Hugging Face's logo

datht
/
vi-modernbert-ViHSD-ep20

Text Classification
Transformers
Safetensors
modernbert
Generated from Trainer
Model card Files Community
datht commited on
Commit
17b3f99
·
verified ·
1 Parent(s): 6cc8566

Training in progress, epoch 0

Browse files
Files changed (12) hide show
  1. README.md +85 -0
  2. all_results.json +27 -0
  3. config.json +57 -0
  4. eval_results.json +13 -0
  5. model.safetensors +3 -0
  6. special_tokens_map.json +62 -0
  7. test_results.json +12 -0
  8. tokenizer.json +0 -0
  9. tokenizer_config.json +286 -0
  10. train_results.json +7 -0
  11. trainer_state.json +2535 -0
  12. training_args.bin +3 -0
README.md ADDED
@@ -0,0 +1,85 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ library_name: transformers
3
+ tags:
4
+ - generated_from_trainer
5
+ model-index:
6
+ - name: vi-modernbert-VLSP2016_SA-ep20
7
+ results: []
8
+ ---
9
+
10
+ <!-- This model card has been generated automatically according to the information the Trainer had access to. You
11
+ should probably proofread and complete it, then remove this comment. -->
12
+
13
+ # vi-modernbert-VLSP2016_SA-ep20
14
+
15
+ This model was trained from scratch on the None dataset.
16
+ It achieves the following results on the evaluation set:
17
+ - Loss: 2.9834
18
+ - Micro F1: 74.4762
19
+ - Micro Precision: 74.4762
20
+ - Micro Recall: 74.4762
21
+ - Macro F1: 74.4952
22
+ - Macro Precision: 74.5167
23
+ - Macro Recall: 74.4762
24
+
25
+ ## Model description
26
+
27
+ More information needed
28
+
29
+ ## Intended uses & limitations
30
+
31
+ More information needed
32
+
33
+ ## Training and evaluation data
34
+
35
+ More information needed
36
+
37
+ ## Training procedure
38
+
39
+ ### Training hyperparameters
40
+
41
+ The following hyperparameters were used during training:
42
+ - learning_rate: 0.0002
43
+ - train_batch_size: 16
44
+ - eval_batch_size: 16
45
+ - seed: 42
46
+ - distributed_type: multi-GPU
47
+ - gradient_accumulation_steps: 4
48
+ - total_train_batch_size: 64
49
+ - optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
50
+ - lr_scheduler_type: cosine
51
+ - lr_scheduler_warmup_ratio: 0.01
52
+ - num_epochs: 20.0
53
+
54
+ ### Training results
55
+
56
+ | Training Loss | Epoch | Step | Validation Loss | Micro F1 | Micro Precision | Micro Recall | Macro F1 | Macro Precision | Macro Recall |
57
+ |:-------------:|:-------:|:----:|:---------------:|:--------:|:---------------:|:------------:|:--------:|:---------------:|:------------:|
58
+ | 2.2395 | 1.0 | 80 | 0.6337 | 73.1429 | 73.1429 | 73.1429 | 73.2675 | 75.2452 | 73.1429 |
59
+ | 1.0835 | 2.0 | 160 | 0.7826 | 73.9048 | 73.9048 | 73.9048 | 73.5690 | 73.7349 | 73.9048 |
60
+ | 0.2647 | 3.0 | 240 | 1.3780 | 73.2381 | 73.2381 | 73.2381 | 73.1154 | 73.2290 | 73.2381 |
61
+ | 0.3228 | 4.0 | 320 | 1.7462 | 72.7619 | 72.7619 | 72.7619 | 72.7257 | 72.7147 | 72.7619 |
62
+ | 0.0747 | 5.0 | 400 | 1.7594 | 74.0 | 74.0 | 74.0 | 74.1705 | 75.0419 | 74.0000 |
63
+ | 0.0868 | 6.0 | 480 | 1.5809 | 73.6190 | 73.6190 | 73.6190 | 73.2438 | 73.6195 | 73.6190 |
64
+ | 0.0801 | 7.0 | 560 | 1.7074 | 74.3810 | 74.3810 | 74.3810 | 74.3996 | 74.4326 | 74.3810 |
65
+ | 0.1489 | 8.0 | 640 | 3.0942 | 73.5238 | 73.5238 | 73.5238 | 73.7559 | 74.9712 | 73.5238 |
66
+ | 0.1558 | 9.0 | 720 | 3.2612 | 72.3810 | 72.3810 | 72.3810 | 72.5636 | 74.4763 | 72.3810 |
67
+ | 0.0 | 10.0 | 800 | 3.0012 | 74.8571 | 74.8571 | 74.8571 | 74.9893 | 75.3648 | 74.8571 |
68
+ | 0.0 | 11.0 | 880 | 2.9445 | 74.0 | 74.0 | 74.0 | 73.9497 | 73.9142 | 74.0000 |
69
+ | 0.0 | 12.0 | 960 | 2.9618 | 74.4762 | 74.4762 | 74.4762 | 74.4845 | 74.4934 | 74.4762 |
70
+ | 0.0 | 13.0 | 1040 | 2.9744 | 74.4762 | 74.4762 | 74.4762 | 74.4952 | 74.5167 | 74.4762 |
71
+ | 0.0 | 14.0 | 1120 | 2.9787 | 74.3810 | 74.3810 | 74.3810 | 74.3974 | 74.4156 | 74.3810 |
72
+ | 0.0 | 15.0 | 1200 | 2.9775 | 74.5714 | 74.5714 | 74.5714 | 74.5929 | 74.6181 | 74.5714 |
73
+ | 0.0 | 16.0 | 1280 | 2.9835 | 74.2857 | 74.2857 | 74.2857 | 74.3051 | 74.3270 | 74.2857 |
74
+ | 0.0 | 17.0 | 1360 | 2.9836 | 74.4762 | 74.4762 | 74.4762 | 74.4869 | 74.4985 | 74.4762 |
75
+ | 0.0 | 18.0 | 1440 | 2.9821 | 74.4762 | 74.4762 | 74.4762 | 74.4944 | 74.5153 | 74.4762 |
76
+ | 0.0 | 19.0 | 1520 | 2.9836 | 74.5714 | 74.5714 | 74.5714 | 74.5929 | 74.6181 | 74.5714 |
77
+ | 0.0 | 19.7524 | 1580 | 2.9834 | 74.4762 | 74.4762 | 74.4762 | 74.4952 | 74.5167 | 74.4762 |
78
+
79
+
80
+ ### Framework versions
81
+
82
+ - Transformers 4.50.0
83
+ - Pytorch 2.6.0+cu124
84
+ - Datasets 2.15.0
85
+ - Tokenizers 0.21.1
all_results.json ADDED
@@ -0,0 +1,27 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 19.752351097178682,
3
+ "eval_loss": 2.983431816101074,
4
+ "eval_macro_f1": 74.495166799055,
5
+ "eval_macro_precision": 74.51671347356447,
6
+ "eval_macro_recall": 74.47619047619047,
7
+ "eval_micro_f1": 74.47619047619047,
8
+ "eval_micro_precision": 74.47619047619047,
9
+ "eval_micro_recall": 74.47619047619047,
10
+ "eval_runtime": 3.4211,
11
+ "eval_samples_per_second": 306.923,
12
+ "eval_steps_per_second": 19.292,
13
+ "test_loss": 2.983431816101074,
14
+ "test_macro_f1": 74.495166799055,
15
+ "test_macro_precision": 74.51671347356447,
16
+ "test_macro_recall": 74.47619047619047,
17
+ "test_micro_f1": 74.47619047619047,
18
+ "test_micro_precision": 74.47619047619047,
19
+ "test_micro_recall": 74.47619047619047,
20
+ "test_runtime": 4.1608,
21
+ "test_samples_per_second": 252.353,
22
+ "test_steps_per_second": 15.862,
23
+ "train_loss": 0.2892922113949427,
24
+ "train_runtime": 830.7771,
25
+ "train_samples_per_second": 122.777,
26
+ "train_steps_per_second": 1.902
27
+ }
config.json ADDED
@@ -0,0 +1,57 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "architectures": [
3
+ "ModernBertForSequenceClassification"
4
+ ],
5
+ "attention_bias": false,
6
+ "attention_dropout": 0.0,
7
+ "bos_token_id": 3,
8
+ "classifier_activation": "gelu",
9
+ "classifier_bias": false,
10
+ "classifier_dropout": 0.0,
11
+ "classifier_pooling": "mean",
12
+ "cls_token_id": 3,
13
+ "decoder_bias": true,
14
+ "deterministic_flash_attn": false,
15
+ "embedding_dropout": 0.0,
16
+ "eos_token_id": 4,
17
+ "global_attn_every_n_layers": 3,
18
+ "global_rope_theta": 160000.0,
19
+ "gradient_checkpointing": false,
20
+ "hidden_activation": "gelu",
21
+ "hidden_size": 768,
22
+ "id2label": {
23
+ "0": "CLEAN",
24
+ "1": "HATE",
25
+ "2": "OFFENSIVE"
26
+ },
27
+ "initializer_cutoff_factor": 2.0,
28
+ "initializer_range": 0.02,
29
+ "intermediate_size": 1152,
30
+ "label2id": {
31
+ "CLEAN": 0,
32
+ "HATE": 1,
33
+ "OFFENSIVE": 2
34
+ },
35
+ "layer_norm_eps": 1e-05,
36
+ "local_attention": 128,
37
+ "local_rope_theta": 10000.0,
38
+ "max_position_embeddings": 8192,
39
+ "mlp_bias": false,
40
+ "mlp_dropout": 0.0,
41
+ "model_type": "modernbert",
42
+ "norm_bias": false,
43
+ "norm_eps": 1e-05,
44
+ "num_attention_heads": 12,
45
+ "num_hidden_layers": 22,
46
+ "pad_token_id": 5,
47
+ "position_embedding_type": "absolute",
48
+ "problem_type": "single_label_classification",
49
+ "reference_compile": true,
50
+ "repad_logits_with_grad": false,
51
+ "sep_token_id": 4,
52
+ "sparse_pred_ignore_index": -100,
53
+ "sparse_prediction": false,
54
+ "torch_dtype": "bfloat16",
55
+ "transformers_version": "4.50.0",
56
+ "vocab_size": 50368
57
+ }
eval_results.json ADDED
@@ -0,0 +1,13 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 19.752351097178682,
3
+ "eval_loss": 2.983431816101074,
4
+ "eval_macro_f1": 74.495166799055,
5
+ "eval_macro_precision": 74.51671347356447,
6
+ "eval_macro_recall": 74.47619047619047,
7
+ "eval_micro_f1": 74.47619047619047,
8
+ "eval_micro_precision": 74.47619047619047,
9
+ "eval_micro_recall": 74.47619047619047,
10
+ "eval_runtime": 3.4211,
11
+ "eval_samples_per_second": 306.923,
12
+ "eval_steps_per_second": 19.292
13
+ }
model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:41892dde0c38d11fc3ad43796e3af4a153a3ca2353abccf77fe45b8c5394189d
3
+ size 299228630
special_tokens_map.json ADDED
@@ -0,0 +1,62 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "additional_special_tokens": [
3
+ "<|UNUSED_0|>",
4
+ "<|UNUSED_1|>",
5
+ "<|UNUSED_2|>",
6
+ "<|UNUSED_3|>",
7
+ "<|UNUSED_4|>",
8
+ "<|UNUSED_5|>",
9
+ "<|UNUSED_6|>",
10
+ "<|UNUSED_7|>",
11
+ "<|UNUSED_8|>",
12
+ "<|UNUSED_9|>",
13
+ "<|UNUSED_10|>",
14
+ "<|UNUSED_11|>",
15
+ "<|UNUSED_12|>",
16
+ "<|UNUSED_13|>",
17
+ "<|UNUSED_14|>",
18
+ "<|UNUSED_15|>",
19
+ "<|HISTORY|>",
20
+ "<|PHONE_NUMBER|>",
21
+ "<|EMAIL|>",
22
+ "<|IP_ADDRESS|>",
23
+ "<|URL|>",
24
+ "<|DATE|>",
25
+ "<|TIME|>"
26
+ ],
27
+ "cls_token": {
28
+ "content": "[CLS]",
29
+ "lstrip": false,
30
+ "normalized": false,
31
+ "rstrip": false,
32
+ "single_word": false
33
+ },
34
+ "mask_token": {
35
+ "content": "[MASK]",
36
+ "lstrip": true,
37
+ "normalized": false,
38
+ "rstrip": false,
39
+ "single_word": false
40
+ },
41
+ "pad_token": {
42
+ "content": "[PAD]",
43
+ "lstrip": false,
44
+ "normalized": false,
45
+ "rstrip": false,
46
+ "single_word": false
47
+ },
48
+ "sep_token": {
49
+ "content": "[SEP]",
50
+ "lstrip": false,
51
+ "normalized": false,
52
+ "rstrip": false,
53
+ "single_word": false
54
+ },
55
+ "unk_token": {
56
+ "content": "[UNK]",
57
+ "lstrip": false,
58
+ "normalized": false,
59
+ "rstrip": false,
60
+ "single_word": false
61
+ }
62
+ }
test_results.json ADDED
@@ -0,0 +1,12 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "test_loss": 2.983431816101074,
3
+ "test_macro_f1": 74.495166799055,
4
+ "test_macro_precision": 74.51671347356447,
5
+ "test_macro_recall": 74.47619047619047,
6
+ "test_micro_f1": 74.47619047619047,
7
+ "test_micro_precision": 74.47619047619047,
8
+ "test_micro_recall": 74.47619047619047,
9
+ "test_runtime": 4.1608,
10
+ "test_samples_per_second": 252.353,
11
+ "test_steps_per_second": 15.862
12
+ }
tokenizer.json ADDED
The diff for this file is too large to render. See raw diff
 
tokenizer_config.json ADDED
@@ -0,0 +1,286 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "added_tokens_decoder": {
3
+ "0": {
4
+ "content": "<|padding|>",
5
+ "lstrip": false,
6
+ "normalized": false,
7
+ "rstrip": false,
8
+ "single_word": false,
9
+ "special": true
10
+ },
11
+ "1": {
12
+ "content": "<|endoftext|>",
13
+ "lstrip": false,
14
+ "normalized": false,
15
+ "rstrip": false,
16
+ "single_word": false,
17
+ "special": true
18
+ },
19
+ "2": {
20
+ "content": "[UNK]",
21
+ "lstrip": false,
22
+ "normalized": false,
23
+ "rstrip": false,
24
+ "single_word": false,
25
+ "special": true
26
+ },
27
+ "3": {
28
+ "content": "[CLS]",
29
+ "lstrip": false,
30
+ "normalized": false,
31
+ "rstrip": false,
32
+ "single_word": false,
33
+ "special": true
34
+ },
35
+ "4": {
36
+ "content": "[SEP]",
37
+ "lstrip": false,
38
+ "normalized": false,
39
+ "rstrip": false,
40
+ "single_word": false,
41
+ "special": true
42
+ },
43
+ "5": {
44
+ "content": "[PAD]",
45
+ "lstrip": false,
46
+ "normalized": false,
47
+ "rstrip": false,
48
+ "single_word": false,
49
+ "special": true
50
+ },
51
+ "6": {
52
+ "content": "[MASK]",
53
+ "lstrip": true,
54
+ "normalized": false,
55
+ "rstrip": false,
56
+ "single_word": false,
57
+ "special": true
58
+ },
59
+ "7": {
60
+ "content": "<|UNUSED_0|>",
61
+ "lstrip": false,
62
+ "normalized": false,
63
+ "rstrip": false,
64
+ "single_word": false,
65
+ "special": true
66
+ },
67
+ "8": {
68
+ "content": "<|UNUSED_1|>",
69
+ "lstrip": false,
70
+ "normalized": false,
71
+ "rstrip": false,
72
+ "single_word": false,
73
+ "special": true
74
+ },
75
+ "9": {
76
+ "content": "<|UNUSED_2|>",
77
+ "lstrip": false,
78
+ "normalized": false,
79
+ "rstrip": false,
80
+ "single_word": false,
81
+ "special": true
82
+ },
83
+ "10": {
84
+ "content": "<|UNUSED_3|>",
85
+ "lstrip": false,
86
+ "normalized": false,
87
+ "rstrip": false,
88
+ "single_word": false,
89
+ "special": true
90
+ },
91
+ "11": {
92
+ "content": "<|UNUSED_4|>",
93
+ "lstrip": false,
94
+ "normalized": false,
95
+ "rstrip": false,
96
+ "single_word": false,
97
+ "special": true
98
+ },
99
+ "12": {
100
+ "content": "<|UNUSED_5|>",
101
+ "lstrip": false,
102
+ "normalized": false,
103
+ "rstrip": false,
104
+ "single_word": false,
105
+ "special": true
106
+ },
107
+ "13": {
108
+ "content": "<|UNUSED_6|>",
109
+ "lstrip": false,
110
+ "normalized": false,
111
+ "rstrip": false,
112
+ "single_word": false,
113
+ "special": true
114
+ },
115
+ "14": {
116
+ "content": "<|UNUSED_7|>",
117
+ "lstrip": false,
118
+ "normalized": false,
119
+ "rstrip": false,
120
+ "single_word": false,
121
+ "special": true
122
+ },
123
+ "15": {
124
+ "content": "<|UNUSED_8|>",
125
+ "lstrip": false,
126
+ "normalized": false,
127
+ "rstrip": false,
128
+ "single_word": false,
129
+ "special": true
130
+ },
131
+ "16": {
132
+ "content": "<|UNUSED_9|>",
133
+ "lstrip": false,
134
+ "normalized": false,
135
+ "rstrip": false,
136
+ "single_word": false,
137
+ "special": true
138
+ },
139
+ "17": {
140
+ "content": "<|UNUSED_10|>",
141
+ "lstrip": false,
142
+ "normalized": false,
143
+ "rstrip": false,
144
+ "single_word": false,
145
+ "special": true
146
+ },
147
+ "18": {
148
+ "content": "<|UNUSED_11|>",
149
+ "lstrip": false,
150
+ "normalized": false,
151
+ "rstrip": false,
152
+ "single_word": false,
153
+ "special": true
154
+ },
155
+ "19": {
156
+ "content": "<|UNUSED_12|>",
157
+ "lstrip": false,
158
+ "normalized": false,
159
+ "rstrip": false,
160
+ "single_word": false,
161
+ "special": true
162
+ },
163
+ "20": {
164
+ "content": "<|UNUSED_13|>",
165
+ "lstrip": false,
166
+ "normalized": false,
167
+ "rstrip": false,
168
+ "single_word": false,
169
+ "special": true
170
+ },
171
+ "21": {
172
+ "content": "<|UNUSED_14|>",
173
+ "lstrip": false,
174
+ "normalized": false,
175
+ "rstrip": false,
176
+ "single_word": false,
177
+ "special": true
178
+ },
179
+ "22": {
180
+ "content": "<|UNUSED_15|>",
181
+ "lstrip": false,
182
+ "normalized": false,
183
+ "rstrip": false,
184
+ "single_word": false,
185
+ "special": true
186
+ },
187
+ "23": {
188
+ "content": "<|HISTORY|>",
189
+ "lstrip": false,
190
+ "normalized": false,
191
+ "rstrip": false,
192
+ "single_word": false,
193
+ "special": true
194
+ },
195
+ "24": {
196
+ "content": "<|PHONE_NUMBER|>",
197
+ "lstrip": false,
198
+ "normalized": false,
199
+ "rstrip": false,
200
+ "single_word": false,
201
+ "special": true
202
+ },
203
+ "25": {
204
+ "content": "<|EMAIL|>",
205
+ "lstrip": false,
206
+ "normalized": false,
207
+ "rstrip": false,
208
+ "single_word": false,
209
+ "special": true
210
+ },
211
+ "26": {
212
+ "content": "<|IP_ADDRESS|>",
213
+ "lstrip": false,
214
+ "normalized": false,
215
+ "rstrip": false,
216
+ "single_word": false,
217
+ "special": true
218
+ },
219
+ "27": {
220
+ "content": "<|URL|>",
221
+ "lstrip": false,
222
+ "normalized": false,
223
+ "rstrip": false,
224
+ "single_word": false,
225
+ "special": true
226
+ },
227
+ "28": {
228
+ "content": "<|DATE|>",
229
+ "lstrip": false,
230
+ "normalized": false,
231
+ "rstrip": false,
232
+ "single_word": false,
233
+ "special": true
234
+ },
235
+ "29": {
236
+ "content": "<|TIME|>",
237
+ "lstrip": false,
238
+ "normalized": false,
239
+ "rstrip": false,
240
+ "single_word": false,
241
+ "special": true
242
+ }
243
+ },
244
+ "additional_special_tokens": [
245
+ "<|UNUSED_0|>",
246
+ "<|UNUSED_1|>",
247
+ "<|UNUSED_2|>",
248
+ "<|UNUSED_3|>",
249
+ "<|UNUSED_4|>",
250
+ "<|UNUSED_5|>",
251
+ "<|UNUSED_6|>",
252
+ "<|UNUSED_7|>",
253
+ "<|UNUSED_8|>",
254
+ "<|UNUSED_9|>",
255
+ "<|UNUSED_10|>",
256
+ "<|UNUSED_11|>",
257
+ "<|UNUSED_12|>",
258
+ "<|UNUSED_13|>",
259
+ "<|UNUSED_14|>",
260
+ "<|UNUSED_15|>",
261
+ "<|HISTORY|>",
262
+ "<|PHONE_NUMBER|>",
263
+ "<|EMAIL|>",
264
+ "<|IP_ADDRESS|>",
265
+ "<|URL|>",
266
+ "<|DATE|>",
267
+ "<|TIME|>"
268
+ ],
269
+ "clean_up_tokenization_spaces": true,
270
+ "cls_token": "[CLS]",
271
+ "extra_special_tokens": {},
272
+ "mask_token": "[MASK]",
273
+ "max_len": 512,
274
+ "model_input_names": [
275
+ "input_ids",
276
+ "attention_mask"
277
+ ],
278
+ "model_max_length": 8192,
279
+ "pad_token": "[PAD]",
280
+ "padding_side": "right",
281
+ "return_tensors": "pt",
282
+ "sep_token": "[SEP]",
283
+ "tokenizer_class": "PreTrainedTokenizer",
284
+ "truncation": true,
285
+ "unk_token": "[UNK]"
286
+ }
train_results.json ADDED
@@ -0,0 +1,7 @@
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 19.752351097178682,
3
+ "train_loss": 0.2892922113949427,
4
+ "train_runtime": 830.7771,
5
+ "train_samples_per_second": 122.777,
6
+ "train_steps_per_second": 1.902
7
+ }
trainer_state.json ADDED
@@ -0,0 +1,2535 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_global_step": 800,
3
+ "best_metric": 74.85714285714286,
4
+ "best_model_checkpoint": "/data/datht163/benchmark/checkpoint-800",
5
+ "epoch": 19.752351097178682,
6
+ "eval_steps": 500,
7
+ "global_step": 1580,
8
+ "is_hyper_param_search": false,
9
+ "is_local_process_zero": true,
10
+ "is_world_process_zero": true,
11
+ "log_history": [
12
+ {
13
+ "epoch": 0.06269592476489028,
14
+ "grad_norm": 20.461719512939453,
15
+ "learning_rate": 6.25e-05,
16
+ "loss": 4.5641,
17
+ "step": 5
18
+ },
19
+ {
20
+ "epoch": 0.12539184952978055,
21
+ "grad_norm": 45.15224075317383,
22
+ "learning_rate": 0.000125,
23
+ "loss": 4.8469,
24
+ "step": 10
25
+ },
26
+ {
27
+ "epoch": 0.18808777429467086,
28
+ "grad_norm": 22.794708251953125,
29
+ "learning_rate": 0.0001875,
30
+ "loss": 4.368,
31
+ "step": 15
32
+ },
33
+ {
34
+ "epoch": 0.2507836990595611,
35
+ "grad_norm": 19.09231185913086,
36
+ "learning_rate": 0.00019999677214588312,
37
+ "loss": 3.9258,
38
+ "step": 20
39
+ },
40
+ {
41
+ "epoch": 0.31347962382445144,
42
+ "grad_norm": 15.92790412902832,
43
+ "learning_rate": 0.0001999836593456696,
44
+ "loss": 3.2234,
45
+ "step": 25
46
+ },
47
+ {
48
+ "epoch": 0.3761755485893417,
49
+ "grad_norm": 19.797821044921875,
50
+ "learning_rate": 0.00019996046118014955,
51
+ "loss": 3.2684,
52
+ "step": 30
53
+ },
54
+ {
55
+ "epoch": 0.438871473354232,
56
+ "grad_norm": 19.880847930908203,
57
+ "learning_rate": 0.00019992717998932507,
58
+ "loss": 3.1656,
59
+ "step": 35
60
+ },
61
+ {
62
+ "epoch": 0.5015673981191222,
63
+ "grad_norm": 19.707406997680664,
64
+ "learning_rate": 0.00019988381913027442,
65
+ "loss": 3.0031,
66
+ "step": 40
67
+ },
68
+ {
69
+ "epoch": 0.5642633228840125,
70
+ "grad_norm": 14.307902336120605,
71
+ "learning_rate": 0.00019983038297681336,
72
+ "loss": 2.7219,
73
+ "step": 45
74
+ },
75
+ {
76
+ "epoch": 0.6269592476489029,
77
+ "grad_norm": 13.697257995605469,
78
+ "learning_rate": 0.00019976687691905393,
79
+ "loss": 2.6117,
80
+ "step": 50
81
+ },
82
+ {
83
+ "epoch": 0.6896551724137931,
84
+ "grad_norm": 20.15327262878418,
85
+ "learning_rate": 0.0001996933073628608,
86
+ "loss": 2.7996,
87
+ "step": 55
88
+ },
89
+ {
90
+ "epoch": 0.7523510971786834,
91
+ "grad_norm": 20.71539878845215,
92
+ "learning_rate": 0.00019960968172920516,
93
+ "loss": 2.5578,
94
+ "step": 60
95
+ },
96
+ {
97
+ "epoch": 0.8150470219435737,
98
+ "grad_norm": 8.937211036682129,
99
+ "learning_rate": 0.00019951600845341595,
100
+ "loss": 2.9141,
101
+ "step": 65
102
+ },
103
+ {
104
+ "epoch": 0.877742946708464,
105
+ "grad_norm": 8.478225708007812,
106
+ "learning_rate": 0.0001994122969843293,
107
+ "loss": 2.8832,
108
+ "step": 70
109
+ },
110
+ {
111
+ "epoch": 0.9404388714733543,
112
+ "grad_norm": 15.695404052734375,
113
+ "learning_rate": 0.00019929855778333516,
114
+ "loss": 2.7437,
115
+ "step": 75
116
+ },
117
+ {
118
+ "epoch": 1.0,
119
+ "grad_norm": 19.672351837158203,
120
+ "learning_rate": 0.00019917480232332224,
121
+ "loss": 2.2395,
122
+ "step": 80
123
+ },
124
+ {
125
+ "epoch": 1.0,
126
+ "eval_loss": 0.6336904764175415,
127
+ "eval_macro_f1": 73.26745178358081,
128
+ "eval_macro_precision": 75.24521017228876,
129
+ "eval_macro_recall": 73.14285714285714,
130
+ "eval_micro_f1": 73.14285714285714,
131
+ "eval_micro_precision": 73.14285714285714,
132
+ "eval_micro_recall": 73.14285714285714,
133
+ "eval_runtime": 5.7289,
134
+ "eval_samples_per_second": 183.283,
135
+ "eval_steps_per_second": 11.521,
136
+ "step": 80
137
+ },
138
+ {
139
+ "epoch": 1.0626959247648902,
140
+ "grad_norm": 14.15684700012207,
141
+ "learning_rate": 0.0001990410430875205,
142
+ "loss": 1.4396,
143
+ "step": 85
144
+ },
145
+ {
146
+ "epoch": 1.1253918495297806,
147
+ "grad_norm": 20.00255012512207,
148
+ "learning_rate": 0.00019889729356824235,
149
+ "loss": 1.3728,
150
+ "step": 90
151
+ },
152
+ {
153
+ "epoch": 1.188087774294671,
154
+ "grad_norm": 13.56872272491455,
155
+ "learning_rate": 0.00019874356826552129,
156
+ "loss": 1.6022,
157
+ "step": 95
158
+ },
159
+ {
160
+ "epoch": 1.250783699059561,
161
+ "grad_norm": 6.213724613189697,
162
+ "learning_rate": 0.00019857988268564953,
163
+ "loss": 1.2819,
164
+ "step": 100
165
+ },
166
+ {
167
+ "epoch": 1.3134796238244515,
168
+ "grad_norm": 28.165010452270508,
169
+ "learning_rate": 0.0001984062533396137,
170
+ "loss": 1.1332,
171
+ "step": 105
172
+ },
173
+ {
174
+ "epoch": 1.3761755485893417,
175
+ "grad_norm": 15.799768447875977,
176
+ "learning_rate": 0.00019822269774142954,
177
+ "loss": 1.244,
178
+ "step": 110
179
+ },
180
+ {
181
+ "epoch": 1.438871473354232,
182
+ "grad_norm": 21.20913314819336,
183
+ "learning_rate": 0.0001980292344063752,
184
+ "loss": 1.4427,
185
+ "step": 115
186
+ },
187
+ {
188
+ "epoch": 1.5015673981191222,
189
+ "grad_norm": 18.147506713867188,
190
+ "learning_rate": 0.0001978258828491236,
191
+ "loss": 1.4596,
192
+ "step": 120
193
+ },
194
+ {
195
+ "epoch": 1.5642633228840124,
196
+ "grad_norm": 17.44733428955078,
197
+ "learning_rate": 0.00019761266358177398,
198
+ "loss": 0.9827,
199
+ "step": 125
200
+ },
201
+ {
202
+ "epoch": 1.626959247648903,
203
+ "grad_norm": 18.33002471923828,
204
+ "learning_rate": 0.00019738959811178272,
205
+ "loss": 1.2366,
206
+ "step": 130
207
+ },
208
+ {
209
+ "epoch": 1.6896551724137931,
210
+ "grad_norm": 16.217641830444336,
211
+ "learning_rate": 0.00019715670893979414,
212
+ "loss": 1.2665,
213
+ "step": 135
214
+ },
215
+ {
216
+ "epoch": 1.7523510971786833,
217
+ "grad_norm": 16.21353530883789,
218
+ "learning_rate": 0.00019691401955737072,
219
+ "loss": 1.5896,
220
+ "step": 140
221
+ },
222
+ {
223
+ "epoch": 1.8150470219435737,
224
+ "grad_norm": 11.688411712646484,
225
+ "learning_rate": 0.0001966615544446234,
226
+ "loss": 1.5775,
227
+ "step": 145
228
+ },
229
+ {
230
+ "epoch": 1.877742946708464,
231
+ "grad_norm": 14.346232414245605,
232
+ "learning_rate": 0.0001963993390677424,
233
+ "loss": 1.1833,
234
+ "step": 150
235
+ },
236
+ {
237
+ "epoch": 1.9404388714733543,
238
+ "grad_norm": 6.273299694061279,
239
+ "learning_rate": 0.00019612739987642845,
240
+ "loss": 1.0552,
241
+ "step": 155
242
+ },
243
+ {
244
+ "epoch": 2.0,
245
+ "grad_norm": 17.62116050720215,
246
+ "learning_rate": 0.00019584576430122473,
247
+ "loss": 1.0835,
248
+ "step": 160
249
+ },
250
+ {
251
+ "epoch": 2.0,
252
+ "eval_loss": 0.7825595140457153,
253
+ "eval_macro_f1": 73.56897234236968,
254
+ "eval_macro_precision": 73.7348812342735,
255
+ "eval_macro_recall": 73.90476190476191,
256
+ "eval_micro_f1": 73.90476190476191,
257
+ "eval_micro_precision": 73.90476190476191,
258
+ "eval_micro_recall": 73.90476190476191,
259
+ "eval_runtime": 3.4148,
260
+ "eval_samples_per_second": 307.482,
261
+ "eval_steps_per_second": 19.327,
262
+ "step": 160
263
+ },
264
+ {
265
+ "epoch": 2.06269592476489,
266
+ "grad_norm": 4.175657749176025,
267
+ "learning_rate": 0.00019555446075075,
268
+ "loss": 0.3124,
269
+ "step": 165
270
+ },
271
+ {
272
+ "epoch": 2.1253918495297803,
273
+ "grad_norm": 9.987613677978516,
274
+ "learning_rate": 0.00019525351860883293,
275
+ "loss": 0.3053,
276
+ "step": 170
277
+ },
278
+ {
279
+ "epoch": 2.188087774294671,
280
+ "grad_norm": 27.39342498779297,
281
+ "learning_rate": 0.00019494296823154835,
282
+ "loss": 0.2825,
283
+ "step": 175
284
+ },
285
+ {
286
+ "epoch": 2.250783699059561,
287
+ "grad_norm": 5.328784942626953,
288
+ "learning_rate": 0.000194622840944155,
289
+ "loss": 0.32,
290
+ "step": 180
291
+ },
292
+ {
293
+ "epoch": 2.3134796238244513,
294
+ "grad_norm": 8.484028816223145,
295
+ "learning_rate": 0.00019429316903793583,
296
+ "loss": 0.0699,
297
+ "step": 185
298
+ },
299
+ {
300
+ "epoch": 2.376175548589342,
301
+ "grad_norm": 16.879404067993164,
302
+ "learning_rate": 0.00019395398576694086,
303
+ "loss": 0.2403,
304
+ "step": 190
305
+ },
306
+ {
307
+ "epoch": 2.438871473354232,
308
+ "grad_norm": 49.66862487792969,
309
+ "learning_rate": 0.0001936053253446327,
310
+ "loss": 0.2737,
311
+ "step": 195
312
+ },
313
+ {
314
+ "epoch": 2.501567398119122,
315
+ "grad_norm": 17.669599533081055,
316
+ "learning_rate": 0.00019324722294043558,
317
+ "loss": 0.4838,
318
+ "step": 200
319
+ },
320
+ {
321
+ "epoch": 2.5642633228840124,
322
+ "grad_norm": 14.710466384887695,
323
+ "learning_rate": 0.00019287971467618766,
324
+ "loss": 0.2973,
325
+ "step": 205
326
+ },
327
+ {
328
+ "epoch": 2.626959247648903,
329
+ "grad_norm": 25.36009979248047,
330
+ "learning_rate": 0.00019250283762249748,
331
+ "loss": 0.4224,
332
+ "step": 210
333
+ },
334
+ {
335
+ "epoch": 2.689655172413793,
336
+ "grad_norm": 12.651631355285645,
337
+ "learning_rate": 0.00019211662979500468,
338
+ "loss": 0.1925,
339
+ "step": 215
340
+ },
341
+ {
342
+ "epoch": 2.7523510971786833,
343
+ "grad_norm": 14.720685958862305,
344
+ "learning_rate": 0.00019172113015054532,
345
+ "loss": 0.3322,
346
+ "step": 220
347
+ },
348
+ {
349
+ "epoch": 2.815047021943574,
350
+ "grad_norm": 15.48449993133545,
351
+ "learning_rate": 0.00019131637858322225,
352
+ "loss": 0.1582,
353
+ "step": 225
354
+ },
355
+ {
356
+ "epoch": 2.877742946708464,
357
+ "grad_norm": 8.015260696411133,
358
+ "learning_rate": 0.00019090241592038113,
359
+ "loss": 0.3334,
360
+ "step": 230
361
+ },
362
+ {
363
+ "epoch": 2.9404388714733543,
364
+ "grad_norm": 20.96394157409668,
365
+ "learning_rate": 0.00019047928391849195,
366
+ "loss": 0.2362,
367
+ "step": 235
368
+ },
369
+ {
370
+ "epoch": 3.0,
371
+ "grad_norm": 37.81004333496094,
372
+ "learning_rate": 0.00019004702525893732,
373
+ "loss": 0.2647,
374
+ "step": 240
375
+ },
376
+ {
377
+ "epoch": 3.0,
378
+ "eval_loss": 1.3779538869857788,
379
+ "eval_macro_f1": 73.11543717930599,
380
+ "eval_macro_precision": 73.228985585862,
381
+ "eval_macro_recall": 73.23809523809524,
382
+ "eval_micro_f1": 73.23809523809524,
383
+ "eval_micro_precision": 73.23809523809524,
384
+ "eval_micro_recall": 73.23809523809524,
385
+ "eval_runtime": 3.6271,
386
+ "eval_samples_per_second": 289.487,
387
+ "eval_steps_per_second": 18.196,
388
+ "step": 240
389
+ },
390
+ {
391
+ "epoch": 3.06269592476489,
392
+ "grad_norm": 8.566853523254395,
393
+ "learning_rate": 0.0001896056835437069,
394
+ "loss": 0.0999,
395
+ "step": 245
396
+ },
397
+ {
398
+ "epoch": 3.1253918495297803,
399
+ "grad_norm": 8.521112442016602,
400
+ "learning_rate": 0.0001891553032909996,
401
+ "loss": 0.0719,
402
+ "step": 250
403
+ },
404
+ {
405
+ "epoch": 3.188087774294671,
406
+ "grad_norm": 16.376962661743164,
407
+ "learning_rate": 0.00018869592993073258,
408
+ "loss": 0.2835,
409
+ "step": 255
410
+ },
411
+ {
412
+ "epoch": 3.250783699059561,
413
+ "grad_norm": 33.38531494140625,
414
+ "learning_rate": 0.0001882276097999592,
415
+ "loss": 0.2702,
416
+ "step": 260
417
+ },
418
+ {
419
+ "epoch": 3.3134796238244513,
420
+ "grad_norm": 24.302696228027344,
421
+ "learning_rate": 0.0001877503901381947,
422
+ "loss": 0.2763,
423
+ "step": 265
424
+ },
425
+ {
426
+ "epoch": 3.376175548589342,
427
+ "grad_norm": 18.85578727722168,
428
+ "learning_rate": 0.0001872643190826512,
429
+ "loss": 0.5174,
430
+ "step": 270
431
+ },
432
+ {
433
+ "epoch": 3.438871473354232,
434
+ "grad_norm": 53.06370162963867,
435
+ "learning_rate": 0.00018676944566338213,
436
+ "loss": 0.2884,
437
+ "step": 275
438
+ },
439
+ {
440
+ "epoch": 3.501567398119122,
441
+ "grad_norm": 34.46812438964844,
442
+ "learning_rate": 0.0001862658197983366,
443
+ "loss": 0.2232,
444
+ "step": 280
445
+ },
446
+ {
447
+ "epoch": 3.5642633228840124,
448
+ "grad_norm": 18.52553939819336,
449
+ "learning_rate": 0.00018575349228832397,
450
+ "loss": 0.1158,
451
+ "step": 285
452
+ },
453
+ {
454
+ "epoch": 3.626959247648903,
455
+ "grad_norm": 0.9173059463500977,
456
+ "learning_rate": 0.00018523251481188986,
457
+ "loss": 0.2123,
458
+ "step": 290
459
+ },
460
+ {
461
+ "epoch": 3.689655172413793,
462
+ "grad_norm": 12.346527099609375,
463
+ "learning_rate": 0.00018470293992010312,
464
+ "loss": 0.2126,
465
+ "step": 295
466
+ },
467
+ {
468
+ "epoch": 3.7523510971786833,
469
+ "grad_norm": 84.08663177490234,
470
+ "learning_rate": 0.00018416482103125506,
471
+ "loss": 0.222,
472
+ "step": 300
473
+ },
474
+ {
475
+ "epoch": 3.815047021943574,
476
+ "grad_norm": 19.547080993652344,
477
+ "learning_rate": 0.0001836182124254711,
478
+ "loss": 0.1389,
479
+ "step": 305
480
+ },
481
+ {
482
+ "epoch": 3.877742946708464,
483
+ "grad_norm": 1.2908215522766113,
484
+ "learning_rate": 0.00018306316923923563,
485
+ "loss": 0.1632,
486
+ "step": 310
487
+ },
488
+ {
489
+ "epoch": 3.9404388714733543,
490
+ "grad_norm": 30.96961784362793,
491
+ "learning_rate": 0.00018249974745983023,
492
+ "loss": 0.2089,
493
+ "step": 315
494
+ },
495
+ {
496
+ "epoch": 4.0,
497
+ "grad_norm": 13.291217803955078,
498
+ "learning_rate": 0.00018192800391968642,
499
+ "loss": 0.3228,
500
+ "step": 320
501
+ },
502
+ {
503
+ "epoch": 4.0,
504
+ "eval_loss": 1.7461718320846558,
505
+ "eval_macro_f1": 72.72572059706258,
506
+ "eval_macro_precision": 72.71473132903009,
507
+ "eval_macro_recall": 72.76190476190476,
508
+ "eval_micro_f1": 72.76190476190476,
509
+ "eval_micro_precision": 72.76190476190476,
510
+ "eval_micro_recall": 72.76190476190476,
511
+ "eval_runtime": 4.7844,
512
+ "eval_samples_per_second": 219.463,
513
+ "eval_steps_per_second": 13.795,
514
+ "step": 320
515
+ },
516
+ {
517
+ "epoch": 4.06269592476489,
518
+ "grad_norm": 22.18653678894043,
519
+ "learning_rate": 0.00018134799629065276,
520
+ "loss": 0.3768,
521
+ "step": 325
522
+ },
523
+ {
524
+ "epoch": 4.12539184952978,
525
+ "grad_norm": 18.046934127807617,
526
+ "learning_rate": 0.00018075978307817764,
527
+ "loss": 0.1879,
528
+ "step": 330
529
+ },
530
+ {
531
+ "epoch": 4.1880877742946705,
532
+ "grad_norm": 8.421941757202148,
533
+ "learning_rate": 0.00018016342361540773,
534
+ "loss": 0.2067,
535
+ "step": 335
536
+ },
537
+ {
538
+ "epoch": 4.250783699059561,
539
+ "grad_norm": 37.9600830078125,
540
+ "learning_rate": 0.0001795589780572031,
541
+ "loss": 0.1612,
542
+ "step": 340
543
+ },
544
+ {
545
+ "epoch": 4.313479623824452,
546
+ "grad_norm": 2.552628993988037,
547
+ "learning_rate": 0.0001789465073740693,
548
+ "loss": 0.0256,
549
+ "step": 345
550
+ },
551
+ {
552
+ "epoch": 4.376175548589342,
553
+ "grad_norm": 45.092559814453125,
554
+ "learning_rate": 0.00017832607334600746,
555
+ "loss": 0.2179,
556
+ "step": 350
557
+ },
558
+ {
559
+ "epoch": 4.438871473354232,
560
+ "grad_norm": 1.6102479696273804,
561
+ "learning_rate": 0.00017769773855628226,
562
+ "loss": 0.1914,
563
+ "step": 355
564
+ },
565
+ {
566
+ "epoch": 4.501567398119122,
567
+ "grad_norm": 27.624589920043945,
568
+ "learning_rate": 0.0001770615663851093,
569
+ "loss": 0.1637,
570
+ "step": 360
571
+ },
572
+ {
573
+ "epoch": 4.564263322884012,
574
+ "grad_norm": 31.627342224121094,
575
+ "learning_rate": 0.00017641762100326208,
576
+ "loss": 0.1678,
577
+ "step": 365
578
+ },
579
+ {
580
+ "epoch": 4.6269592476489025,
581
+ "grad_norm": 17.774938583374023,
582
+ "learning_rate": 0.0001757659673655986,
583
+ "loss": 0.0717,
584
+ "step": 370
585
+ },
586
+ {
587
+ "epoch": 4.689655172413794,
588
+ "grad_norm": 5.605461597442627,
589
+ "learning_rate": 0.00017510667120450998,
590
+ "loss": 0.2968,
591
+ "step": 375
592
+ },
593
+ {
594
+ "epoch": 4.752351097178684,
595
+ "grad_norm": 27.00714683532715,
596
+ "learning_rate": 0.00017443979902328956,
597
+ "loss": 0.5085,
598
+ "step": 380
599
+ },
600
+ {
601
+ "epoch": 4.815047021943574,
602
+ "grad_norm": 26.765687942504883,
603
+ "learning_rate": 0.0001737654180894249,
604
+ "loss": 0.1005,
605
+ "step": 385
606
+ },
607
+ {
608
+ "epoch": 4.877742946708464,
609
+ "grad_norm": 15.505696296691895,
610
+ "learning_rate": 0.00017308359642781242,
611
+ "loss": 0.197,
612
+ "step": 390
613
+ },
614
+ {
615
+ "epoch": 4.940438871473354,
616
+ "grad_norm": 7.485856056213379,
617
+ "learning_rate": 0.00017239440281389582,
618
+ "loss": 0.1499,
619
+ "step": 395
620
+ },
621
+ {
622
+ "epoch": 5.0,
623
+ "grad_norm": 2.3405601978302,
624
+ "learning_rate": 0.00017169790676672858,
625
+ "loss": 0.0747,
626
+ "step": 400
627
+ },
628
+ {
629
+ "epoch": 5.0,
630
+ "eval_loss": 1.7594363689422607,
631
+ "eval_macro_f1": 74.17047440340497,
632
+ "eval_macro_precision": 75.04188054945206,
633
+ "eval_macro_recall": 73.99999999999999,
634
+ "eval_micro_f1": 74.0,
635
+ "eval_micro_precision": 74.0,
636
+ "eval_micro_recall": 74.0,
637
+ "eval_runtime": 4.2956,
638
+ "eval_samples_per_second": 244.435,
639
+ "eval_steps_per_second": 15.364,
640
+ "step": 400
641
+ },
642
+ {
643
+ "epoch": 5.06269592476489,
644
+ "grad_norm": 11.809943199157715,
645
+ "learning_rate": 0.00017099417854196165,
646
+ "loss": 0.0426,
647
+ "step": 405
648
+ },
649
+ {
650
+ "epoch": 5.12539184952978,
651
+ "grad_norm": 4.720785140991211,
652
+ "learning_rate": 0.00017028328912475668,
653
+ "loss": 0.044,
654
+ "step": 410
655
+ },
656
+ {
657
+ "epoch": 5.1880877742946705,
658
+ "grad_norm": 14.604973793029785,
659
+ "learning_rate": 0.00016956531022262585,
660
+ "loss": 0.1949,
661
+ "step": 415
662
+ },
663
+ {
664
+ "epoch": 5.250783699059561,
665
+ "grad_norm": 0.6709997653961182,
666
+ "learning_rate": 0.00016884031425819853,
667
+ "loss": 0.295,
668
+ "step": 420
669
+ },
670
+ {
671
+ "epoch": 5.313479623824452,
672
+ "grad_norm": 69.35284423828125,
673
+ "learning_rate": 0.00016810837436191623,
674
+ "loss": 0.4807,
675
+ "step": 425
676
+ },
677
+ {
678
+ "epoch": 5.376175548589342,
679
+ "grad_norm": 24.146665573120117,
680
+ "learning_rate": 0.00016736956436465573,
681
+ "loss": 0.5015,
682
+ "step": 430
683
+ },
684
+ {
685
+ "epoch": 5.438871473354232,
686
+ "grad_norm": 3.2245795726776123,
687
+ "learning_rate": 0.0001666239587902819,
688
+ "loss": 0.1359,
689
+ "step": 435
690
+ },
691
+ {
692
+ "epoch": 5.501567398119122,
693
+ "grad_norm": 8.290742874145508,
694
+ "learning_rate": 0.00016587163284813032,
695
+ "loss": 0.0412,
696
+ "step": 440
697
+ },
698
+ {
699
+ "epoch": 5.564263322884012,
700
+ "grad_norm": 21.256576538085938,
701
+ "learning_rate": 0.00016511266242542102,
702
+ "loss": 0.0626,
703
+ "step": 445
704
+ },
705
+ {
706
+ "epoch": 5.6269592476489025,
707
+ "grad_norm": 1.178815245628357,
708
+ "learning_rate": 0.00016434712407960373,
709
+ "loss": 0.0422,
710
+ "step": 450
711
+ },
712
+ {
713
+ "epoch": 5.689655172413794,
714
+ "grad_norm": 19.474632263183594,
715
+ "learning_rate": 0.00016357509503063538,
716
+ "loss": 0.3287,
717
+ "step": 455
718
+ },
719
+ {
720
+ "epoch": 5.752351097178684,
721
+ "grad_norm": 38.11343002319336,
722
+ "learning_rate": 0.00016279665315319114,
723
+ "loss": 0.3171,
724
+ "step": 460
725
+ },
726
+ {
727
+ "epoch": 5.815047021943574,
728
+ "grad_norm": 63.41005325317383,
729
+ "learning_rate": 0.00016201187696880887,
730
+ "loss": 0.2117,
731
+ "step": 465
732
+ },
733
+ {
734
+ "epoch": 5.877742946708464,
735
+ "grad_norm": 1.7413610219955444,
736
+ "learning_rate": 0.00016122084563796905,
737
+ "loss": 0.0836,
738
+ "step": 470
739
+ },
740
+ {
741
+ "epoch": 5.940438871473354,
742
+ "grad_norm": 25.483036041259766,
743
+ "learning_rate": 0.00016042363895210946,
744
+ "loss": 0.4536,
745
+ "step": 475
746
+ },
747
+ {
748
+ "epoch": 6.0,
749
+ "grad_norm": 13.102269172668457,
750
+ "learning_rate": 0.00015962033732557686,
751
+ "loss": 0.0868,
752
+ "step": 480
753
+ },
754
+ {
755
+ "epoch": 6.0,
756
+ "eval_loss": 1.5808621644973755,
757
+ "eval_macro_f1": 73.24384906714903,
758
+ "eval_macro_precision": 73.61945929207427,
759
+ "eval_macro_recall": 73.61904761904762,
760
+ "eval_micro_f1": 73.61904761904762,
761
+ "eval_micro_precision": 73.61904761904762,
762
+ "eval_micro_recall": 73.61904761904762,
763
+ "eval_runtime": 5.0263,
764
+ "eval_samples_per_second": 208.902,
765
+ "eval_steps_per_second": 13.131,
766
+ "step": 480
767
+ },
768
+ {
769
+ "epoch": 6.06269592476489,
770
+ "grad_norm": 6.51555871963501,
771
+ "learning_rate": 0.00015881102178751553,
772
+ "loss": 0.0678,
773
+ "step": 485
774
+ },
775
+ {
776
+ "epoch": 6.12539184952978,
777
+ "grad_norm": 19.54594612121582,
778
+ "learning_rate": 0.00015799577397369375,
779
+ "loss": 0.0731,
780
+ "step": 490
781
+ },
782
+ {
783
+ "epoch": 6.1880877742946705,
784
+ "grad_norm": 7.0642266273498535,
785
+ "learning_rate": 0.0001571746761182693,
786
+ "loss": 0.0848,
787
+ "step": 495
788
+ },
789
+ {
790
+ "epoch": 6.250783699059561,
791
+ "grad_norm": 4.012495040893555,
792
+ "learning_rate": 0.00015634781104549442,
793
+ "loss": 0.1141,
794
+ "step": 500
795
+ },
796
+ {
797
+ "epoch": 6.313479623824452,
798
+ "grad_norm": 2.1113383769989014,
799
+ "learning_rate": 0.00015551526216136144,
800
+ "loss": 0.0945,
801
+ "step": 505
802
+ },
803
+ {
804
+ "epoch": 6.376175548589342,
805
+ "grad_norm": 37.096473693847656,
806
+ "learning_rate": 0.00015467711344518942,
807
+ "loss": 0.225,
808
+ "step": 510
809
+ },
810
+ {
811
+ "epoch": 6.438871473354232,
812
+ "grad_norm": 25.69189453125,
813
+ "learning_rate": 0.00015383344944115324,
814
+ "loss": 0.0556,
815
+ "step": 515
816
+ },
817
+ {
818
+ "epoch": 6.501567398119122,
819
+ "grad_norm": 4.062742710113525,
820
+ "learning_rate": 0.00015298435524975572,
821
+ "loss": 0.0765,
822
+ "step": 520
823
+ },
824
+ {
825
+ "epoch": 6.564263322884012,
826
+ "grad_norm": 14.947588920593262,
827
+ "learning_rate": 0.00015212991651924326,
828
+ "loss": 0.0302,
829
+ "step": 525
830
+ },
831
+ {
832
+ "epoch": 6.6269592476489025,
833
+ "grad_norm": 28.04601287841797,
834
+ "learning_rate": 0.0001512702194369668,
835
+ "loss": 0.4446,
836
+ "step": 530
837
+ },
838
+ {
839
+ "epoch": 6.689655172413794,
840
+ "grad_norm": 0.8309017419815063,
841
+ "learning_rate": 0.00015040535072068778,
842
+ "loss": 0.1228,
843
+ "step": 535
844
+ },
845
+ {
846
+ "epoch": 6.752351097178684,
847
+ "grad_norm": 0.3538524806499481,
848
+ "learning_rate": 0.00014953539760983122,
849
+ "loss": 0.1258,
850
+ "step": 540
851
+ },
852
+ {
853
+ "epoch": 6.815047021943574,
854
+ "grad_norm": 0.2524064779281616,
855
+ "learning_rate": 0.00014866044785668563,
856
+ "loss": 0.036,
857
+ "step": 545
858
+ },
859
+ {
860
+ "epoch": 6.877742946708464,
861
+ "grad_norm": 46.764183044433594,
862
+ "learning_rate": 0.00014778058971755154,
863
+ "loss": 0.294,
864
+ "step": 550
865
+ },
866
+ {
867
+ "epoch": 6.940438871473354,
868
+ "grad_norm": 33.56895446777344,
869
+ "learning_rate": 0.00014689591194383896,
870
+ "loss": 0.1452,
871
+ "step": 555
872
+ },
873
+ {
874
+ "epoch": 7.0,
875
+ "grad_norm": 0.07464821636676788,
876
+ "learning_rate": 0.00014600650377311522,
877
+ "loss": 0.0801,
878
+ "step": 560
879
+ },
880
+ {
881
+ "epoch": 7.0,
882
+ "eval_loss": 1.7074089050292969,
883
+ "eval_macro_f1": 74.3995628020244,
884
+ "eval_macro_precision": 74.43261914638138,
885
+ "eval_macro_recall": 74.38095238095238,
886
+ "eval_micro_f1": 74.38095238095238,
887
+ "eval_micro_precision": 74.38095238095238,
888
+ "eval_micro_recall": 74.38095238095238,
889
+ "eval_runtime": 5.0045,
890
+ "eval_samples_per_second": 209.812,
891
+ "eval_steps_per_second": 13.188,
892
+ "step": 560
893
+ },
894
+ {
895
+ "epoch": 7.06269592476489,
896
+ "grad_norm": 24.801523208618164,
897
+ "learning_rate": 0.00014511245492010335,
898
+ "loss": 0.0224,
899
+ "step": 565
900
+ },
901
+ {
902
+ "epoch": 7.12539184952978,
903
+ "grad_norm": 2.976022720336914,
904
+ "learning_rate": 0.00014421385556763266,
905
+ "loss": 0.1211,
906
+ "step": 570
907
+ },
908
+ {
909
+ "epoch": 7.1880877742946705,
910
+ "grad_norm": 0.37286555767059326,
911
+ "learning_rate": 0.00014331079635754193,
912
+ "loss": 0.0327,
913
+ "step": 575
914
+ },
915
+ {
916
+ "epoch": 7.250783699059561,
917
+ "grad_norm": 0.7439866662025452,
918
+ "learning_rate": 0.0001424033683815365,
919
+ "loss": 0.0382,
920
+ "step": 580
921
+ },
922
+ {
923
+ "epoch": 7.313479623824452,
924
+ "grad_norm": 24.909921646118164,
925
+ "learning_rate": 0.00014149166317199954,
926
+ "loss": 0.0603,
927
+ "step": 585
928
+ },
929
+ {
930
+ "epoch": 7.376175548589342,
931
+ "grad_norm": 9.340410232543945,
932
+ "learning_rate": 0.0001405757726927595,
933
+ "loss": 0.2319,
934
+ "step": 590
935
+ },
936
+ {
937
+ "epoch": 7.438871473354232,
938
+ "grad_norm": 3.947765350341797,
939
+ "learning_rate": 0.00013965578932981346,
940
+ "loss": 0.0448,
941
+ "step": 595
942
+ },
943
+ {
944
+ "epoch": 7.501567398119122,
945
+ "grad_norm": 0.29921436309814453,
946
+ "learning_rate": 0.00013873180588200827,
947
+ "loss": 0.006,
948
+ "step": 600
949
+ },
950
+ {
951
+ "epoch": 7.564263322884012,
952
+ "grad_norm": 0.11227085441350937,
953
+ "learning_rate": 0.00013780391555167992,
954
+ "loss": 0.0037,
955
+ "step": 605
956
+ },
957
+ {
958
+ "epoch": 7.6269592476489025,
959
+ "grad_norm": 0.041259847581386566,
960
+ "learning_rate": 0.0001368722119352521,
961
+ "loss": 0.0149,
962
+ "step": 610
963
+ },
964
+ {
965
+ "epoch": 7.689655172413794,
966
+ "grad_norm": 41.212154388427734,
967
+ "learning_rate": 0.00013593678901379524,
968
+ "loss": 0.0205,
969
+ "step": 615
970
+ },
971
+ {
972
+ "epoch": 7.752351097178684,
973
+ "grad_norm": 12.241500854492188,
974
+ "learning_rate": 0.00013499774114354655,
975
+ "loss": 0.0038,
976
+ "step": 620
977
+ },
978
+ {
979
+ "epoch": 7.815047021943574,
980
+ "grad_norm": 55.03153991699219,
981
+ "learning_rate": 0.00013405516304639234,
982
+ "loss": 0.029,
983
+ "step": 625
984
+ },
985
+ {
986
+ "epoch": 7.877742946708464,
987
+ "grad_norm": 0.5877093076705933,
988
+ "learning_rate": 0.00013310914980031334,
989
+ "loss": 0.0072,
990
+ "step": 630
991
+ },
992
+ {
993
+ "epoch": 7.940438871473354,
994
+ "grad_norm": 34.05458068847656,
995
+ "learning_rate": 0.00013215979682979428,
996
+ "loss": 0.0398,
997
+ "step": 635
998
+ },
999
+ {
1000
+ "epoch": 8.0,
1001
+ "grad_norm": 86.24744415283203,
1002
+ "learning_rate": 0.00013120719989619833,
1003
+ "loss": 0.1489,
1004
+ "step": 640
1005
+ },
1006
+ {
1007
+ "epoch": 8.0,
1008
+ "eval_loss": 3.0941507816314697,
1009
+ "eval_macro_f1": 73.75591198712105,
1010
+ "eval_macro_precision": 74.97115819170575,
1011
+ "eval_macro_recall": 73.52380952380952,
1012
+ "eval_micro_f1": 73.52380952380952,
1013
+ "eval_micro_precision": 73.52380952380952,
1014
+ "eval_micro_recall": 73.52380952380952,
1015
+ "eval_runtime": 4.8846,
1016
+ "eval_samples_per_second": 214.962,
1017
+ "eval_steps_per_second": 13.512,
1018
+ "step": 640
1019
+ },
1020
+ {
1021
+ "epoch": 8.06269592476489,
1022
+ "grad_norm": 0.29403501749038696,
1023
+ "learning_rate": 0.0001302514550881076,
1024
+ "loss": 0.1469,
1025
+ "step": 645
1026
+ },
1027
+ {
1028
+ "epoch": 8.12539184952978,
1029
+ "grad_norm": 0.0023457759525626898,
1030
+ "learning_rate": 0.0001292926588116308,
1031
+ "loss": 0.0521,
1032
+ "step": 650
1033
+ },
1034
+ {
1035
+ "epoch": 8.18808777429467,
1036
+ "grad_norm": 0.0002847153227776289,
1037
+ "learning_rate": 0.00012833090778067863,
1038
+ "loss": 0.0004,
1039
+ "step": 655
1040
+ },
1041
+ {
1042
+ "epoch": 8.25078369905956,
1043
+ "grad_norm": 0.07453305274248123,
1044
+ "learning_rate": 0.0001273662990072083,
1045
+ "loss": 0.0002,
1046
+ "step": 660
1047
+ },
1048
+ {
1049
+ "epoch": 8.31347962382445,
1050
+ "grad_norm": 0.015363110229372978,
1051
+ "learning_rate": 0.00012639892979143789,
1052
+ "loss": 0.0349,
1053
+ "step": 665
1054
+ },
1055
+ {
1056
+ "epoch": 8.376175548589341,
1057
+ "grad_norm": 0.003746453206986189,
1058
+ "learning_rate": 0.00012542889771203166,
1059
+ "loss": 0.0002,
1060
+ "step": 670
1061
+ },
1062
+ {
1063
+ "epoch": 8.438871473354231,
1064
+ "grad_norm": 0.07255641371011734,
1065
+ "learning_rate": 0.00012445630061625729,
1066
+ "loss": 0.0391,
1067
+ "step": 675
1068
+ },
1069
+ {
1070
+ "epoch": 8.501567398119121,
1071
+ "grad_norm": 0.006769082974642515,
1072
+ "learning_rate": 0.00012348123661011601,
1073
+ "loss": 0.0626,
1074
+ "step": 680
1075
+ },
1076
+ {
1077
+ "epoch": 8.564263322884013,
1078
+ "grad_norm": 1.8504444360733032,
1079
+ "learning_rate": 0.00012250380404844665,
1080
+ "loss": 0.1377,
1081
+ "step": 685
1082
+ },
1083
+ {
1084
+ "epoch": 8.626959247648903,
1085
+ "grad_norm": 30.328672409057617,
1086
+ "learning_rate": 0.00012152410152500453,
1087
+ "loss": 0.0094,
1088
+ "step": 690
1089
+ },
1090
+ {
1091
+ "epoch": 8.689655172413794,
1092
+ "grad_norm": 0.01078982837498188,
1093
+ "learning_rate": 0.00012054222786251634,
1094
+ "loss": 0.0014,
1095
+ "step": 695
1096
+ },
1097
+ {
1098
+ "epoch": 8.752351097178684,
1099
+ "grad_norm": 0.02426566183567047,
1100
+ "learning_rate": 0.00011955828210271187,
1101
+ "loss": 0.0001,
1102
+ "step": 700
1103
+ },
1104
+ {
1105
+ "epoch": 8.815047021943574,
1106
+ "grad_norm": 0.3005918562412262,
1107
+ "learning_rate": 0.00011857236349633358,
1108
+ "loss": 0.1305,
1109
+ "step": 705
1110
+ },
1111
+ {
1112
+ "epoch": 8.877742946708464,
1113
+ "grad_norm": 40.52849578857422,
1114
+ "learning_rate": 0.00011758457149312538,
1115
+ "loss": 0.1226,
1116
+ "step": 710
1117
+ },
1118
+ {
1119
+ "epoch": 8.940438871473354,
1120
+ "grad_norm": 0.03640785068273544,
1121
+ "learning_rate": 0.0001165950057318008,
1122
+ "loss": 0.0081,
1123
+ "step": 715
1124
+ },
1125
+ {
1126
+ "epoch": 9.0,
1127
+ "grad_norm": 0.8364064693450928,
1128
+ "learning_rate": 0.00011560376602999272,
1129
+ "loss": 0.1558,
1130
+ "step": 720
1131
+ },
1132
+ {
1133
+ "epoch": 9.0,
1134
+ "eval_loss": 3.2611937522888184,
1135
+ "eval_macro_f1": 72.56355683169605,
1136
+ "eval_macro_precision": 74.47633152198945,
1137
+ "eval_macro_recall": 72.3809523809524,
1138
+ "eval_micro_f1": 72.38095238095238,
1139
+ "eval_micro_precision": 72.38095238095238,
1140
+ "eval_micro_recall": 72.38095238095238,
1141
+ "eval_runtime": 4.5879,
1142
+ "eval_samples_per_second": 228.86,
1143
+ "eval_steps_per_second": 14.386,
1144
+ "step": 720
1145
+ },
1146
+ {
1147
+ "epoch": 9.06269592476489,
1148
+ "grad_norm": 0.017020411789417267,
1149
+ "learning_rate": 0.00011461095237418454,
1150
+ "loss": 0.0155,
1151
+ "step": 725
1152
+ },
1153
+ {
1154
+ "epoch": 9.12539184952978,
1155
+ "grad_norm": 0.0012181233614683151,
1156
+ "learning_rate": 0.00011361666490962468,
1157
+ "loss": 0.0176,
1158
+ "step": 730
1159
+ },
1160
+ {
1161
+ "epoch": 9.18808777429467,
1162
+ "grad_norm": 0.04490913078188896,
1163
+ "learning_rate": 0.00011262100393022482,
1164
+ "loss": 0.0084,
1165
+ "step": 735
1166
+ },
1167
+ {
1168
+ "epoch": 9.25078369905956,
1169
+ "grad_norm": 0.0009038946591317654,
1170
+ "learning_rate": 0.00011162406986844323,
1171
+ "loss": 0.0,
1172
+ "step": 740
1173
+ },
1174
+ {
1175
+ "epoch": 9.31347962382445,
1176
+ "grad_norm": 0.0007440209737978876,
1177
+ "learning_rate": 0.00011062596328515424,
1178
+ "loss": 0.0207,
1179
+ "step": 745
1180
+ },
1181
+ {
1182
+ "epoch": 9.376175548589341,
1183
+ "grad_norm": 0.00041356749716214836,
1184
+ "learning_rate": 0.00010962678485950455,
1185
+ "loss": 0.0018,
1186
+ "step": 750
1187
+ },
1188
+ {
1189
+ "epoch": 9.438871473354231,
1190
+ "grad_norm": 0.0054762535728514194,
1191
+ "learning_rate": 0.00010862663537875775,
1192
+ "loss": 0.0,
1193
+ "step": 755
1194
+ },
1195
+ {
1196
+ "epoch": 9.501567398119121,
1197
+ "grad_norm": 0.027165431529283524,
1198
+ "learning_rate": 0.00010762561572812788,
1199
+ "loss": 0.0,
1200
+ "step": 760
1201
+ },
1202
+ {
1203
+ "epoch": 9.564263322884013,
1204
+ "grad_norm": 0.002276531420648098,
1205
+ "learning_rate": 0.0001066238268806032,
1206
+ "loss": 0.0102,
1207
+ "step": 765
1208
+ },
1209
+ {
1210
+ "epoch": 9.626959247648903,
1211
+ "grad_norm": 0.14534221589565277,
1212
+ "learning_rate": 0.00010562136988676078,
1213
+ "loss": 0.0001,
1214
+ "step": 770
1215
+ },
1216
+ {
1217
+ "epoch": 9.689655172413794,
1218
+ "grad_norm": 0.0010948021663352847,
1219
+ "learning_rate": 0.00010461834586457398,
1220
+ "loss": 0.0,
1221
+ "step": 775
1222
+ },
1223
+ {
1224
+ "epoch": 9.752351097178684,
1225
+ "grad_norm": 5.5901953601278365e-05,
1226
+ "learning_rate": 0.00010361485598921212,
1227
+ "loss": 0.104,
1228
+ "step": 780
1229
+ },
1230
+ {
1231
+ "epoch": 9.815047021943574,
1232
+ "grad_norm": 0.8184991478919983,
1233
+ "learning_rate": 0.00010261100148283538,
1234
+ "loss": 0.0106,
1235
+ "step": 785
1236
+ },
1237
+ {
1238
+ "epoch": 9.877742946708464,
1239
+ "grad_norm": 0.006693361327052116,
1240
+ "learning_rate": 0.00010160688360438419,
1241
+ "loss": 0.0,
1242
+ "step": 790
1243
+ },
1244
+ {
1245
+ "epoch": 9.940438871473354,
1246
+ "grad_norm": 59.32970428466797,
1247
+ "learning_rate": 0.00010060260363936547,
1248
+ "loss": 0.052,
1249
+ "step": 795
1250
+ },
1251
+ {
1252
+ "epoch": 10.0,
1253
+ "grad_norm": 0.021533485502004623,
1254
+ "learning_rate": 9.959826288963571e-05,
1255
+ "loss": 0.0,
1256
+ "step": 800
1257
+ },
1258
+ {
1259
+ "epoch": 10.0,
1260
+ "eval_loss": 3.001224994659424,
1261
+ "eval_macro_f1": 74.98933185346014,
1262
+ "eval_macro_precision": 75.36484226701617,
1263
+ "eval_macro_recall": 74.85714285714286,
1264
+ "eval_micro_f1": 74.85714285714286,
1265
+ "eval_micro_precision": 74.85714285714286,
1266
+ "eval_micro_recall": 74.85714285714286,
1267
+ "eval_runtime": 4.9316,
1268
+ "eval_samples_per_second": 212.912,
1269
+ "eval_steps_per_second": 13.383,
1270
+ "step": 800
1271
+ },
1272
+ {
1273
+ "epoch": 10.06269592476489,
1274
+ "grad_norm": 0.10546494275331497,
1275
+ "learning_rate": 9.85939626631829e-05,
1276
+ "loss": 0.0001,
1277
+ "step": 805
1278
+ },
1279
+ {
1280
+ "epoch": 10.12539184952978,
1281
+ "grad_norm": 0.0011004558764398098,
1282
+ "learning_rate": 9.758980426390732e-05,
1283
+ "loss": 0.0,
1284
+ "step": 810
1285
+ },
1286
+ {
1287
+ "epoch": 10.18808777429467,
1288
+ "grad_norm": 0.02645757608115673,
1289
+ "learning_rate": 9.658588898140322e-05,
1290
+ "loss": 0.0,
1291
+ "step": 815
1292
+ },
1293
+ {
1294
+ "epoch": 10.25078369905956,
1295
+ "grad_norm": 0.06658319383859634,
1296
+ "learning_rate": 9.558231808074156e-05,
1297
+ "loss": 0.0001,
1298
+ "step": 820
1299
+ },
1300
+ {
1301
+ "epoch": 10.31347962382445,
1302
+ "grad_norm": 0.1821487843990326,
1303
+ "learning_rate": 9.457919279225548e-05,
1304
+ "loss": 0.0,
1305
+ "step": 825
1306
+ },
1307
+ {
1308
+ "epoch": 10.376175548589341,
1309
+ "grad_norm": 0.0002453463093843311,
1310
+ "learning_rate": 9.357661430132915e-05,
1311
+ "loss": 0.0,
1312
+ "step": 830
1313
+ },
1314
+ {
1315
+ "epoch": 10.438871473354231,
1316
+ "grad_norm": 0.0003151444543618709,
1317
+ "learning_rate": 9.257468373819123e-05,
1318
+ "loss": 0.0,
1319
+ "step": 835
1320
+ },
1321
+ {
1322
+ "epoch": 10.501567398119121,
1323
+ "grad_norm": 4.588846059050411e-05,
1324
+ "learning_rate": 9.157350216771378e-05,
1325
+ "loss": 0.0,
1326
+ "step": 840
1327
+ },
1328
+ {
1329
+ "epoch": 10.564263322884013,
1330
+ "grad_norm": 0.0018681439105421305,
1331
+ "learning_rate": 9.057317057921787e-05,
1332
+ "loss": 0.0,
1333
+ "step": 845
1334
+ },
1335
+ {
1336
+ "epoch": 10.626959247648903,
1337
+ "grad_norm": 0.0006441921577788889,
1338
+ "learning_rate": 8.957378987628682e-05,
1339
+ "loss": 0.0,
1340
+ "step": 850
1341
+ },
1342
+ {
1343
+ "epoch": 10.689655172413794,
1344
+ "grad_norm": 0.0007152777980081737,
1345
+ "learning_rate": 8.857546086658789e-05,
1346
+ "loss": 0.0,
1347
+ "step": 855
1348
+ },
1349
+ {
1350
+ "epoch": 10.752351097178684,
1351
+ "grad_norm": 0.0002644763153512031,
1352
+ "learning_rate": 8.757828425170404e-05,
1353
+ "loss": 0.0173,
1354
+ "step": 860
1355
+ },
1356
+ {
1357
+ "epoch": 10.815047021943574,
1358
+ "grad_norm": 0.0004763313045259565,
1359
+ "learning_rate": 8.658236061697586e-05,
1360
+ "loss": 0.0,
1361
+ "step": 865
1362
+ },
1363
+ {
1364
+ "epoch": 10.877742946708464,
1365
+ "grad_norm": 0.07049086689949036,
1366
+ "learning_rate": 8.55877904213558e-05,
1367
+ "loss": 0.0,
1368
+ "step": 870
1369
+ },
1370
+ {
1371
+ "epoch": 10.940438871473354,
1372
+ "grad_norm": 0.00021191804262343794,
1373
+ "learning_rate": 8.459467398727462e-05,
1374
+ "loss": 0.0,
1375
+ "step": 875
1376
+ },
1377
+ {
1378
+ "epoch": 11.0,
1379
+ "grad_norm": 0.0024338788352906704,
1380
+ "learning_rate": 8.360311149052205e-05,
1381
+ "loss": 0.0,
1382
+ "step": 880
1383
+ },
1384
+ {
1385
+ "epoch": 11.0,
1386
+ "eval_loss": 2.944465160369873,
1387
+ "eval_macro_f1": 73.94965715163707,
1388
+ "eval_macro_precision": 73.91418752481633,
1389
+ "eval_macro_recall": 73.99999999999999,
1390
+ "eval_micro_f1": 74.0,
1391
+ "eval_micro_precision": 74.0,
1392
+ "eval_micro_recall": 74.0,
1393
+ "eval_runtime": 3.4955,
1394
+ "eval_samples_per_second": 300.386,
1395
+ "eval_steps_per_second": 18.881,
1396
+ "step": 880
1397
+ },
1398
+ {
1399
+ "epoch": 11.06269592476489,
1400
+ "grad_norm": 0.0007475401507690549,
1401
+ "learning_rate": 8.261320295014182e-05,
1402
+ "loss": 0.0,
1403
+ "step": 885
1404
+ },
1405
+ {
1406
+ "epoch": 11.12539184952978,
1407
+ "grad_norm": 5.158692511031404e-05,
1408
+ "learning_rate": 8.162504821834295e-05,
1409
+ "loss": 0.0,
1410
+ "step": 890
1411
+ },
1412
+ {
1413
+ "epoch": 11.18808777429467,
1414
+ "grad_norm": 0.0004298978892620653,
1415
+ "learning_rate": 8.06387469704276e-05,
1416
+ "loss": 0.0,
1417
+ "step": 895
1418
+ },
1419
+ {
1420
+ "epoch": 11.25078369905956,
1421
+ "grad_norm": 0.005391178652644157,
1422
+ "learning_rate": 7.965439869473664e-05,
1423
+ "loss": 0.0,
1424
+ "step": 900
1425
+ },
1426
+ {
1427
+ "epoch": 11.31347962382445,
1428
+ "grad_norm": 0.00016025469813030213,
1429
+ "learning_rate": 7.867210268261439e-05,
1430
+ "loss": 0.0,
1431
+ "step": 905
1432
+ },
1433
+ {
1434
+ "epoch": 11.376175548589341,
1435
+ "grad_norm": 0.00021810800535604358,
1436
+ "learning_rate": 7.769195801839313e-05,
1437
+ "loss": 0.0,
1438
+ "step": 910
1439
+ },
1440
+ {
1441
+ "epoch": 11.438871473354231,
1442
+ "grad_norm": 0.0006552350823767483,
1443
+ "learning_rate": 7.671406356939836e-05,
1444
+ "loss": 0.0,
1445
+ "step": 915
1446
+ },
1447
+ {
1448
+ "epoch": 11.501567398119121,
1449
+ "grad_norm": 0.001337722409516573,
1450
+ "learning_rate": 7.573851797597602e-05,
1451
+ "loss": 0.0,
1452
+ "step": 920
1453
+ },
1454
+ {
1455
+ "epoch": 11.564263322884013,
1456
+ "grad_norm": 0.00711465161293745,
1457
+ "learning_rate": 7.476541964154269e-05,
1458
+ "loss": 0.0,
1459
+ "step": 925
1460
+ },
1461
+ {
1462
+ "epoch": 11.626959247648903,
1463
+ "grad_norm": 3.7509220419451594e-05,
1464
+ "learning_rate": 7.379486672265964e-05,
1465
+ "loss": 0.0,
1466
+ "step": 930
1467
+ },
1468
+ {
1469
+ "epoch": 11.689655172413794,
1470
+ "grad_norm": 0.00123989034909755,
1471
+ "learning_rate": 7.28269571191317e-05,
1472
+ "loss": 0.0,
1473
+ "step": 935
1474
+ },
1475
+ {
1476
+ "epoch": 11.752351097178684,
1477
+ "grad_norm": 0.00043695452040992677,
1478
+ "learning_rate": 7.186178846413214e-05,
1479
+ "loss": 0.0,
1480
+ "step": 940
1481
+ },
1482
+ {
1483
+ "epoch": 11.815047021943574,
1484
+ "grad_norm": 0.0020732777193188667,
1485
+ "learning_rate": 7.089945811435433e-05,
1486
+ "loss": 0.0,
1487
+ "step": 945
1488
+ },
1489
+ {
1490
+ "epoch": 11.877742946708464,
1491
+ "grad_norm": 0.0006961887702345848,
1492
+ "learning_rate": 6.994006314019141e-05,
1493
+ "loss": 0.0,
1494
+ "step": 950
1495
+ },
1496
+ {
1497
+ "epoch": 11.940438871473354,
1498
+ "grad_norm": 0.00014696276048198342,
1499
+ "learning_rate": 6.898370031594487e-05,
1500
+ "loss": 0.0,
1501
+ "step": 955
1502
+ },
1503
+ {
1504
+ "epoch": 12.0,
1505
+ "grad_norm": 0.004987122491002083,
1506
+ "learning_rate": 6.803046611006278e-05,
1507
+ "loss": 0.0,
1508
+ "step": 960
1509
+ },
1510
+ {
1511
+ "epoch": 12.0,
1512
+ "eval_loss": 2.9617700576782227,
1513
+ "eval_macro_f1": 74.48450982070439,
1514
+ "eval_macro_precision": 74.49341191282451,
1515
+ "eval_macro_recall": 74.47619047619047,
1516
+ "eval_micro_f1": 74.47619047619047,
1517
+ "eval_micro_precision": 74.47619047619047,
1518
+ "eval_micro_recall": 74.47619047619047,
1519
+ "eval_runtime": 3.4656,
1520
+ "eval_samples_per_second": 302.974,
1521
+ "eval_steps_per_second": 19.044,
1522
+ "step": 960
1523
+ },
1524
+ {
1525
+ "epoch": 12.06269592476489,
1526
+ "grad_norm": 3.519210804370232e-05,
1527
+ "learning_rate": 6.708045667540897e-05,
1528
+ "loss": 0.0,
1529
+ "step": 965
1530
+ },
1531
+ {
1532
+ "epoch": 12.12539184952978,
1533
+ "grad_norm": 0.0011250913375988603,
1534
+ "learning_rate": 6.613376783956423e-05,
1535
+ "loss": 0.0,
1536
+ "step": 970
1537
+ },
1538
+ {
1539
+ "epoch": 12.18808777429467,
1540
+ "grad_norm": 5.886521466891281e-05,
1541
+ "learning_rate": 6.519049509516013e-05,
1542
+ "loss": 0.0,
1543
+ "step": 975
1544
+ },
1545
+ {
1546
+ "epoch": 12.25078369905956,
1547
+ "grad_norm": 0.0004661143757402897,
1548
+ "learning_rate": 6.425073359024663e-05,
1549
+ "loss": 0.0,
1550
+ "step": 980
1551
+ },
1552
+ {
1553
+ "epoch": 12.31347962382445,
1554
+ "grad_norm": 0.00029624096350744367,
1555
+ "learning_rate": 6.331457811869437e-05,
1556
+ "loss": 0.0,
1557
+ "step": 985
1558
+ },
1559
+ {
1560
+ "epoch": 12.376175548589341,
1561
+ "grad_norm": 0.00012905469338875264,
1562
+ "learning_rate": 6.2382123110633e-05,
1563
+ "loss": 0.0,
1564
+ "step": 990
1565
+ },
1566
+ {
1567
+ "epoch": 12.438871473354231,
1568
+ "grad_norm": 0.00013497307372745126,
1569
+ "learning_rate": 6.145346262292595e-05,
1570
+ "loss": 0.0,
1571
+ "step": 995
1572
+ },
1573
+ {
1574
+ "epoch": 12.501567398119121,
1575
+ "grad_norm": 0.0008577414555475116,
1576
+ "learning_rate": 6.052869032968285e-05,
1577
+ "loss": 0.0,
1578
+ "step": 1000
1579
+ },
1580
+ {
1581
+ "epoch": 12.564263322884013,
1582
+ "grad_norm": 0.0001506131811765954,
1583
+ "learning_rate": 5.960789951281052e-05,
1584
+ "loss": 0.0,
1585
+ "step": 1005
1586
+ },
1587
+ {
1588
+ "epoch": 12.626959247648903,
1589
+ "grad_norm": 0.00014009448932483792,
1590
+ "learning_rate": 5.8691183052603834e-05,
1591
+ "loss": 0.0,
1592
+ "step": 1010
1593
+ },
1594
+ {
1595
+ "epoch": 12.689655172413794,
1596
+ "grad_norm": 2.588312781881541e-05,
1597
+ "learning_rate": 5.777863341837675e-05,
1598
+ "loss": 0.0,
1599
+ "step": 1015
1600
+ },
1601
+ {
1602
+ "epoch": 12.752351097178684,
1603
+ "grad_norm": 0.00011443781841080636,
1604
+ "learning_rate": 5.687034265913485e-05,
1605
+ "loss": 0.0,
1606
+ "step": 1020
1607
+ },
1608
+ {
1609
+ "epoch": 12.815047021943574,
1610
+ "grad_norm": 0.00010904014925472438,
1611
+ "learning_rate": 5.596640239429051e-05,
1612
+ "loss": 0.0,
1613
+ "step": 1025
1614
+ },
1615
+ {
1616
+ "epoch": 12.877742946708464,
1617
+ "grad_norm": 5.243328268988989e-05,
1618
+ "learning_rate": 5.5066903804421025e-05,
1619
+ "loss": 0.0,
1620
+ "step": 1030
1621
+ },
1622
+ {
1623
+ "epoch": 12.940438871473354,
1624
+ "grad_norm": 0.0001976548373932019,
1625
+ "learning_rate": 5.4171937622071435e-05,
1626
+ "loss": 0.0,
1627
+ "step": 1035
1628
+ },
1629
+ {
1630
+ "epoch": 13.0,
1631
+ "grad_norm": 1.4458280929829925e-05,
1632
+ "learning_rate": 5.32815941226022e-05,
1633
+ "loss": 0.0,
1634
+ "step": 1040
1635
+ },
1636
+ {
1637
+ "epoch": 13.0,
1638
+ "eval_loss": 2.9744179248809814,
1639
+ "eval_macro_f1": 74.495166799055,
1640
+ "eval_macro_precision": 74.51671347356447,
1641
+ "eval_macro_recall": 74.47619047619047,
1642
+ "eval_micro_f1": 74.47619047619047,
1643
+ "eval_micro_precision": 74.47619047619047,
1644
+ "eval_micro_recall": 74.47619047619047,
1645
+ "eval_runtime": 3.7357,
1646
+ "eval_samples_per_second": 281.069,
1647
+ "eval_steps_per_second": 17.667,
1648
+ "step": 1040
1649
+ },
1650
+ {
1651
+ "epoch": 13.06269592476489,
1652
+ "grad_norm": 0.00015813493519090116,
1653
+ "learning_rate": 5.2395963115083104e-05,
1654
+ "loss": 0.0,
1655
+ "step": 1045
1656
+ },
1657
+ {
1658
+ "epoch": 13.12539184952978,
1659
+ "grad_norm": 5.232647708908189e-06,
1660
+ "learning_rate": 5.151513393323426e-05,
1661
+ "loss": 0.0,
1662
+ "step": 1050
1663
+ },
1664
+ {
1665
+ "epoch": 13.18808777429467,
1666
+ "grad_norm": 0.0007055670721456409,
1667
+ "learning_rate": 5.06391954264149e-05,
1668
+ "loss": 0.0,
1669
+ "step": 1055
1670
+ },
1671
+ {
1672
+ "epoch": 13.25078369905956,
1673
+ "grad_norm": 0.0001715083053568378,
1674
+ "learning_rate": 4.976823595066128e-05,
1675
+ "loss": 0.0,
1676
+ "step": 1060
1677
+ },
1678
+ {
1679
+ "epoch": 13.31347962382445,
1680
+ "grad_norm": 0.0010762359015643597,
1681
+ "learning_rate": 4.8902343359774085e-05,
1682
+ "loss": 0.0,
1683
+ "step": 1065
1684
+ },
1685
+ {
1686
+ "epoch": 13.376175548589341,
1687
+ "grad_norm": 0.0003584644291549921,
1688
+ "learning_rate": 4.804160499645667e-05,
1689
+ "loss": 0.0,
1690
+ "step": 1070
1691
+ },
1692
+ {
1693
+ "epoch": 13.438871473354231,
1694
+ "grad_norm": 0.00013147966819815338,
1695
+ "learning_rate": 4.7186107683504656e-05,
1696
+ "loss": 0.0,
1697
+ "step": 1075
1698
+ },
1699
+ {
1700
+ "epoch": 13.501567398119121,
1701
+ "grad_norm": 0.00026007898850366473,
1702
+ "learning_rate": 4.6335937715048306e-05,
1703
+ "loss": 0.0,
1704
+ "step": 1080
1705
+ },
1706
+ {
1707
+ "epoch": 13.564263322884013,
1708
+ "grad_norm": 8.957670797826722e-05,
1709
+ "learning_rate": 4.549118084784788e-05,
1710
+ "loss": 0.0,
1711
+ "step": 1085
1712
+ },
1713
+ {
1714
+ "epoch": 13.626959247648903,
1715
+ "grad_norm": 7.390981409116648e-06,
1716
+ "learning_rate": 4.465192229264337e-05,
1717
+ "loss": 0.0,
1718
+ "step": 1090
1719
+ },
1720
+ {
1721
+ "epoch": 13.689655172413794,
1722
+ "grad_norm": 0.0010093646124005318,
1723
+ "learning_rate": 4.381824670555934e-05,
1724
+ "loss": 0.0,
1725
+ "step": 1095
1726
+ },
1727
+ {
1728
+ "epoch": 13.752351097178684,
1729
+ "grad_norm": 0.0001321593881584704,
1730
+ "learning_rate": 4.29902381795655e-05,
1731
+ "loss": 0.0,
1732
+ "step": 1100
1733
+ },
1734
+ {
1735
+ "epoch": 13.815047021943574,
1736
+ "grad_norm": 0.0003493933181744069,
1737
+ "learning_rate": 4.216798023599441e-05,
1738
+ "loss": 0.0,
1739
+ "step": 1105
1740
+ },
1741
+ {
1742
+ "epoch": 13.877742946708464,
1743
+ "grad_norm": 0.00017224319162778556,
1744
+ "learning_rate": 4.135155581611661e-05,
1745
+ "loss": 0.0,
1746
+ "step": 1110
1747
+ },
1748
+ {
1749
+ "epoch": 13.940438871473354,
1750
+ "grad_norm": 0.0007415884756483138,
1751
+ "learning_rate": 4.0541047272774315e-05,
1752
+ "loss": 0.0,
1753
+ "step": 1115
1754
+ },
1755
+ {
1756
+ "epoch": 14.0,
1757
+ "grad_norm": 0.00015337667718995363,
1758
+ "learning_rate": 3.973653636207437e-05,
1759
+ "loss": 0.0,
1760
+ "step": 1120
1761
+ },
1762
+ {
1763
+ "epoch": 14.0,
1764
+ "eval_loss": 2.978665351867676,
1765
+ "eval_macro_f1": 74.39737250761942,
1766
+ "eval_macro_precision": 74.41561381963534,
1767
+ "eval_macro_recall": 74.38095238095238,
1768
+ "eval_micro_f1": 74.38095238095238,
1769
+ "eval_micro_precision": 74.38095238095238,
1770
+ "eval_micro_recall": 74.38095238095238,
1771
+ "eval_runtime": 4.1748,
1772
+ "eval_samples_per_second": 251.508,
1773
+ "eval_steps_per_second": 15.809,
1774
+ "step": 1120
1775
+ },
1776
+ {
1777
+ "epoch": 14.06269592476489,
1778
+ "grad_norm": 0.00025141274090856314,
1779
+ "learning_rate": 3.893810423514172e-05,
1780
+ "loss": 0.0,
1781
+ "step": 1125
1782
+ },
1783
+ {
1784
+ "epoch": 14.12539184952978,
1785
+ "grad_norm": 0.0002585667825769633,
1786
+ "learning_rate": 3.814583142993352e-05,
1787
+ "loss": 0.0,
1788
+ "step": 1130
1789
+ },
1790
+ {
1791
+ "epoch": 14.18808777429467,
1792
+ "grad_norm": 1.575879105075728e-05,
1793
+ "learning_rate": 3.7359797863115283e-05,
1794
+ "loss": 0.0,
1795
+ "step": 1135
1796
+ },
1797
+ {
1798
+ "epoch": 14.25078369905956,
1799
+ "grad_norm": 3.0083991077844985e-05,
1800
+ "learning_rate": 3.6580082821999786e-05,
1801
+ "loss": 0.0,
1802
+ "step": 1140
1803
+ },
1804
+ {
1805
+ "epoch": 14.31347962382445,
1806
+ "grad_norm": 0.0006605722010135651,
1807
+ "learning_rate": 3.580676495654911e-05,
1808
+ "loss": 0.0,
1809
+ "step": 1145
1810
+ },
1811
+ {
1812
+ "epoch": 14.376175548589341,
1813
+ "grad_norm": 0.00018690273282118142,
1814
+ "learning_rate": 3.503992227144147e-05,
1815
+ "loss": 0.0,
1816
+ "step": 1150
1817
+ },
1818
+ {
1819
+ "epoch": 14.438871473354231,
1820
+ "grad_norm": 0.00042844025301747024,
1821
+ "learning_rate": 3.427963211820274e-05,
1822
+ "loss": 0.0,
1823
+ "step": 1155
1824
+ },
1825
+ {
1826
+ "epoch": 14.501567398119121,
1827
+ "grad_norm": 0.00033945144969038665,
1828
+ "learning_rate": 3.352597118740404e-05,
1829
+ "loss": 0.0,
1830
+ "step": 1160
1831
+ },
1832
+ {
1833
+ "epoch": 14.564263322884013,
1834
+ "grad_norm": 5.347471233108081e-05,
1835
+ "learning_rate": 3.277901550092581e-05,
1836
+ "loss": 0.0,
1837
+ "step": 1165
1838
+ },
1839
+ {
1840
+ "epoch": 14.626959247648903,
1841
+ "grad_norm": 0.0009476160048507154,
1842
+ "learning_rate": 3.2038840404289705e-05,
1843
+ "loss": 0.0,
1844
+ "step": 1170
1845
+ },
1846
+ {
1847
+ "epoch": 14.689655172413794,
1848
+ "grad_norm": 0.0006543208146467805,
1849
+ "learning_rate": 3.13055205590583e-05,
1850
+ "loss": 0.0,
1851
+ "step": 1175
1852
+ },
1853
+ {
1854
+ "epoch": 14.752351097178684,
1855
+ "grad_norm": 6.901170854689553e-05,
1856
+ "learning_rate": 3.0579129935304066e-05,
1857
+ "loss": 0.0,
1858
+ "step": 1180
1859
+ },
1860
+ {
1861
+ "epoch": 14.815047021943574,
1862
+ "grad_norm": 0.0006354337092489004,
1863
+ "learning_rate": 2.9859741804147957e-05,
1864
+ "loss": 0.0,
1865
+ "step": 1185
1866
+ },
1867
+ {
1868
+ "epoch": 14.877742946708464,
1869
+ "grad_norm": 0.0009218246559612453,
1870
+ "learning_rate": 2.9147428730368475e-05,
1871
+ "loss": 0.0,
1872
+ "step": 1190
1873
+ },
1874
+ {
1875
+ "epoch": 14.940438871473354,
1876
+ "grad_norm": 0.0008917151717469096,
1877
+ "learning_rate": 2.844226256508221e-05,
1878
+ "loss": 0.0,
1879
+ "step": 1195
1880
+ },
1881
+ {
1882
+ "epoch": 15.0,
1883
+ "grad_norm": 4.064366657985374e-05,
1884
+ "learning_rate": 2.7744314438496088e-05,
1885
+ "loss": 0.0,
1886
+ "step": 1200
1887
+ },
1888
+ {
1889
+ "epoch": 15.0,
1890
+ "eval_loss": 2.9775209426879883,
1891
+ "eval_macro_f1": 74.59288643616037,
1892
+ "eval_macro_precision": 74.61805414620287,
1893
+ "eval_macro_recall": 74.57142857142857,
1894
+ "eval_micro_f1": 74.57142857142857,
1895
+ "eval_micro_precision": 74.57142857142857,
1896
+ "eval_micro_recall": 74.57142857142857,
1897
+ "eval_runtime": 4.2475,
1898
+ "eval_samples_per_second": 247.204,
1899
+ "eval_steps_per_second": 15.539,
1900
+ "step": 1200
1901
+ },
1902
+ {
1903
+ "epoch": 15.06269592476489,
1904
+ "grad_norm": 0.0005438003572635353,
1905
+ "learning_rate": 2.7053654752732528e-05,
1906
+ "loss": 0.0,
1907
+ "step": 1205
1908
+ },
1909
+ {
1910
+ "epoch": 15.12539184952978,
1911
+ "grad_norm": 6.86753774061799e-05,
1912
+ "learning_rate": 2.6370353174727836e-05,
1913
+ "loss": 0.0,
1914
+ "step": 1210
1915
+ },
1916
+ {
1917
+ "epoch": 15.18808777429467,
1918
+ "grad_norm": 0.00032200937857851386,
1919
+ "learning_rate": 2.5694478629205078e-05,
1920
+ "loss": 0.0,
1921
+ "step": 1215
1922
+ },
1923
+ {
1924
+ "epoch": 15.25078369905956,
1925
+ "grad_norm": 0.0002235895226476714,
1926
+ "learning_rate": 2.5026099291721516e-05,
1927
+ "loss": 0.0,
1928
+ "step": 1220
1929
+ },
1930
+ {
1931
+ "epoch": 15.31347962382445,
1932
+ "grad_norm": 0.00034938243334181607,
1933
+ "learning_rate": 2.4365282581791782e-05,
1934
+ "loss": 0.0,
1935
+ "step": 1225
1936
+ },
1937
+ {
1938
+ "epoch": 15.376175548589341,
1939
+ "grad_norm": 0.00023688429791945964,
1940
+ "learning_rate": 2.371209515608718e-05,
1941
+ "loss": 0.0,
1942
+ "step": 1230
1943
+ },
1944
+ {
1945
+ "epoch": 15.438871473354231,
1946
+ "grad_norm": 0.0008163132588379085,
1947
+ "learning_rate": 2.3066602901712108e-05,
1948
+ "loss": 0.0,
1949
+ "step": 1235
1950
+ },
1951
+ {
1952
+ "epoch": 15.501567398119121,
1953
+ "grad_norm": 0.00019654417701531202,
1954
+ "learning_rate": 2.242887092955801e-05,
1955
+ "loss": 0.0,
1956
+ "step": 1240
1957
+ },
1958
+ {
1959
+ "epoch": 15.564263322884013,
1960
+ "grad_norm": 2.644214873726014e-05,
1961
+ "learning_rate": 2.1798963567735608e-05,
1962
+ "loss": 0.0,
1963
+ "step": 1245
1964
+ },
1965
+ {
1966
+ "epoch": 15.626959247648903,
1967
+ "grad_norm": 0.0005443547270260751,
1968
+ "learning_rate": 2.1176944355086058e-05,
1969
+ "loss": 0.0,
1970
+ "step": 1250
1971
+ },
1972
+ {
1973
+ "epoch": 15.689655172413794,
1974
+ "grad_norm": 0.00014729479153174907,
1975
+ "learning_rate": 2.0562876034771882e-05,
1976
+ "loss": 0.0,
1977
+ "step": 1255
1978
+ },
1979
+ {
1980
+ "epoch": 15.752351097178684,
1981
+ "grad_norm": 0.0004818035813514143,
1982
+ "learning_rate": 1.995682054794803e-05,
1983
+ "loss": 0.0,
1984
+ "step": 1260
1985
+ },
1986
+ {
1987
+ "epoch": 15.815047021943574,
1988
+ "grad_norm": 0.0002842153771780431,
1989
+ "learning_rate": 1.935883902751382e-05,
1990
+ "loss": 0.0,
1991
+ "step": 1265
1992
+ },
1993
+ {
1994
+ "epoch": 15.877742946708464,
1995
+ "grad_norm": 0.00016755808610469103,
1996
+ "learning_rate": 1.8768991791946456e-05,
1997
+ "loss": 0.0,
1998
+ "step": 1270
1999
+ },
2000
+ {
2001
+ "epoch": 15.940438871473354,
2002
+ "grad_norm": 0.0001340256567345932,
2003
+ "learning_rate": 1.8187338339216775e-05,
2004
+ "loss": 0.0,
2005
+ "step": 1275
2006
+ },
2007
+ {
2008
+ "epoch": 16.0,
2009
+ "grad_norm": 0.00031726626912131906,
2010
+ "learning_rate": 1.76139373407876e-05,
2011
+ "loss": 0.0,
2012
+ "step": 1280
2013
+ },
2014
+ {
2015
+ "epoch": 16.0,
2016
+ "eval_loss": 2.9834632873535156,
2017
+ "eval_macro_f1": 74.30509548505665,
2018
+ "eval_macro_precision": 74.3270405312326,
2019
+ "eval_macro_recall": 74.28571428571429,
2020
+ "eval_micro_f1": 74.28571428571429,
2021
+ "eval_micro_precision": 74.28571428571429,
2022
+ "eval_micro_recall": 74.28571428571429,
2023
+ "eval_runtime": 5.3043,
2024
+ "eval_samples_per_second": 197.954,
2025
+ "eval_steps_per_second": 12.443,
2026
+ "step": 1280
2027
+ },
2028
+ {
2029
+ "epoch": 16.062695924764892,
2030
+ "grad_norm": 0.0002320996136404574,
2031
+ "learning_rate": 1.7048846635695602e-05,
2032
+ "loss": 0.0,
2033
+ "step": 1285
2034
+ },
2035
+ {
2036
+ "epoch": 16.12539184952978,
2037
+ "grad_norm": 0.0005247556255199015,
2038
+ "learning_rate": 1.649212322471695e-05,
2039
+ "loss": 0.0,
2040
+ "step": 1290
2041
+ },
2042
+ {
2043
+ "epoch": 16.188087774294672,
2044
+ "grad_norm": 6.593632861040533e-05,
2045
+ "learning_rate": 1.5943823264617796e-05,
2046
+ "loss": 0.0,
2047
+ "step": 1295
2048
+ },
2049
+ {
2050
+ "epoch": 16.25078369905956,
2051
+ "grad_norm": 2.0959419998689555e-05,
2052
+ "learning_rate": 1.540400206248963e-05,
2053
+ "loss": 0.0,
2054
+ "step": 1300
2055
+ },
2056
+ {
2057
+ "epoch": 16.313479623824453,
2058
+ "grad_norm": 0.0002277921448694542,
2059
+ "learning_rate": 1.4872714070170468e-05,
2060
+ "loss": 0.0,
2061
+ "step": 1305
2062
+ },
2063
+ {
2064
+ "epoch": 16.37617554858934,
2065
+ "grad_norm": 5.7650511735118926e-05,
2066
+ "learning_rate": 1.435001287875234e-05,
2067
+ "loss": 0.0,
2068
+ "step": 1310
2069
+ },
2070
+ {
2071
+ "epoch": 16.438871473354233,
2072
+ "grad_norm": 0.0003665912081487477,
2073
+ "learning_rate": 1.3835951213175413e-05,
2074
+ "loss": 0.0,
2075
+ "step": 1315
2076
+ },
2077
+ {
2078
+ "epoch": 16.50156739811912,
2079
+ "grad_norm": 0.00018977168656419963,
2080
+ "learning_rate": 1.3330580926909763e-05,
2081
+ "loss": 0.0,
2082
+ "step": 1320
2083
+ },
2084
+ {
2085
+ "epoch": 16.564263322884013,
2086
+ "grad_norm": 2.3261072783498093e-05,
2087
+ "learning_rate": 1.2833952996724863e-05,
2088
+ "loss": 0.0,
2089
+ "step": 1325
2090
+ },
2091
+ {
2092
+ "epoch": 16.6269592476489,
2093
+ "grad_norm": 0.00015886298206169158,
2094
+ "learning_rate": 1.2346117517547551e-05,
2095
+ "loss": 0.0,
2096
+ "step": 1330
2097
+ },
2098
+ {
2099
+ "epoch": 16.689655172413794,
2100
+ "grad_norm": 0.00010217857925454155,
2101
+ "learning_rate": 1.1867123697408854e-05,
2102
+ "loss": 0.0,
2103
+ "step": 1335
2104
+ },
2105
+ {
2106
+ "epoch": 16.752351097178682,
2107
+ "grad_norm": 0.000697963812854141,
2108
+ "learning_rate": 1.139701985248055e-05,
2109
+ "loss": 0.0,
2110
+ "step": 1340
2111
+ },
2112
+ {
2113
+ "epoch": 16.815047021943574,
2114
+ "grad_norm": 0.0002654240815900266,
2115
+ "learning_rate": 1.0935853402201335e-05,
2116
+ "loss": 0.0,
2117
+ "step": 1345
2118
+ },
2119
+ {
2120
+ "epoch": 16.877742946708462,
2121
+ "grad_norm": 0.0005684046191163361,
2122
+ "learning_rate": 1.0483670864493778e-05,
2123
+ "loss": 0.0,
2124
+ "step": 1350
2125
+ },
2126
+ {
2127
+ "epoch": 16.940438871473354,
2128
+ "grad_norm": 0.00042359635699540377,
2129
+ "learning_rate": 1.004051785107184e-05,
2130
+ "loss": 0.0,
2131
+ "step": 1355
2132
+ },
2133
+ {
2134
+ "epoch": 17.0,
2135
+ "grad_norm": 8.337834879057482e-05,
2136
+ "learning_rate": 9.606439062840256e-06,
2137
+ "loss": 0.0,
2138
+ "step": 1360
2139
+ },
2140
+ {
2141
+ "epoch": 17.0,
2142
+ "eval_loss": 2.983590602874756,
2143
+ "eval_macro_f1": 74.48693497595121,
2144
+ "eval_macro_precision": 74.49848858038513,
2145
+ "eval_macro_recall": 74.47619047619048,
2146
+ "eval_micro_f1": 74.47619047619047,
2147
+ "eval_micro_precision": 74.47619047619047,
2148
+ "eval_micro_recall": 74.47619047619047,
2149
+ "eval_runtime": 5.1696,
2150
+ "eval_samples_per_second": 203.111,
2151
+ "eval_steps_per_second": 12.767,
2152
+ "step": 1360
2153
+ },
2154
+ {
2155
+ "epoch": 17.062695924764892,
2156
+ "grad_norm": 0.0003075774875469506,
2157
+ "learning_rate": 9.181478285385381e-06,
2158
+ "loss": 0.0,
2159
+ "step": 1365
2160
+ },
2161
+ {
2162
+ "epoch": 17.12539184952978,
2163
+ "grad_norm": 8.384697139263153e-05,
2164
+ "learning_rate": 8.765678384558607e-06,
2165
+ "loss": 0.0,
2166
+ "step": 1370
2167
+ },
2168
+ {
2169
+ "epoch": 17.188087774294672,
2170
+ "grad_norm": 0.00010231971828034148,
2171
+ "learning_rate": 8.359081302152394e-06,
2172
+ "loss": 0.0,
2173
+ "step": 1375
2174
+ },
2175
+ {
2176
+ "epoch": 17.25078369905956,
2177
+ "grad_norm": 0.0003627596015576273,
2178
+ "learning_rate": 7.961728051669737e-06,
2179
+ "loss": 0.0,
2180
+ "step": 1380
2181
+ },
2182
+ {
2183
+ "epoch": 17.313479623824453,
2184
+ "grad_norm": 3.662167000584304e-05,
2185
+ "learning_rate": 7.5736587141870155e-06,
2186
+ "loss": 0.0,
2187
+ "step": 1385
2188
+ },
2189
+ {
2190
+ "epoch": 17.37617554858934,
2191
+ "grad_norm": 0.0005837052594870329,
2192
+ "learning_rate": 7.194912434311052e-06,
2193
+ "loss": 0.0,
2194
+ "step": 1390
2195
+ },
2196
+ {
2197
+ "epoch": 17.438871473354233,
2198
+ "grad_norm": 7.577840005978942e-05,
2199
+ "learning_rate": 6.8255274162305374e-06,
2200
+ "loss": 0.0,
2201
+ "step": 1395
2202
+ },
2203
+ {
2204
+ "epoch": 17.50156739811912,
2205
+ "grad_norm": 0.00022036675363779068,
2206
+ "learning_rate": 6.465540919862456e-06,
2207
+ "loss": 0.0,
2208
+ "step": 1400
2209
+ },
2210
+ {
2211
+ "epoch": 17.564263322884013,
2212
+ "grad_norm": 7.557481876574457e-05,
2213
+ "learning_rate": 6.11498925709364e-06,
2214
+ "loss": 0.0,
2215
+ "step": 1405
2216
+ },
2217
+ {
2218
+ "epoch": 17.6269592476489,
2219
+ "grad_norm": 0.00023109870380721986,
2220
+ "learning_rate": 5.77390778811796e-06,
2221
+ "loss": 0.0,
2222
+ "step": 1410
2223
+ },
2224
+ {
2225
+ "epoch": 17.689655172413794,
2226
+ "grad_norm": 0.0002563658345025033,
2227
+ "learning_rate": 5.44233091786951e-06,
2228
+ "loss": 0.0,
2229
+ "step": 1415
2230
+ },
2231
+ {
2232
+ "epoch": 17.752351097178682,
2233
+ "grad_norm": 1.977133797481656e-05,
2234
+ "learning_rate": 5.12029209255227e-06,
2235
+ "loss": 0.0,
2236
+ "step": 1420
2237
+ },
2238
+ {
2239
+ "epoch": 17.815047021943574,
2240
+ "grad_norm": 0.00025862394249998033,
2241
+ "learning_rate": 4.807823796266331e-06,
2242
+ "loss": 0.0,
2243
+ "step": 1425
2244
+ },
2245
+ {
2246
+ "epoch": 17.877742946708462,
2247
+ "grad_norm": 0.00048169083311222494,
2248
+ "learning_rate": 4.504957547731214e-06,
2249
+ "loss": 0.0,
2250
+ "step": 1430
2251
+ },
2252
+ {
2253
+ "epoch": 17.940438871473354,
2254
+ "grad_norm": 1.255560255231103e-05,
2255
+ "learning_rate": 4.211723897106534e-06,
2256
+ "loss": 0.0,
2257
+ "step": 1435
2258
+ },
2259
+ {
2260
+ "epoch": 18.0,
2261
+ "grad_norm": 0.0004092851304449141,
2262
+ "learning_rate": 3.928152422910491e-06,
2263
+ "loss": 0.0,
2264
+ "step": 1440
2265
+ },
2266
+ {
2267
+ "epoch": 18.0,
2268
+ "eval_loss": 2.982109308242798,
2269
+ "eval_macro_f1": 74.49440308940996,
2270
+ "eval_macro_precision": 74.5153120821081,
2271
+ "eval_macro_recall": 74.47619047619047,
2272
+ "eval_micro_f1": 74.47619047619047,
2273
+ "eval_micro_precision": 74.47619047619047,
2274
+ "eval_micro_recall": 74.47619047619047,
2275
+ "eval_runtime": 5.0341,
2276
+ "eval_samples_per_second": 208.578,
2277
+ "eval_steps_per_second": 13.111,
2278
+ "step": 1440
2279
+ },
2280
+ {
2281
+ "epoch": 18.062695924764892,
2282
+ "grad_norm": 9.531569230603054e-05,
2283
+ "learning_rate": 3.6542717290362515e-06,
2284
+ "loss": 0.0,
2285
+ "step": 1445
2286
+ },
2287
+ {
2288
+ "epoch": 18.12539184952978,
2289
+ "grad_norm": 5.50920121895615e-05,
2290
+ "learning_rate": 3.390109441866618e-06,
2291
+ "loss": 0.0,
2292
+ "step": 1450
2293
+ },
2294
+ {
2295
+ "epoch": 18.188087774294672,
2296
+ "grad_norm": 0.00010016823216574267,
2297
+ "learning_rate": 3.135692207487373e-06,
2298
+ "loss": 0.0,
2299
+ "step": 1455
2300
+ },
2301
+ {
2302
+ "epoch": 18.25078369905956,
2303
+ "grad_norm": 0.0006045085028745234,
2304
+ "learning_rate": 2.8910456889995498e-06,
2305
+ "loss": 0.0,
2306
+ "step": 1460
2307
+ },
2308
+ {
2309
+ "epoch": 18.313479623824453,
2310
+ "grad_norm": 0.00010283043229719624,
2311
+ "learning_rate": 2.656194563930714e-06,
2312
+ "loss": 0.0,
2313
+ "step": 1465
2314
+ },
2315
+ {
2316
+ "epoch": 18.37617554858934,
2317
+ "grad_norm": 0.00047241951688192785,
2318
+ "learning_rate": 2.4311625217457778e-06,
2319
+ "loss": 0.0,
2320
+ "step": 1470
2321
+ },
2322
+ {
2323
+ "epoch": 18.438871473354233,
2324
+ "grad_norm": 0.0003394366940483451,
2325
+ "learning_rate": 2.2159722614573996e-06,
2326
+ "loss": 0.0,
2327
+ "step": 1475
2328
+ },
2329
+ {
2330
+ "epoch": 18.50156739811912,
2331
+ "grad_norm": 0.000773964449763298,
2332
+ "learning_rate": 2.010645489336382e-06,
2333
+ "loss": 0.0,
2334
+ "step": 1480
2335
+ },
2336
+ {
2337
+ "epoch": 18.564263322884013,
2338
+ "grad_norm": 6.701203528791666e-05,
2339
+ "learning_rate": 1.8152029167221475e-06,
2340
+ "loss": 0.0,
2341
+ "step": 1485
2342
+ },
2343
+ {
2344
+ "epoch": 18.6269592476489,
2345
+ "grad_norm": 0.0003426434122957289,
2346
+ "learning_rate": 1.6296642579335496e-06,
2347
+ "loss": 0.0,
2348
+ "step": 1490
2349
+ },
2350
+ {
2351
+ "epoch": 18.689655172413794,
2352
+ "grad_norm": 0.00010157287761103362,
2353
+ "learning_rate": 1.4540482282803137e-06,
2354
+ "loss": 0.0,
2355
+ "step": 1495
2356
+ },
2357
+ {
2358
+ "epoch": 18.752351097178682,
2359
+ "grad_norm": 0.0003385106392670423,
2360
+ "learning_rate": 1.2883725421752201e-06,
2361
+ "loss": 0.0,
2362
+ "step": 1500
2363
+ },
2364
+ {
2365
+ "epoch": 18.815047021943574,
2366
+ "grad_norm": 0.00031495324219577014,
2367
+ "learning_rate": 1.132653911347248e-06,
2368
+ "loss": 0.0,
2369
+ "step": 1505
2370
+ },
2371
+ {
2372
+ "epoch": 18.877742946708462,
2373
+ "grad_norm": 2.4275641408166848e-05,
2374
+ "learning_rate": 9.869080431558542e-07,
2375
+ "loss": 0.0,
2376
+ "step": 1510
2377
+ },
2378
+ {
2379
+ "epoch": 18.940438871473354,
2380
+ "grad_norm": 0.0007264981977641582,
2381
+ "learning_rate": 8.511496390065543e-07,
2382
+ "loss": 0.0,
2383
+ "step": 1515
2384
+ },
2385
+ {
2386
+ "epoch": 19.0,
2387
+ "grad_norm": 3.2049887522589415e-05,
2388
+ "learning_rate": 7.253923928680406e-07,
2389
+ "loss": 0.0,
2390
+ "step": 1520
2391
+ },
2392
+ {
2393
+ "epoch": 19.0,
2394
+ "eval_loss": 2.983579158782959,
2395
+ "eval_macro_f1": 74.59288643616037,
2396
+ "eval_macro_precision": 74.61805414620287,
2397
+ "eval_macro_recall": 74.57142857142857,
2398
+ "eval_micro_f1": 74.57142857142857,
2399
+ "eval_micro_precision": 74.57142857142857,
2400
+ "eval_micro_recall": 74.57142857142857,
2401
+ "eval_runtime": 3.5338,
2402
+ "eval_samples_per_second": 297.133,
2403
+ "eval_steps_per_second": 18.677,
2404
+ "step": 1520
2405
+ },
2406
+ {
2407
+ "epoch": 19.062695924764892,
2408
+ "grad_norm": 0.00015775053179822862,
2409
+ "learning_rate": 6.096489898908208e-07,
2410
+ "loss": 0.0,
2411
+ "step": 1525
2412
+ },
2413
+ {
2414
+ "epoch": 19.12539184952978,
2415
+ "grad_norm": 0.00020885077537968755,
2416
+ "learning_rate": 5.039311051276752e-07,
2417
+ "loss": 0.0,
2418
+ "step": 1530
2419
+ },
2420
+ {
2421
+ "epoch": 19.188087774294672,
2422
+ "grad_norm": 0.0004631892079487443,
2423
+ "learning_rate": 4.082494023560091e-07,
2424
+ "loss": 0.0,
2425
+ "step": 1535
2426
+ },
2427
+ {
2428
+ "epoch": 19.25078369905956,
2429
+ "grad_norm": 7.329711661441252e-05,
2430
+ "learning_rate": 3.2261353300219176e-07,
2431
+ "loss": 0.0,
2432
+ "step": 1540
2433
+ },
2434
+ {
2435
+ "epoch": 19.313479623824453,
2436
+ "grad_norm": 0.000103279686300084,
2437
+ "learning_rate": 2.4703213516799053e-07,
2438
+ "loss": 0.0,
2439
+ "step": 1545
2440
+ },
2441
+ {
2442
+ "epoch": 19.37617554858934,
2443
+ "grad_norm": 0.00035780860343948007,
2444
+ "learning_rate": 1.8151283275928964e-07,
2445
+ "loss": 0.0,
2446
+ "step": 1550
2447
+ },
2448
+ {
2449
+ "epoch": 19.438871473354233,
2450
+ "grad_norm": 0.0009104281198233366,
2451
+ "learning_rate": 1.2606223471702817e-07,
2452
+ "loss": 0.0,
2453
+ "step": 1555
2454
+ },
2455
+ {
2456
+ "epoch": 19.50156739811912,
2457
+ "grad_norm": 0.0008115767268463969,
2458
+ "learning_rate": 8.068593435055505e-08,
2459
+ "loss": 0.0,
2460
+ "step": 1560
2461
+ },
2462
+ {
2463
+ "epoch": 19.564263322884013,
2464
+ "grad_norm": 2.108391527144704e-05,
2465
+ "learning_rate": 4.5388508773469564e-08,
2466
+ "loss": 0.0,
2467
+ "step": 1565
2468
+ },
2469
+ {
2470
+ "epoch": 19.6269592476489,
2471
+ "grad_norm": 0.00026582309510558844,
2472
+ "learning_rate": 2.0173518441868324e-08,
2473
+ "loss": 0.0,
2474
+ "step": 1570
2475
+ },
2476
+ {
2477
+ "epoch": 19.689655172413794,
2478
+ "grad_norm": 0.0001829984103096649,
2479
+ "learning_rate": 5.043506795276987e-09,
2480
+ "loss": 0.0,
2481
+ "step": 1575
2482
+ },
2483
+ {
2484
+ "epoch": 19.752351097178682,
2485
+ "grad_norm": 0.00031997705809772015,
2486
+ "learning_rate": 0.0,
2487
+ "loss": 0.0,
2488
+ "step": 1580
2489
+ },
2490
+ {
2491
+ "epoch": 19.752351097178682,
2492
+ "eval_loss": 2.983431816101074,
2493
+ "eval_macro_f1": 74.495166799055,
2494
+ "eval_macro_precision": 74.51671347356447,
2495
+ "eval_macro_recall": 74.47619047619047,
2496
+ "eval_micro_f1": 74.47619047619047,
2497
+ "eval_micro_precision": 74.47619047619047,
2498
+ "eval_micro_recall": 74.47619047619047,
2499
+ "eval_runtime": 5.0918,
2500
+ "eval_samples_per_second": 206.215,
2501
+ "eval_steps_per_second": 12.962,
2502
+ "step": 1580
2503
+ },
2504
+ {
2505
+ "epoch": 19.752351097178682,
2506
+ "step": 1580,
2507
+ "total_flos": 8695327329615872.0,
2508
+ "train_loss": 0.2892922113949427,
2509
+ "train_runtime": 830.7771,
2510
+ "train_samples_per_second": 122.777,
2511
+ "train_steps_per_second": 1.902
2512
+ }
2513
+ ],
2514
+ "logging_steps": 5,
2515
+ "max_steps": 1580,
2516
+ "num_input_tokens_seen": 0,
2517
+ "num_train_epochs": 20,
2518
+ "save_steps": 50.0,
2519
+ "stateful_callbacks": {
2520
+ "TrainerControl": {
2521
+ "args": {
2522
+ "should_epoch_stop": false,
2523
+ "should_evaluate": false,
2524
+ "should_log": false,
2525
+ "should_save": true,
2526
+ "should_training_stop": true
2527
+ },
2528
+ "attributes": {}
2529
+ }
2530
+ },
2531
+ "total_flos": 8695327329615872.0,
2532
+ "train_batch_size": 16,
2533
+ "trial_name": null,
2534
+ "trial_params": null
2535
+ }
training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b08767f894ce7bf64160856bec942bbf746f1db02b5ef253c80e5a909f773cb7
3
+ size 6840