-05,
358
+ "loss": 1.416,
359
+ "step": 100
360
+ },
361
+ {
362
+ "epoch": 0.6754966887417219,
363
+ "grad_norm": 10.29113483428955,
364
+ "learning_rate": 4.8085419734904275e-05,
365
+ "loss": 1.6586,
366
+ "step": 102
367
+ },
368
+ {
369
+ "epoch": 0.6887417218543046,
370
+ "grad_norm": 9.081689834594727,
371
+ "learning_rate": 4.793814432989691e-05,
372
+ "loss": 1.305,
373
+ "step": 104
374
+ },
375
+ {
376
+ "epoch": 0.7019867549668874,
377
+ "grad_norm": 6.368233680725098,
378
+ "learning_rate": 4.779086892488955e-05,
379
+ "loss": 1.1645,
380
+ "step": 106
381
+ },
382
+ {
383
+ "epoch": 0.7152317880794702,
384
+ "grad_norm": 12.125802040100098,
385
+ "learning_rate": 4.764359351988218e-05,
386
+ "loss": 0.8825,
387
+ "step": 108
388
+ },
389
+ {
390
+ "epoch": 0.7284768211920529,
391
+ "grad_norm": 14.471185684204102,
392
+ "learning_rate": 4.7496318114874815e-05,
393
+ "loss": 1.3133,
394
+ "step": 110
395
+ },
396
+ {
397
+ "epoch": 0.7417218543046358,
398
+ "grad_norm": 10.047285079956055,
399
+ "learning_rate": 4.734904270986745e-05,
400
+ "loss": 1.0271,
401
+ "step": 112
402
+ },
403
+ {
404
+ "epoch": 0.7549668874172185,
405
+ "grad_norm": 10.376700401306152,
406
+ "learning_rate": 4.720176730486009e-05,
407
+ "loss": 1.5487,
408
+ "step": 114
409
+ },
410
+ {
411
+ "epoch": 0.7682119205298014,
412
+ "grad_norm": 9.097859382629395,
413
+ "learning_rate": 4.7054491899852726e-05,
414
+ "loss": 1.7211,
415
+ "step": 116
416
+ },
417
+ {
418
+ "epoch": 0.7814569536423841,
419
+ "grad_norm": 11.168022155761719,
420
+ "learning_rate": 4.690721649484536e-05,
421
+ "loss": 1.3699,
422
+ "step": 118
423
+ },
424
+ {
425
+ "epoch": 0.7947019867549668,
426
+ "grad_norm": 7.571075439453125,
427
+ "learning_rate": 4.6759941089838e-05,
428
+ "loss": 0.7628,
429
+ "step": 120
430
+ },
431
+ {
432
+ "epoch": 0.8079470198675497,
433
+ "grad_norm": 7.332944393157959,
434
+ "learning_rate": 4.661266568483064e-05,
435
+ "loss": 1.2271,
436
+ "step": 122
437
+ },
438
+ {
439
+ "epoch": 0.8211920529801324,
440
+ "grad_norm": 9.472811698913574,
441
+ "learning_rate": 4.6465390279823274e-05,
442
+ "loss": 0.9225,
443
+ "step": 124
444
+ },
445
+ {
446
+ "epoch": 0.8344370860927153,
447
+ "grad_norm": 7.298526287078857,
448
+ "learning_rate": 4.631811487481591e-05,
449
+ "loss": 1.4189,
450
+ "step": 126
451
+ },
452
+ {
453
+ "epoch": 0.847682119205298,
454
+ "grad_norm": 6.718891620635986,
455
+ "learning_rate": 4.617083946980855e-05,
456
+ "loss": 0.9269,
457
+ "step": 128
458
+ },
459
+ {
460
+ "epoch": 0.8609271523178808,
461
+ "grad_norm": 5.722424507141113,
462
+ "learning_rate": 4.602356406480118e-05,
463
+ "loss": 1.5321,
464
+ "step": 130
465
+ },
466
+ {
467
+ "epoch": 0.8741721854304636,
468
+ "grad_norm": 8.258366584777832,
469
+ "learning_rate": 4.5876288659793814e-05,
470
+ "loss": 0.9162,
471
+ "step": 132
472
+ },
473
+ {
474
+ "epoch": 0.8874172185430463,
475
+ "grad_norm": 6.919400691986084,
476
+ "learning_rate": 4.572901325478645e-05,
477
+ "loss": 1.0685,
478
+ "step": 134
479
+ },
480
+ {
481
+ "epoch": 0.9006622516556292,
482
+ "grad_norm": 6.770501613616943,
483
+ "learning_rate": 4.558173784977909e-05,
484
+ "loss": 0.7407,
485
+ "step": 136
486
+ },
487
+ {
488
+ "epoch": 0.9139072847682119,
489
+ "grad_norm": 6.387173652648926,
490
+ "learning_rate": 4.5434462444771725e-05,
491
+ "loss": 0.6707,
492
+ "step": 138
493
+ },
494
+ {
495
+ "epoch": 0.9271523178807947,
496
+ "grad_norm": 11.18732738494873,
497
+ "learning_rate": 4.528718703976436e-05,
498
+ "loss": 1.3069,
499
+ "step": 140
500
+ },
501
+ {
502
+ "epoch": 0.9403973509933775,
503
+ "grad_norm": 4.17434549331665,
504
+ "learning_rate": 4.5139911634757e-05,
505
+ "loss": 0.7223,
506
+ "step": 142
507
+ },
508
+ {
509
+ "epoch": 0.9536423841059603,
510
+ "grad_norm": 7.86276388168335,
511
+ "learning_rate": 4.4992636229749635e-05,
512
+ "loss": 0.7461,
513
+ "step": 144
514
+ },
515
+ {
516
+ "epoch": 0.9668874172185431,
517
+ "grad_norm": 11.436422348022461,
518
+ "learning_rate": 4.484536082474227e-05,
519
+ "loss": 1.4973,
520
+ "step": 146
521
+ },
522
+ {
523
+ "epoch": 0.9801324503311258,
524
+ "grad_norm": 11.492631912231445,
525
+ "learning_rate": 4.469808541973491e-05,
526
+ "loss": 1.1266,
527
+ "step": 148
528
+ },
529
+ {
530
+ "epoch": 0.9933774834437086,
531
+ "grad_norm": 6.025942802429199,
532
+ "learning_rate": 4.4550810014727546e-05,
533
+ "loss": 0.5809,
534
+ "step": 150
535
+ },
536
+ {
537
+ "epoch": 1.0,
538
+ "eval_accuracy": 0.62,
539
+ "eval_f1_macro": 0.2182729551150604,
540
+ "eval_f1_micro": 0.62,
541
+ "eval_f1_weighted": 0.5708288904078378,
542
+ "eval_loss": 1.3821334838867188,
543
+ "eval_precision_macro": 0.20881118881118882,
544
+ "eval_precision_micro": 0.62,
545
+ "eval_precision_weighted": 0.5551748251748252,
546
+ "eval_recall_macro": 0.25717532467532467,
547
+ "eval_recall_micro": 0.62,
548
+ "eval_recall_weighted": 0.62,
549
+ "eval_runtime": 0.1587,
550
+ "eval_samples_per_second": 630.071,
551
+ "eval_steps_per_second": 44.105,
552
+ "step": 151
553
+ },
554
+ {
555
+ "epoch": 1.0066225165562914,
556
+ "grad_norm": 6.717867374420166,
557
+ "learning_rate": 4.447717231222386e-05,
558
+ "loss": 0.8661,
559
+ "step": 152
560
+ },
561
+ {
562
+ "epoch": 1.0198675496688743,
563
+ "grad_norm": 5.969238758087158,
564
+ "learning_rate": 4.4329896907216494e-05,
565
+ "loss": 0.6295,
566
+ "step": 154
567
+ },
568
+ {
569
+ "epoch": 1.033112582781457,
570
+ "grad_norm": 5.343296527862549,
571
+ "learning_rate": 4.418262150220913e-05,
572
+ "loss": 1.1331,
573
+ "step": 156
574
+ },
575
+ {
576
+ "epoch": 1.0463576158940397,
577
+ "grad_norm": 5.666139602661133,
578
+ "learning_rate": 4.403534609720177e-05,
579
+ "loss": 0.8978,
580
+ "step": 158
581
+ },
582
+ {
583
+ "epoch": 1.0596026490066226,
584
+ "grad_norm": 5.301098823547363,
585
+ "learning_rate": 4.3888070692194405e-05,
586
+ "loss": 0.6172,
587
+ "step": 160
588
+ },
589
+ {
590
+ "epoch": 1.0728476821192052,
591
+ "grad_norm": 6.197515964508057,
592
+ "learning_rate": 4.374079528718704e-05,
593
+ "loss": 0.6356,
594
+ "step": 162
595
+ },
596
+ {
597
+ "epoch": 1.086092715231788,
598
+ "grad_norm": 9.58243179321289,
599
+ "learning_rate": 4.359351988217968e-05,
600
+ "loss": 1.5982,
601
+ "step": 164
602
+ },
603
+ {
604
+ "epoch": 1.099337748344371,
605
+ "grad_norm": 5.4112629890441895,
606
+ "learning_rate": 4.3446244477172316e-05,
607
+ "loss": 0.5373,
608
+ "step": 166
609
+ },
610
+ {
611
+ "epoch": 1.1125827814569536,
612
+ "grad_norm": 6.890096664428711,
613
+ "learning_rate": 4.329896907216495e-05,
614
+ "loss": 1.594,
615
+ "step": 168
616
+ },
617
+ {
618
+ "epoch": 1.1258278145695364,
619
+ "grad_norm": 13.663280487060547,
620
+ "learning_rate": 4.315169366715759e-05,
621
+ "loss": 0.6526,
622
+ "step": 170
623
+ },
624
+ {
625
+ "epoch": 1.1390728476821192,
626
+ "grad_norm": 5.038116455078125,
627
+ "learning_rate": 4.3004418262150226e-05,
628
+ "loss": 0.8239,
629
+ "step": 172
630
+ },
631
+ {
632
+ "epoch": 1.152317880794702,
633
+ "grad_norm": 10.813724517822266,
634
+ "learning_rate": 4.2857142857142856e-05,
635
+ "loss": 0.9156,
636
+ "step": 174
637
+ },
638
+ {
639
+ "epoch": 1.1655629139072847,
640
+ "grad_norm": 3.2282865047454834,
641
+ "learning_rate": 4.270986745213549e-05,
642
+ "loss": 0.6795,
643
+ "step": 176
644
+ },
645
+ {
646
+ "epoch": 1.1788079470198676,
647
+ "grad_norm": 5.9910197257995605,
648
+ "learning_rate": 4.256259204712813e-05,
649
+ "loss": 1.0571,
650
+ "step": 178
651
+ },
652
+ {
653
+ "epoch": 1.1920529801324504,
654
+ "grad_norm": 4.989276885986328,
655
+ "learning_rate": 4.241531664212077e-05,
656
+ "loss": 1.1591,
657
+ "step": 180
658
+ },
659
+ {
660
+ "epoch": 1.205298013245033,
661
+ "grad_norm": 12.654097557067871,
662
+ "learning_rate": 4.2268041237113404e-05,
663
+ "loss": 1.3218,
664
+ "step": 182
665
+ },
666
+ {
667
+ "epoch": 1.218543046357616,
668
+ "grad_norm": 4.979647636413574,
669
+ "learning_rate": 4.212076583210604e-05,
670
+ "loss": 0.6925,
671
+ "step": 184
672
+ },
673
+ {
674
+ "epoch": 1.2317880794701987,
675
+ "grad_norm": 9.59062385559082,
676
+ "learning_rate": 4.197349042709868e-05,
677
+ "loss": 1.2088,
678
+ "step": 186
679
+ },
680
+ {
681
+ "epoch": 1.2450331125827814,
682
+ "grad_norm": 8.358089447021484,
683
+ "learning_rate": 4.1826215022091314e-05,
684
+ "loss": 0.7255,
685
+ "step": 188
686
+ },
687
+ {
688
+ "epoch": 1.2582781456953642,
689
+ "grad_norm": 1.6286242008209229,
690
+ "learning_rate": 4.167893961708395e-05,
691
+ "loss": 0.6235,
692
+ "step": 190
693
+ },
694
+ {
695
+ "epoch": 1.271523178807947,
696
+ "grad_norm": 7.9347662925720215,
697
+ "learning_rate": 4.153166421207659e-05,
698
+ "loss": 0.964,
699
+ "step": 192
700
+ },
701
+ {
702
+ "epoch": 1.2847682119205297,
703
+ "grad_norm": 6.71475887298584,
704
+ "learning_rate": 4.138438880706922e-05,
705
+ "loss": 0.9162,
706
+ "step": 194
707
+ },
708
+ {
709
+ "epoch": 1.2980132450331126,
710
+ "grad_norm": 2.8124191761016846,
711
+ "learning_rate": 4.1237113402061855e-05,
712
+ "loss": 0.3924,
713
+ "step": 196
714
+ },
715
+ {
716
+ "epoch": 1.3112582781456954,
717
+ "grad_norm": 8.39048957824707,
718
+ "learning_rate": 4.108983799705449e-05,
719
+ "loss": 0.68,
720
+ "step": 198
721
+ },
722
+ {
723
+ "epoch": 1.3245033112582782,
724
+ "grad_norm": 10.033766746520996,
725
+ "learning_rate": 4.094256259204713e-05,
726
+ "loss": 0.6962,
727
+ "step": 200
728
+ },
729
+ {
730
+ "epoch": 1.3377483443708609,
731
+ "grad_norm": 6.6539106369018555,
732
+ "learning_rate": 4.0795287187039766e-05,
733
+ "loss": 1.0892,
734
+ "step": 202
735
+ },
736
+ {
737
+ "epoch": 1.3509933774834437,
738
+ "grad_norm": 5.977195739746094,
739
+ "learning_rate": 4.06480117820324e-05,
740
+ "loss": 0.5886,
741
+ "step": 204
742
+ },
743
+ {
744
+ "epoch": 1.3642384105960264,
745
+ "grad_norm": 5.548197269439697,
746
+ "learning_rate": 4.050073637702504e-05,
747
+ "loss": 0.9609,
748
+ "step": 206
749
+ },
750
+ {
751
+ "epoch": 1.3774834437086092,
752
+ "grad_norm": 6.958375453948975,
753
+ "learning_rate": 4.0353460972017676e-05,
754
+ "loss": 1.0345,
755
+ "step": 208
756
+ },
757
+ {
758
+ "epoch": 1.390728476821192,
759
+ "grad_norm": 14.204687118530273,
760
+ "learning_rate": 4.020618556701031e-05,
761
+ "loss": 0.7327,
762
+ "step": 210
763
+ },
764
+ {
765
+ "epoch": 1.403973509933775,
766
+ "grad_norm": 8.209890365600586,
767
+ "learning_rate": 4.005891016200295e-05,
768
+ "loss": 0.5259,
769
+ "step": 212
770
+ },
771
+ {
772
+ "epoch": 1.4172185430463577,
773
+ "grad_norm": 6.424736976623535,
774
+ "learning_rate": 3.991163475699559e-05,
775
+ "loss": 0.9341,
776
+ "step": 214
777
+ },
778
+ {
779
+ "epoch": 1.4304635761589404,
780
+ "grad_norm": 7.625380039215088,
781
+ "learning_rate": 3.976435935198822e-05,
782
+ "loss": 1.1303,
783
+ "step": 216
784
+ },
785
+ {
786
+ "epoch": 1.4437086092715232,
787
+ "grad_norm": 7.861330509185791,
788
+ "learning_rate": 3.9617083946980854e-05,
789
+ "loss": 0.5599,
790
+ "step": 218
791
+ },
792
+ {
793
+ "epoch": 1.4569536423841059,
794
+ "grad_norm": 11.318681716918945,
795
+ "learning_rate": 3.946980854197349e-05,
796
+ "loss": 0.665,
797
+ "step": 220
798
+ },
799
+ {
800
+ "epoch": 1.4701986754966887,
801
+ "grad_norm": 8.797566413879395,
802
+ "learning_rate": 3.932253313696613e-05,
803
+ "loss": 0.9297,
804
+ "step": 222
805
+ },
806
+ {
807
+ "epoch": 1.4834437086092715,
808
+ "grad_norm": 9.536187171936035,
809
+ "learning_rate": 3.9175257731958764e-05,
810
+ "loss": 0.5168,
811
+ "step": 224
812
+ },
813
+ {
814
+ "epoch": 1.4966887417218544,
815
+ "grad_norm": 5.4594807624816895,
816
+ "learning_rate": 3.90279823269514e-05,
817
+ "loss": 0.6562,
818
+ "step": 226
819
+ },
820
+ {
821
+ "epoch": 1.5099337748344372,
822
+ "grad_norm": 6.416939735412598,
823
+ "learning_rate": 3.888070692194404e-05,
824
+ "loss": 0.4248,
825
+ "step": 228
826
+ },
827
+ {
828
+ "epoch": 1.5231788079470199,
829
+ "grad_norm": 8.049701690673828,
830
+ "learning_rate": 3.8733431516936675e-05,
831
+ "loss": 0.6311,
832
+ "step": 230
833
+ },
834
+ {
835
+ "epoch": 1.5364238410596025,
836
+ "grad_norm": 6.978274345397949,
837
+ "learning_rate": 3.858615611192931e-05,
838
+ "loss": 0.5238,
839
+ "step": 232
840
+ },
841
+ {
842
+ "epoch": 1.5496688741721854,
843
+ "grad_norm": 4.730326175689697,
844
+ "learning_rate": 3.843888070692195e-05,
845
+ "loss": 0.9127,
846
+ "step": 234
847
+ },
848
+ {
849
+ "epoch": 1.5629139072847682,
850
+ "grad_norm": 6.017629623413086,
851
+ "learning_rate": 3.836524300441827e-05,
852
+ "loss": 0.6197,
853
+ "step": 236
854
+ },
855
+ {
856
+ "epoch": 1.576158940397351,
857
+ "grad_norm": 1.4615082740783691,
858
+ "learning_rate": 3.82179675994109e-05,
859
+ "loss": 0.1855,
860
+ "step": 238
861
+ },
862
+ {
863
+ "epoch": 1.589403973509934,
864
+ "grad_norm": 17.785310745239258,
865
+ "learning_rate": 3.8070692194403534e-05,
866
+ "loss": 1.4609,
867
+ "step": 240
868
+ },
869
+ {
870
+ "epoch": 1.6026490066225165,
871
+ "grad_norm": 11.241979598999023,
872
+ "learning_rate": 3.792341678939617e-05,
873
+ "loss": 0.8828,
874
+ "step": 242
875
+ },
876
+ {
877
+ "epoch": 1.6158940397350994,
878
+ "grad_norm": 13.630949974060059,
879
+ "learning_rate": 3.777614138438881e-05,
880
+ "loss": 1.374,
881
+ "step": 244
882
+ },
883
+ {
884
+ "epoch": 1.629139072847682,
885
+ "grad_norm": 13.52291488647461,
886
+ "learning_rate": 3.7628865979381445e-05,
887
+ "loss": 0.7898,
888
+ "step": 246
889
+ },
890
+ {
891
+ "epoch": 1.6423841059602649,
892
+ "grad_norm": 6.53799295425415,
893
+ "learning_rate": 3.748159057437408e-05,
894
+ "loss": 0.7232,
895
+ "step": 248
896
+ },
897
+ {
898
+ "epoch": 1.6556291390728477,
899
+ "grad_norm": 7.03351354598999,
900
+ "learning_rate": 3.733431516936672e-05,
901
+ "loss": 1.5363,
902
+ "step": 250
903
+ },
904
+ {
905
+ "epoch": 1.6688741721854305,
906
+ "grad_norm": 10.971170425415039,
907
+ "learning_rate": 3.7187039764359355e-05,
908
+ "loss": 0.3952,
909
+ "step": 252
910
+ },
911
+ {
912
+ "epoch": 1.6821192052980134,
913
+ "grad_norm": 10.918181419372559,
914
+ "learning_rate": 3.703976435935199e-05,
915
+ "loss": 1.0573,
916
+ "step": 254
917
+ },
918
+ {
919
+ "epoch": 1.695364238410596,
920
+ "grad_norm": 5.845329284667969,
921
+ "learning_rate": 3.689248895434463e-05,
922
+ "loss": 0.9152,
923
+ "step": 256
924
+ },
925
+ {
926
+ "epoch": 1.7086092715231787,
927
+ "grad_norm": 10.043814659118652,
928
+ "learning_rate": 3.6745213549337266e-05,
929
+ "loss": 0.7483,
930
+ "step": 258
931
+ },
932
+ {
933
+ "epoch": 1.7218543046357615,
934
+ "grad_norm": 9.504009246826172,
935
+ "learning_rate": 3.6597938144329896e-05,
936
+ "loss": 1.3825,
937
+ "step": 260
938
+ },
939
+ {
940
+ "epoch": 1.7350993377483444,
941
+ "grad_norm": 7.6096415519714355,
942
+ "learning_rate": 3.645066273932253e-05,
943
+ "loss": 0.3722,
944
+ "step": 262
945
+ },
946
+ {
947
+ "epoch": 1.7483443708609272,
948
+ "grad_norm": 15.977928161621094,
949
+ "learning_rate": 3.630338733431517e-05,
950
+ "loss": 0.7817,
951
+ "step": 264
952
+ },
953
+ {
954
+ "epoch": 1.76158940397351,
955
+ "grad_norm": 12.720684051513672,
956
+ "learning_rate": 3.6156111929307806e-05,
957
+ "loss": 0.5112,
958
+ "step": 266
959
+ },
960
+ {
961
+ "epoch": 1.7748344370860927,
962
+ "grad_norm": 8.58932113647461,
963
+ "learning_rate": 3.600883652430044e-05,
964
+ "loss": 0.7,
965
+ "step": 268
966
+ },
967
+ {
968
+ "epoch": 1.7880794701986755,
969
+ "grad_norm": 13.423791885375977,
970
+ "learning_rate": 3.586156111929308e-05,
971
+ "loss": 1.6471,
972
+ "step": 270
973
+ },
974
+ {
975
+ "epoch": 1.8013245033112582,
976
+ "grad_norm": 6.468234062194824,
977
+ "learning_rate": 3.571428571428572e-05,
978
+ "loss": 0.66,
979
+ "step": 272
980
+ },
981
+ {
982
+ "epoch": 1.814569536423841,
983
+ "grad_norm": 9.493569374084473,
984
+ "learning_rate": 3.5567010309278354e-05,
985
+ "loss": 1.1906,
986
+ "step": 274
987
+ },
988
+ {
989
+ "epoch": 1.8278145695364238,
990
+ "grad_norm": 4.703698635101318,
991
+ "learning_rate": 3.541973490427099e-05,
992
+ "loss": 0.4843,
993
+ "step": 276
994
+ },
995
+ {
996
+ "epoch": 1.8410596026490067,
997
+ "grad_norm": 7.946201324462891,
998
+ "learning_rate": 3.527245949926363e-05,
999
+ "loss": 0.5529,
1000
+ "step": 278
1001
+ },
1002
+ {
1003
+ "epoch": 1.8543046357615895,
1004
+ "grad_norm": 6.270992279052734,
1005
+ "learning_rate": 3.512518409425626e-05,
1006
+ "loss": 0.9139,
1007
+ "step": 280
1008
+ },
1009
+ {
1010
+ "epoch": 1.8675496688741722,
1011
+ "grad_norm": 12.685065269470215,
1012
+ "learning_rate": 3.4977908689248894e-05,
1013
+ "loss": 0.4489,
1014
+ "step": 282
1015
+ },
1016
+ {
1017
+ "epoch": 1.8807947019867548,
1018
+ "grad_norm": 9.88779354095459,
1019
+ "learning_rate": 3.483063328424153e-05,
1020
+ "loss": 1.0594,
1021
+ "step": 284
1022
+ },
1023
+ {
1024
+ "epoch": 1.8940397350993377,
1025
+ "grad_norm": 15.420830726623535,
1026
+ "learning_rate": 3.468335787923417e-05,
1027
+ "loss": 1.5748,
1028
+ "step": 286
1029
+ },
1030
+ {
1031
+ "epoch": 1.9072847682119205,
1032
+ "grad_norm": 12.748838424682617,
1033
+ "learning_rate": 3.4536082474226805e-05,
1034
+ "loss": 0.3906,
1035
+ "step": 288
1036
+ },
1037
+ {
1038
+ "epoch": 1.9205298013245033,
1039
+ "grad_norm": 8.6655912399292,
1040
+ "learning_rate": 3.438880706921944e-05,
1041
+ "loss": 0.7372,
1042
+ "step": 290
1043
+ },
1044
+ {
1045
+ "epoch": 1.9337748344370862,
1046
+ "grad_norm": 2.1088364124298096,
1047
+ "learning_rate": 3.424153166421208e-05,
1048
+ "loss": 0.6368,
1049
+ "step": 292
1050
+ },
1051
+ {
1052
+ "epoch": 1.9470198675496688,
1053
+ "grad_norm": 5.693451404571533,
1054
+ "learning_rate": 3.4094256259204716e-05,
1055
+ "loss": 1.1065,
1056
+ "step": 294
1057
+ },
1058
+ {
1059
+ "epoch": 1.9602649006622517,
1060
+ "grad_norm": 5.4681315422058105,
1061
+ "learning_rate": 3.394698085419735e-05,
1062
+ "loss": 0.2881,
1063
+ "step": 296
1064
+ },
1065
+ {
1066
+ "epoch": 1.9735099337748343,
1067
+ "grad_norm": 3.1230578422546387,
1068
+ "learning_rate": 3.379970544918999e-05,
1069
+ "loss": 0.7056,
1070
+ "step": 298
1071
+ },
1072
+ {
1073
+ "epoch": 1.9867549668874172,
1074
+ "grad_norm": 8.198813438415527,
1075
+ "learning_rate": 3.3652430044182626e-05,
1076
+ "loss": 0.3768,
1077
+ "step": 300
1078
+ },
1079
+ {
1080
+ "epoch": 2.0,
1081
+ "grad_norm": 5.530179023742676,
1082
+ "learning_rate": 3.3505154639175256e-05,
1083
+ "loss": 0.3156,
1084
+ "step": 302
1085
+ },
1086
+ {
1087
+ "epoch": 2.0,
1088
+ "eval_accuracy": 0.74,
1089
+ "eval_f1_macro": 0.45774191750278714,
1090
+ "eval_f1_micro": 0.74,
1091
+ "eval_f1_weighted": 0.7020952062430323,
1092
+ "eval_loss": 1.118362307548523,
1093
+ "eval_precision_macro": 0.5131296992481202,
1094
+ "eval_precision_micro": 0.74,
1095
+ "eval_precision_weighted": 0.7261616541353384,
1096
+ "eval_recall_macro": 0.48530303030303035,
1097
+ "eval_recall_micro": 0.74,
1098
+ "eval_recall_weighted": 0.74,
1099
+ "eval_runtime": 0.1632,
1100
+ "eval_samples_per_second": 612.634,
1101
+ "eval_steps_per_second": 42.884,
1102
+ "step": 302
1103
+ },
1104
+ {
1105
+ "epoch": 2.013245033112583,
1106
+ "grad_norm": 1.5750505924224854,
1107
+ "learning_rate": 3.335787923416789e-05,
1108
+ "loss": 0.1854,
1109
+ "step": 304
1110
+ },
1111
+ {
1112
+ "epoch": 2.0264900662251657,
1113
+ "grad_norm": 3.4210903644561768,
1114
+ "learning_rate": 3.321060382916053e-05,
1115
+ "loss": 0.3148,
1116
+ "step": 306
1117
+ },
1118
+ {
1119
+ "epoch": 2.0397350993377485,
1120
+ "grad_norm": 7.4305033683776855,
1121
+ "learning_rate": 3.306332842415317e-05,
1122
+ "loss": 0.7654,
1123
+ "step": 308
1124
+ },
1125
+ {
1126
+ "epoch": 2.052980132450331,
1127
+ "grad_norm": 2.1427576541900635,
1128
+ "learning_rate": 3.2916053019145804e-05,
1129
+ "loss": 0.2679,
1130
+ "step": 310
1131
+ },
1132
+ {
1133
+ "epoch": 2.066225165562914,
1134
+ "grad_norm": 4.614518165588379,
1135
+ "learning_rate": 3.276877761413844e-05,
1136
+ "loss": 0.722,
1137
+ "step": 312
1138
+ },
1139
+ {
1140
+ "epoch": 2.0794701986754967,
1141
+ "grad_norm": 5.343958377838135,
1142
+ "learning_rate": 3.262150220913108e-05,
1143
+ "loss": 0.658,
1144
+ "step": 314
1145
+ },
1146
+ {
1147
+ "epoch": 2.0927152317880795,
1148
+ "grad_norm": 7.354612827301025,
1149
+ "learning_rate": 3.2474226804123714e-05,
1150
+ "loss": 0.5283,
1151
+ "step": 316
1152
+ },
1153
+ {
1154
+ "epoch": 2.1059602649006623,
1155
+ "grad_norm": 3.1576592922210693,
1156
+ "learning_rate": 3.232695139911635e-05,
1157
+ "loss": 0.096,
1158
+ "step": 318
1159
+ },
1160
+ {
1161
+ "epoch": 2.119205298013245,
1162
+ "grad_norm": 3.2874159812927246,
1163
+ "learning_rate": 3.217967599410899e-05,
1164
+ "loss": 0.2096,
1165
+ "step": 320
1166
+ },
1167
+ {
1168
+ "epoch": 2.1324503311258276,
1169
+ "grad_norm": 0.5562194585800171,
1170
+ "learning_rate": 3.2032400589101625e-05,
1171
+ "loss": 0.1196,
1172
+ "step": 322
1173
+ },
1174
+ {
1175
+ "epoch": 2.1456953642384105,
1176
+ "grad_norm": 13.274640083312988,
1177
+ "learning_rate": 3.1885125184094255e-05,
1178
+ "loss": 1.067,
1179
+ "step": 324
1180
+ },
1181
+ {
1182
+ "epoch": 2.1589403973509933,
1183
+ "grad_norm": 8.369612693786621,
1184
+ "learning_rate": 3.173784977908689e-05,
1185
+ "loss": 0.6852,
1186
+ "step": 326
1187
+ },
1188
+ {
1189
+ "epoch": 2.172185430463576,
1190
+ "grad_norm": 10.826526641845703,
1191
+ "learning_rate": 3.159057437407953e-05,
1192
+ "loss": 0.5073,
1193
+ "step": 328
1194
+ },
1195
+ {
1196
+ "epoch": 2.185430463576159,
1197
+ "grad_norm": 6.016572952270508,
1198
+ "learning_rate": 3.1443298969072166e-05,
1199
+ "loss": 0.5949,
1200
+ "step": 330
1201
+ },
1202
+ {
1203
+ "epoch": 2.198675496688742,
1204
+ "grad_norm": 10.392781257629395,
1205
+ "learning_rate": 3.12960235640648e-05,
1206
+ "loss": 0.7787,
1207
+ "step": 332
1208
+ },
1209
+ {
1210
+ "epoch": 2.2119205298013247,
1211
+ "grad_norm": 4.342617511749268,
1212
+ "learning_rate": 3.114874815905744e-05,
1213
+ "loss": 0.4602,
1214
+ "step": 334
1215
+ },
1216
+ {
1217
+ "epoch": 2.225165562913907,
1218
+ "grad_norm": 7.711668014526367,
1219
+ "learning_rate": 3.1001472754050076e-05,
1220
+ "loss": 0.3792,
1221
+ "step": 336
1222
+ },
1223
+ {
1224
+ "epoch": 2.23841059602649,
1225
+ "grad_norm": 5.162533283233643,
1226
+ "learning_rate": 3.085419734904271e-05,
1227
+ "loss": 0.2087,
1228
+ "step": 338
1229
+ },
1230
+ {
1231
+ "epoch": 2.251655629139073,
1232
+ "grad_norm": 7.244391441345215,
1233
+ "learning_rate": 3.070692194403535e-05,
1234
+ "loss": 0.298,
1235
+ "step": 340
1236
+ },
1237
+ {
1238
+ "epoch": 2.2649006622516556,
1239
+ "grad_norm": 10.348755836486816,
1240
+ "learning_rate": 3.055964653902799e-05,
1241
+ "loss": 0.3321,
1242
+ "step": 342
1243
+ },
1244
+ {
1245
+ "epoch": 2.2781456953642385,
1246
+ "grad_norm": 1.1933870315551758,
1247
+ "learning_rate": 3.0412371134020617e-05,
1248
+ "loss": 0.0756,
1249
+ "step": 344
1250
+ },
1251
+ {
1252
+ "epoch": 2.2913907284768213,
1253
+ "grad_norm": 9.06655216217041,
1254
+ "learning_rate": 3.0265095729013254e-05,
1255
+ "loss": 0.7712,
1256
+ "step": 346
1257
+ },
1258
+ {
1259
+ "epoch": 2.304635761589404,
1260
+ "grad_norm": 8.230104446411133,
1261
+ "learning_rate": 3.011782032400589e-05,
1262
+ "loss": 0.9383,
1263
+ "step": 348
1264
+ },
1265
+ {
1266
+ "epoch": 2.3178807947019866,
1267
+ "grad_norm": 10.278660774230957,
1268
+ "learning_rate": 2.9970544918998527e-05,
1269
+ "loss": 0.6566,
1270
+ "step": 350
1271
+ },
1272
+ {
1273
+ "epoch": 2.3311258278145695,
1274
+ "grad_norm": 1.700579047203064,
1275
+ "learning_rate": 2.9823269513991164e-05,
1276
+ "loss": 0.1095,
1277
+ "step": 352
1278
+ },
1279
+ {
1280
+ "epoch": 2.3443708609271523,
1281
+ "grad_norm": 4.130746841430664,
1282
+ "learning_rate": 2.96759941089838e-05,
1283
+ "loss": 0.7986,
1284
+ "step": 354
1285
+ },
1286
+ {
1287
+ "epoch": 2.357615894039735,
1288
+ "grad_norm": 9.603137016296387,
1289
+ "learning_rate": 2.9528718703976438e-05,
1290
+ "loss": 0.6449,
1291
+ "step": 356
1292
+ },
1293
+ {
1294
+ "epoch": 2.370860927152318,
1295
+ "grad_norm": 13.046420097351074,
1296
+ "learning_rate": 2.9381443298969075e-05,
1297
+ "loss": 0.7776,
1298
+ "step": 358
1299
+ },
1300
+ {
1301
+ "epoch": 2.384105960264901,
1302
+ "grad_norm": 4.095331192016602,
1303
+ "learning_rate": 2.9234167893961712e-05,
1304
+ "loss": 0.2639,
1305
+ "step": 360
1306
+ },
1307
+ {
1308
+ "epoch": 2.3973509933774833,
1309
+ "grad_norm": 0.3949756324291229,
1310
+ "learning_rate": 2.908689248895435e-05,
1311
+ "loss": 0.1134,
1312
+ "step": 362
1313
+ },
1314
+ {
1315
+ "epoch": 2.410596026490066,
1316
+ "grad_norm": 3.804518699645996,
1317
+ "learning_rate": 2.8939617083946985e-05,
1318
+ "loss": 0.3421,
1319
+ "step": 364
1320
+ },
1321
+ {
1322
+ "epoch": 2.423841059602649,
1323
+ "grad_norm": 12.5763521194458,
1324
+ "learning_rate": 2.8792341678939616e-05,
1325
+ "loss": 0.7242,
1326
+ "step": 366
1327
+ },
1328
+ {
1329
+ "epoch": 2.437086092715232,
1330
+ "grad_norm": 5.6927032470703125,
1331
+ "learning_rate": 2.8645066273932252e-05,
1332
+ "loss": 0.9591,
1333
+ "step": 368
1334
+ },
1335
+ {
1336
+ "epoch": 2.4503311258278146,
1337
+ "grad_norm": 13.04416275024414,
1338
+ "learning_rate": 2.849779086892489e-05,
1339
+ "loss": 0.4636,
1340
+ "step": 370
1341
+ },
1342
+ {
1343
+ "epoch": 2.4635761589403975,
1344
+ "grad_norm": 4.799520015716553,
1345
+ "learning_rate": 2.8350515463917526e-05,
1346
+ "loss": 0.4037,
1347
+ "step": 372
1348
+ },
1349
+ {
1350
+ "epoch": 2.47682119205298,
1351
+ "grad_norm": 8.108109474182129,
1352
+ "learning_rate": 2.8203240058910163e-05,
1353
+ "loss": 1.2403,
1354
+ "step": 374
1355
+ },
1356
+ {
1357
+ "epoch": 2.4900662251655628,
1358
+ "grad_norm": 3.977107286453247,
1359
+ "learning_rate": 2.80559646539028e-05,
1360
+ "loss": 0.1793,
1361
+ "step": 376
1362
+ },
1363
+ {
1364
+ "epoch": 2.5033112582781456,
1365
+ "grad_norm": 8.900064468383789,
1366
+ "learning_rate": 2.7908689248895437e-05,
1367
+ "loss": 0.6712,
1368
+ "step": 378
1369
+ },
1370
+ {
1371
+ "epoch": 2.5165562913907285,
1372
+ "grad_norm": 5.152413368225098,
1373
+ "learning_rate": 2.7761413843888074e-05,
1374
+ "loss": 0.0786,
1375
+ "step": 380
1376
+ },
1377
+ {
1378
+ "epoch": 2.5298013245033113,
1379
+ "grad_norm": 1.7878741025924683,
1380
+ "learning_rate": 2.761413843888071e-05,
1381
+ "loss": 0.3663,
1382
+ "step": 382
1383
+ },
1384
+ {
1385
+ "epoch": 2.543046357615894,
1386
+ "grad_norm": 4.110722064971924,
1387
+ "learning_rate": 2.7466863033873347e-05,
1388
+ "loss": 0.3523,
1389
+ "step": 384
1390
+ },
1391
+ {
1392
+ "epoch": 2.556291390728477,
1393
+ "grad_norm": 15.2096586227417,
1394
+ "learning_rate": 2.7319587628865977e-05,
1395
+ "loss": 0.233,
1396
+ "step": 386
1397
+ },
1398
+ {
1399
+ "epoch": 2.5695364238410594,
1400
+ "grad_norm": 4.752151012420654,
1401
+ "learning_rate": 2.7172312223858614e-05,
1402
+ "loss": 0.5043,
1403
+ "step": 388
1404
+ },
1405
+ {
1406
+ "epoch": 2.5827814569536423,
1407
+ "grad_norm": 9.852655410766602,
1408
+ "learning_rate": 2.702503681885125e-05,
1409
+ "loss": 0.1858,
1410
+ "step": 390
1411
+ },
1412
+ {
1413
+ "epoch": 2.596026490066225,
1414
+ "grad_norm": 13.454380989074707,
1415
+ "learning_rate": 2.6877761413843888e-05,
1416
+ "loss": 0.597,
1417
+ "step": 392
1418
+ },
1419
+ {
1420
+ "epoch": 2.609271523178808,
1421
+ "grad_norm": 7.8776326179504395,
1422
+ "learning_rate": 2.6730486008836525e-05,
1423
+ "loss": 0.7841,
1424
+ "step": 394
1425
+ },
1426
+ {
1427
+ "epoch": 2.622516556291391,
1428
+ "grad_norm": 9.048988342285156,
1429
+ "learning_rate": 2.6583210603829162e-05,
1430
+ "loss": 0.8305,
1431
+ "step": 396
1432
+ },
1433
+ {
1434
+ "epoch": 2.6357615894039736,
1435
+ "grad_norm": 13.187280654907227,
1436
+ "learning_rate": 2.64359351988218e-05,
1437
+ "loss": 1.0179,
1438
+ "step": 398
1439
+ },
1440
+ {
1441
+ "epoch": 2.6490066225165565,
1442
+ "grad_norm": 13.392657279968262,
1443
+ "learning_rate": 2.6288659793814435e-05,
1444
+ "loss": 0.1317,
1445
+ "step": 400
1446
+ },
1447
+ {
1448
+ "epoch": 2.662251655629139,
1449
+ "grad_norm": 11.062936782836914,
1450
+ "learning_rate": 2.6141384388807072e-05,
1451
+ "loss": 0.6804,
1452
+ "step": 402
1453
+ },
1454
+ {
1455
+ "epoch": 2.6754966887417218,
1456
+ "grad_norm": 11.463652610778809,
1457
+ "learning_rate": 2.599410898379971e-05,
1458
+ "loss": 0.5558,
1459
+ "step": 404
1460
+ },
1461
+ {
1462
+ "epoch": 2.6887417218543046,
1463
+ "grad_norm": 5.696853160858154,
1464
+ "learning_rate": 2.5846833578792346e-05,
1465
+ "loss": 0.7812,
1466
+ "step": 406
1467
+ },
1468
+ {
1469
+ "epoch": 2.7019867549668874,
1470
+ "grad_norm": 15.077645301818848,
1471
+ "learning_rate": 2.5699558173784976e-05,
1472
+ "loss": 0.6744,
1473
+ "step": 408
1474
+ },
1475
+ {
1476
+ "epoch": 2.7152317880794703,
1477
+ "grad_norm": 4.043674468994141,
1478
+ "learning_rate": 2.5552282768777613e-05,
1479
+ "loss": 0.3354,
1480
+ "step": 410
1481
+ },
1482
+ {
1483
+ "epoch": 2.7284768211920527,
1484
+ "grad_norm": 1.150486946105957,
1485
+ "learning_rate": 2.540500736377025e-05,
1486
+ "loss": 0.3314,
1487
+ "step": 412
1488
+ },
1489
+ {
1490
+ "epoch": 2.741721854304636,
1491
+ "grad_norm": 12.133696556091309,
1492
+ "learning_rate": 2.5257731958762887e-05,
1493
+ "loss": 0.526,
1494
+ "step": 414
1495
+ },
1496
+ {
1497
+ "epoch": 2.7549668874172184,
1498
+ "grad_norm": 11.533574104309082,
1499
+ "learning_rate": 2.5110456553755524e-05,
1500
+ "loss": 0.8745,
1501
+ "step": 416
1502
+ },
1503
+ {
1504
+ "epoch": 2.7682119205298013,
1505
+ "grad_norm": 1.3846099376678467,
1506
+ "learning_rate": 2.496318114874816e-05,
1507
+ "loss": 0.204,
1508
+ "step": 418
1509
+ },
1510
+ {
1511
+ "epoch": 2.781456953642384,
1512
+ "grad_norm": 4.653772830963135,
1513
+ "learning_rate": 2.4815905743740797e-05,
1514
+ "loss": 0.3621,
1515
+ "step": 420
1516
+ },
1517
+ {
1518
+ "epoch": 2.794701986754967,
1519
+ "grad_norm": 8.693764686584473,
1520
+ "learning_rate": 2.4668630338733434e-05,
1521
+ "loss": 0.4153,
1522
+ "step": 422
1523
+ },
1524
+ {
1525
+ "epoch": 2.80794701986755,
1526
+ "grad_norm": 11.415797233581543,
1527
+ "learning_rate": 2.4521354933726068e-05,
1528
+ "loss": 0.6997,
1529
+ "step": 424
1530
+ },
1531
+ {
1532
+ "epoch": 2.821192052980132,
1533
+ "grad_norm": 3.682908773422241,
1534
+ "learning_rate": 2.4374079528718704e-05,
1535
+ "loss": 0.3039,
1536
+ "step": 426
1537
+ },
1538
+ {
1539
+ "epoch": 2.8344370860927155,
1540
+ "grad_norm": 9.39224624633789,
1541
+ "learning_rate": 2.422680412371134e-05,
1542
+ "loss": 0.6302,
1543
+ "step": 428
1544
+ },
1545
+ {
1546
+ "epoch": 2.847682119205298,
1547
+ "grad_norm": 3.271193742752075,
1548
+ "learning_rate": 2.4079528718703978e-05,
1549
+ "loss": 0.2706,
1550
+ "step": 430
1551
+ },
1552
+ {
1553
+ "epoch": 2.8609271523178808,
1554
+ "grad_norm": 7.453132629394531,
1555
+ "learning_rate": 2.3932253313696615e-05,
1556
+ "loss": 0.4594,
1557
+ "step": 432
1558
+ },
1559
+ {
1560
+ "epoch": 2.8741721854304636,
1561
+ "grad_norm": 16.519481658935547,
1562
+ "learning_rate": 2.378497790868925e-05,
1563
+ "loss": 0.2437,
1564
+ "step": 434
1565
+ },
1566
+ {
1567
+ "epoch": 2.8874172185430464,
1568
+ "grad_norm": 17.087709426879883,
1569
+ "learning_rate": 2.3637702503681885e-05,
1570
+ "loss": 0.9733,
1571
+ "step": 436
1572
+ },
1573
+ {
1574
+ "epoch": 2.9006622516556293,
1575
+ "grad_norm": 1.0074738264083862,
1576
+ "learning_rate": 2.3490427098674522e-05,
1577
+ "loss": 0.649,
1578
+ "step": 438
1579
+ },
1580
+ {
1581
+ "epoch": 2.9139072847682117,
1582
+ "grad_norm": 4.366696834564209,
1583
+ "learning_rate": 2.334315169366716e-05,
1584
+ "loss": 0.5205,
1585
+ "step": 440
1586
+ },
1587
+ {
1588
+ "epoch": 2.9271523178807946,
1589
+ "grad_norm": 20.609182357788086,
1590
+ "learning_rate": 2.3195876288659796e-05,
1591
+ "loss": 0.2493,
1592
+ "step": 442
1593
+ },
1594
+ {
1595
+ "epoch": 2.9403973509933774,
1596
+ "grad_norm": 9.770134925842285,
1597
+ "learning_rate": 2.3048600883652433e-05,
1598
+ "loss": 0.1564,
1599
+ "step": 444
1600
+ },
1601
+ {
1602
+ "epoch": 2.9536423841059603,
1603
+ "grad_norm": 8.714824676513672,
1604
+ "learning_rate": 2.2901325478645066e-05,
1605
+ "loss": 0.6449,
1606
+ "step": 446
1607
+ },
1608
+ {
1609
+ "epoch": 2.966887417218543,
1610
+ "grad_norm": 7.8644514083862305,
1611
+ "learning_rate": 2.2754050073637703e-05,
1612
+ "loss": 1.3438,
1613
+ "step": 448
1614
+ },
1615
+ {
1616
+ "epoch": 2.980132450331126,
1617
+ "grad_norm": 5.881997108459473,
1618
+ "learning_rate": 2.260677466863034e-05,
1619
+ "loss": 0.3383,
1620
+ "step": 450
1621
+ },
1622
+ {
1623
+ "epoch": 2.993377483443709,
1624
+ "grad_norm": 12.096423149108887,
1625
+ "learning_rate": 2.2459499263622977e-05,
1626
+ "loss": 0.7644,
1627
+ "step": 452
1628
+ },
1629
+ {
1630
+ "epoch": 3.0,
1631
+ "eval_accuracy": 0.74,
1632
+ "eval_f1_macro": 0.4558641367469575,
1633
+ "eval_f1_micro": 0.74,
1634
+ "eval_f1_weighted": 0.7165269403625714,
1635
+ "eval_loss": 1.0647395849227905,
1636
+ "eval_precision_macro": 0.47941017316017315,
1637
+ "eval_precision_micro": 0.74,
1638
+ "eval_precision_weighted": 0.7261709956709957,
1639
+ "eval_recall_macro": 0.4931601731601732,
1640
+ "eval_recall_micro": 0.74,
1641
+ "eval_recall_weighted": 0.74,
1642
+ "eval_runtime": 0.1879,
1643
+ "eval_samples_per_second": 532.193,
1644
+ "eval_steps_per_second": 37.254,
1645
+ "step": 453
1646
+ }
1647
+ ],
1648
+ "logging_steps": 2,
1649
+ "max_steps": 755,
1650
+ "num_input_tokens_seen": 0,
1651
+ "num_train_epochs": 5,
1652
+ "save_steps": 500,
1653
+ "total_flos": 119358311592960.0,
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 sha256:63c1572573d5153d2a501e4fe87e44065a9b3b0887edf39cacfe426b94fb3e3c
3
+ size 5048
config.json ADDED
@@ -0,0 +1,71 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "medicalai/ClinicalBERT",
3
+ "_num_labels": 20,
4
+ "activation": "gelu",
5
+ "architectures": [
6
+ "DistilBertForSequenceClassification"
7
+ ],
8
+ "attention_dropout": 0.1,
9
+ "dim": 768,
10
+ "dropout": 0.1,
11
+ "hidden_dim": 3072,
12
+ "id2label": {
13
+ "0": "ALLERGY",
14
+ "1": "ASSESSMENT",
15
+ "2": "CC",
16
+ "3": "DIAGNOSIS",
17
+ "4": "DISPOSITION",
18
+ "5": "EDCOURSE",
19
+ "6": "EXAM",
20
+ "7": "FAM/SOCHX",
21
+ "8": "GENHX",
22
+ "9": "GYNHX",
23
+ "10": "IMAGING",
24
+ "11": "IMMUNIZATIONS",
25
+ "12": "LABS",
26
+ "13": "MEDICATIONS",
27
+ "14": "OTHER_HISTORY",
28
+ "15": "PASTMEDICALHX",
29
+ "16": "PASTSURGICAL",
30
+ "17": "PLAN",
31
+ "18": "PROCEDURES",
32
+ "19": "ROS"
33
+ },
34
+ "initializer_range": 0.02,
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
+ "max_position_embeddings": 512,
58
+ "model_type": "distilbert",
59
+ "n_heads": 12,
60
+ "n_layers": 6,
61
+ "output_past": true,
62
+ "pad_token_id": 0,
63
+ "problem_type": "single_label_classification",
64
+ "qa_dropout": 0.1,
65
+ "seq_classif_dropout": 0.2,
66
+ "sinusoidal_pos_embds": false,
67
+ "tie_weights_": true,
68
+ "torch_dtype": "float32",
69
+ "transformers_version": "4.40.1",
70
+ "vocab_size": 119547
71
+ }
model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:fad10e6d7c07618a95f6302513c82383b7479673f4b1b5770ab408d882e4de0a
3
+ size 541372744
runs/Apr28_03-26-11_r-dhananjay2912-deberta-section-classifier-d65r8w9r-def20-gpx5z/events.out.tfevents.1714274772.r-dhananjay2912-deberta-section-classifier-d65r8w9r-def20-gpx5z.63.0 CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:ba1cf3f37dd6fdbcebdd7251909ae05b04da14c8e7434d95c8f85d4a4b9a5c81
3
- size 5517
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:6fa6eb777eb5e45ec07d7c8e2a4dbb4b4311783f47c8a6e93d7c60ecc4cfb41f
3
+ size 89331
runs/Apr28_03-26-11_r-dhananjay2912-deberta-section-classifier-d65r8w9r-def20-gpx5z/events.out.tfevents.1714274854.r-dhananjay2912-deberta-section-classifier-d65r8w9r-def20-gpx5z.63.1 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d4985fa14ae6db72968a95d0e16b3ce5f9d9d24259179224ea1aedc80cea6e6a
3
+ size 921
special_tokens_map.json ADDED
@@ -0,0 +1,7 @@
 
 
 
 
 
 
 
 
1
+ {
2
+ "cls_token": "[CLS]",
3
+ "mask_token": "[MASK]",
4
+ "pad_token": "[PAD]",
5
+ "sep_token": "[SEP]",
6
+ "unk_token": "[UNK]"
7
+ }
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
+ "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": true,
48
+ "full_tokenizer_file": null,
49
+ "mask_token": "[MASK]",
50
+ "model_max_length": 1000000000000000019884624838656,
51
+ "never_split": null,
52
+ "pad_token": "[PAD]",
53
+ "sep_token": "[SEP]",
54
+ "strip_accents": null,
55
+ "tokenize_chinese_chars": true,
56
+ "tokenizer_class": "DistilBertTokenizer",
57
+ "unk_token": "[UNK]"
58
+ }
training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:63c1572573d5153d2a501e4fe87e44065a9b3b0887edf39cacfe426b94fb3e3c
3
+ size 5048
training_params.json ADDED
@@ -0,0 +1,28 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "data_path": "autotrain-qs45q-p9syi/autotrain-data",
3
+ "model": "medicalai/ClinicalBERT",
4
+ "lr": 5e-05,
5
+ "epochs": 5,
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-qs45q-p9syi",
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
+ }
vocab.txt ADDED
The diff for this file is too large to render. See raw diff