jakelever commited on
Commit
673ee94
·
verified ·
1 Parent(s): 2dd2dda

Upload folder using huggingface_hub

Browse files
README.md ADDED
@@ -0,0 +1,75 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ task: token-classification
3
+ tags:
4
+ - biomedical
5
+ - bionlp
6
+ license: mit
7
+ base_model: microsoft/BiomedNLP-BiomedBERT-base-uncased-abstract-fulltext
8
+ ---
9
+
10
+ # bioner_tmvar3
11
+
12
+ This is a named entity recognition model fine-tuned from the [microsoft/BiomedNLP-BiomedBERT-base-uncased-abstract-fulltext](https://huggingface.co/microsoft/BiomedNLP-BiomedBERT-base-uncased-abstract-fulltext) model. It predicts spans with 10 possible labels. The labels are **AcidChange, CellLine, DNAAllele, DNAMutation, Gene, OtherMutation, ProteinAllele, ProteinMutation, SNP and Species**.
13
+
14
+ The code used for training this model can be found at https://github.com/Glasgow-AI4BioMed/bioner along with links to other biomedical NER models trained on well-known biomedical corpora. The source dataset information is below.
15
+
16
+ ## Example Usage
17
+
18
+ The code below will load up the model and apply it to the provided text. It uses a simple aggregation strategy to post-process the individual tokens into larger multi-token entities where needed.
19
+
20
+ ```python
21
+ from transformers import pipeline
22
+
23
+ # Load the model as part of an NER pipeline
24
+ ner_pipeline = pipeline("token-classification",
25
+ model="Glasgow-AI4BioMed/bioner_tmvar3",
26
+ aggregation_strategy="max")
27
+
28
+ # Apply it to some text
29
+ ner_pipeline("EGFR T790M mutations have been known to affect treatment outcomes for NSCLC patients receiving erlotinib.")
30
+
31
+ # Output:
32
+ # [ {"entity_group": "Gene", "score": 0.78493, "word": "egfr", "start": 0, "end": 4},
33
+ # {"entity_group": "Species", "score": 0.99795, "word": "patients", "start": 57, "end": 65} ]
34
+ ```
35
+
36
+ ## Dataset Info
37
+
38
+ **Source:** The tmVar3 dataset was downloaded from: https://ftp.ncbi.nlm.nih.gov/pub/lu/tmVar3/
39
+
40
+ The dataset should be cited with: Wei, Chih-Hsuan, et al. "tmVar 3.0: an improved variant concept recognition and normalization tool." Bioinformatics 38.18 (2022): 4449-4451. DOI: [10.1093/bioinformatics/btac537](https://doi.org/10.1093/bioinformatics/btac537)
41
+
42
+ **Preprocessing:** The training/validation/test split was maintained from the original dataset. No changes were made to the annotations. The preprocessing script for this dataset is [prepare_tmvar.py](https://github.com/Glasgow-AI4BioMed/bioner/blob/main/prepare_tmvar.py).
43
+
44
+ ## Performance
45
+
46
+ The span-level performance on the test split for the different labels are shown in the tables below. The full performance results are available in the model repo in Markdown format for viewing and JSON format for easier loading. These include the performance at token level (with individual B- and I- labels as the token classifier uses IOB2 token labelling).
47
+
48
+ | Label | Precision | Recall | F1-score | Support |
49
+ | --- | --- | --- | --- | --- |
50
+ | AcidChange | 0.562 | 0.818 | 0.667 | 11 |
51
+ | CellLine | 0.550 | 0.688 | 0.611 | 16 |
52
+ | DNAAllele | 0.786 | 0.733 | 0.759 | 15 |
53
+ | DNAMutation | 0.824 | 0.864 | 0.844 | 125 |
54
+ | Gene | 0.886 | 0.907 | 0.897 | 766 |
55
+ | OtherMutation | 0.410 | 0.561 | 0.474 | 57 |
56
+ | ProteinAllele | 0.857 | 0.706 | 0.774 | 17 |
57
+ | ProteinMutation | 0.904 | 0.851 | 0.877 | 121 |
58
+ | SNP | 0.815 | 1.000 | 0.898 | 22 |
59
+ | Species | 0.955 | 0.949 | 0.952 | 333 |
60
+ | macro avg | 0.755 | 0.808 | 0.775 | 1483 |
61
+ | weighted avg | 0.871 | 0.889 | 0.879 | 1483 |
62
+
63
+
64
+ ## Hyperparameters
65
+
66
+ Hyperparameter tuning was done with [optuna](https://optuna.org/) and the [hyperparameter_search](https://huggingface.co/docs/transformers/en/hpo_train) functionality. 100 trials were run. Early stopping was applied during training. The best performing model was selected using the macro F1 performance on the validation set. The selected hyperparameters are in the table below.
67
+
68
+ | Hyperparameter | Value |
69
+ |----------------|-------|
70
+ | epochs | 17.0 |
71
+ | learning_rate | 9.821730986128684e-05 |
72
+ | per_device_train_batch_size | 16 |
73
+ | weight_decay | 0.2819767106649835 |
74
+ | warmup_ratio | 0.10342300416492177 |
75
+
best_hyperparameters.json ADDED
@@ -0,0 +1,7 @@
 
 
 
 
 
 
 
 
1
+ {
2
+ "epochs": 17.0,
3
+ "learning_rate": 9.821730986128684e-05,
4
+ "per_device_train_batch_size": 16,
5
+ "weight_decay": 0.2819767106649835,
6
+ "warmup_ratio": 0.10342300416492177
7
+ }
config.json ADDED
@@ -0,0 +1,48 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "microsoft/BiomedNLP-BiomedBERT-base-uncased-abstract-fulltext",
3
+ "architectures": [
4
+ "BertForTokenClassification"
5
+ ],
6
+ "attention_probs_dropout_prob": 0.1,
7
+ "classifier_dropout": null,
8
+ "hidden_act": "gelu",
9
+ "hidden_dropout_prob": 0.1,
10
+ "hidden_size": 768,
11
+ "id2label": {
12
+ "0": "O",
13
+ "1": "B-AcidChange",
14
+ "2": "I-AcidChange",
15
+ "3": "B-CellLine",
16
+ "4": "I-CellLine",
17
+ "5": "B-DNAAllele",
18
+ "6": "I-DNAAllele",
19
+ "7": "B-DNAMutation",
20
+ "8": "I-DNAMutation",
21
+ "9": "B-Gene",
22
+ "10": "I-Gene",
23
+ "11": "B-OtherMutation",
24
+ "12": "I-OtherMutation",
25
+ "13": "B-ProteinAllele",
26
+ "14": "I-ProteinAllele",
27
+ "15": "B-ProteinMutation",
28
+ "16": "I-ProteinMutation",
29
+ "17": "B-SNP",
30
+ "18": "I-SNP",
31
+ "19": "B-Species",
32
+ "20": "I-Species"
33
+ },
34
+ "initializer_range": 0.02,
35
+ "intermediate_size": 3072,
36
+ "layer_norm_eps": 1e-12,
37
+ "max_position_embeddings": 512,
38
+ "model_type": "bert",
39
+ "num_attention_heads": 12,
40
+ "num_hidden_layers": 12,
41
+ "pad_token_id": 0,
42
+ "position_embedding_type": "absolute",
43
+ "torch_dtype": "float32",
44
+ "transformers_version": "4.48.1",
45
+ "type_vocab_size": 2,
46
+ "use_cache": true,
47
+ "vocab_size": 30522
48
+ }
model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8801c9be6f28c8c6cd2db3ddc727b5ba68f85efb12389ca3c6c562e7b44da179
3
+ size 435654532
performance_report.json ADDED
@@ -0,0 +1,653 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "train": {
3
+ "token_level": {
4
+ "O": {
5
+ "precision": 0.9994677158065263,
6
+ "recall": 0.9999884225759769,
7
+ "f1-score": 0.999728001388929,
8
+ "support": 86375.0
9
+ },
10
+ "B-AcidChange": {
11
+ "precision": 1.0,
12
+ "recall": 1.0,
13
+ "f1-score": 1.0,
14
+ "support": 57.0
15
+ },
16
+ "I-AcidChange": {
17
+ "precision": 1.0,
18
+ "recall": 1.0,
19
+ "f1-score": 1.0,
20
+ "support": 170.0
21
+ },
22
+ "B-CellLine": {
23
+ "precision": 1.0,
24
+ "recall": 1.0,
25
+ "f1-score": 1.0,
26
+ "support": 23.0
27
+ },
28
+ "I-CellLine": {
29
+ "precision": 1.0,
30
+ "recall": 1.0,
31
+ "f1-score": 1.0,
32
+ "support": 33.0
33
+ },
34
+ "B-DNAAllele": {
35
+ "precision": 1.0,
36
+ "recall": 1.0,
37
+ "f1-score": 1.0,
38
+ "support": 35.0
39
+ },
40
+ "I-DNAAllele": {
41
+ "precision": 1.0,
42
+ "recall": 1.0,
43
+ "f1-score": 1.0,
44
+ "support": 92.0
45
+ },
46
+ "B-DNAMutation": {
47
+ "precision": 1.0,
48
+ "recall": 1.0,
49
+ "f1-score": 1.0,
50
+ "support": 391.0
51
+ },
52
+ "I-DNAMutation": {
53
+ "precision": 1.0,
54
+ "recall": 1.0,
55
+ "f1-score": 1.0,
56
+ "support": 2067.0
57
+ },
58
+ "B-Gene": {
59
+ "precision": 1.0,
60
+ "recall": 0.9983799108950993,
61
+ "f1-score": 0.999189298743413,
62
+ "support": 2469.0
63
+ },
64
+ "I-Gene": {
65
+ "precision": 0.9994079336885732,
66
+ "recall": 0.9888693614528412,
67
+ "f1-score": 0.9941107184923439,
68
+ "support": 3414.0
69
+ },
70
+ "B-OtherMutation": {
71
+ "precision": 1.0,
72
+ "recall": 1.0,
73
+ "f1-score": 1.0,
74
+ "support": 147.0
75
+ },
76
+ "I-OtherMutation": {
77
+ "precision": 1.0,
78
+ "recall": 1.0,
79
+ "f1-score": 1.0,
80
+ "support": 508.0
81
+ },
82
+ "B-ProteinAllele": {
83
+ "precision": 1.0,
84
+ "recall": 1.0,
85
+ "f1-score": 1.0,
86
+ "support": 43.0
87
+ },
88
+ "I-ProteinAllele": {
89
+ "precision": 1.0,
90
+ "recall": 1.0,
91
+ "f1-score": 1.0,
92
+ "support": 92.0
93
+ },
94
+ "B-ProteinMutation": {
95
+ "precision": 1.0,
96
+ "recall": 1.0,
97
+ "f1-score": 1.0,
98
+ "support": 386.0
99
+ },
100
+ "I-ProteinMutation": {
101
+ "precision": 1.0,
102
+ "recall": 0.9992603550295858,
103
+ "f1-score": 0.9996300406955235,
104
+ "support": 1352.0
105
+ },
106
+ "B-SNP": {
107
+ "precision": 1.0,
108
+ "recall": 1.0,
109
+ "f1-score": 1.0,
110
+ "support": 96.0
111
+ },
112
+ "I-SNP": {
113
+ "precision": 1.0,
114
+ "recall": 1.0,
115
+ "f1-score": 1.0,
116
+ "support": 279.0
117
+ },
118
+ "B-Species": {
119
+ "precision": 0.9990610328638497,
120
+ "recall": 0.99812382739212,
121
+ "f1-score": 0.998592210229939,
122
+ "support": 1066.0
123
+ },
124
+ "I-Species": {
125
+ "precision": 1.0,
126
+ "recall": 0.9863636363636363,
127
+ "f1-score": 0.9931350114416476,
128
+ "support": 220.0
129
+ },
130
+ "accuracy": 0.9995066203493933,
131
+ "macro avg": {
132
+ "precision": 0.9999017467789975,
133
+ "recall": 0.998618357795679,
134
+ "f1-score": 0.9992564419519904,
135
+ "support": 99315.0
136
+ },
137
+ "weighted avg": {
138
+ "precision": 0.9995066374609511,
139
+ "recall": 0.9995066203493933,
140
+ "f1-score": 0.9995054856279689,
141
+ "support": 99315.0
142
+ }
143
+ },
144
+ "span_level": {
145
+ "AcidChange": {
146
+ "precision": 1.0,
147
+ "recall": 1.0,
148
+ "f1-score": 1.0,
149
+ "support": 57
150
+ },
151
+ "CellLine": {
152
+ "precision": 1.0,
153
+ "recall": 1.0,
154
+ "f1-score": 1.0,
155
+ "support": 23
156
+ },
157
+ "DNAAllele": {
158
+ "precision": 1.0,
159
+ "recall": 1.0,
160
+ "f1-score": 1.0,
161
+ "support": 35
162
+ },
163
+ "DNAMutation": {
164
+ "precision": 0.9974424552429667,
165
+ "recall": 0.9974424552429667,
166
+ "f1-score": 0.9974424552429667,
167
+ "support": 391
168
+ },
169
+ "Gene": {
170
+ "precision": 0.9959349593495935,
171
+ "recall": 0.9863123993558777,
172
+ "f1-score": 0.9911003236245955,
173
+ "support": 2484
174
+ },
175
+ "OtherMutation": {
176
+ "precision": 0.9716312056737588,
177
+ "recall": 0.9319727891156463,
178
+ "f1-score": 0.951388888888889,
179
+ "support": 147
180
+ },
181
+ "ProteinAllele": {
182
+ "precision": 1.0,
183
+ "recall": 1.0,
184
+ "f1-score": 1.0,
185
+ "support": 43
186
+ },
187
+ "ProteinMutation": {
188
+ "precision": 0.9974093264248705,
189
+ "recall": 0.9771573604060914,
190
+ "f1-score": 0.9871794871794872,
191
+ "support": 394
192
+ },
193
+ "SNP": {
194
+ "precision": 1.0,
195
+ "recall": 1.0,
196
+ "f1-score": 1.0,
197
+ "support": 96
198
+ },
199
+ "Species": {
200
+ "precision": 0.9990539262062441,
201
+ "recall": 0.9841565703634669,
202
+ "f1-score": 0.9915492957746479,
203
+ "support": 1073
204
+ },
205
+ "macro avg": {
206
+ "precision": 0.9961471872897434,
207
+ "recall": 0.987704157448405,
208
+ "f1-score": 0.9918660450710588,
209
+ "support": 4743
210
+ },
211
+ "weighted avg": {
212
+ "precision": 0.9963517528334667,
213
+ "recall": 0.9850305713683323,
214
+ "f1-score": 0.990644841422112,
215
+ "support": 4743
216
+ }
217
+ }
218
+ },
219
+ "val": {
220
+ "token_level": {
221
+ "O": {
222
+ "precision": 0.9925305486058232,
223
+ "recall": 0.9874107207272202,
224
+ "f1-score": 0.9899640151172673,
225
+ "support": 27722.0
226
+ },
227
+ "B-AcidChange": {
228
+ "precision": 0.76,
229
+ "recall": 0.76,
230
+ "f1-score": 0.76,
231
+ "support": 25.0
232
+ },
233
+ "I-AcidChange": {
234
+ "precision": 0.6944444444444444,
235
+ "recall": 0.7575757575757576,
236
+ "f1-score": 0.7246376811594203,
237
+ "support": 66.0
238
+ },
239
+ "B-CellLine": {
240
+ "precision": 0.42105263157894735,
241
+ "recall": 1.0,
242
+ "f1-score": 0.5925925925925926,
243
+ "support": 8.0
244
+ },
245
+ "I-CellLine": {
246
+ "precision": 0.7142857142857143,
247
+ "recall": 1.0,
248
+ "f1-score": 0.8333333333333334,
249
+ "support": 10.0
250
+ },
251
+ "B-DNAAllele": {
252
+ "precision": 0.9285714285714286,
253
+ "recall": 0.7222222222222222,
254
+ "f1-score": 0.8125,
255
+ "support": 18.0
256
+ },
257
+ "I-DNAAllele": {
258
+ "precision": 0.9,
259
+ "recall": 0.6923076923076923,
260
+ "f1-score": 0.782608695652174,
261
+ "support": 52.0
262
+ },
263
+ "B-DNAMutation": {
264
+ "precision": 0.9016393442622951,
265
+ "recall": 0.9322033898305084,
266
+ "f1-score": 0.9166666666666666,
267
+ "support": 118.0
268
+ },
269
+ "I-DNAMutation": {
270
+ "precision": 0.9351701782820098,
271
+ "recall": 0.8560830860534124,
272
+ "f1-score": 0.8938807126258714,
273
+ "support": 674.0
274
+ },
275
+ "B-Gene": {
276
+ "precision": 0.9036004645760743,
277
+ "recall": 0.9688667496886675,
278
+ "f1-score": 0.9350961538461539,
279
+ "support": 803.0
280
+ },
281
+ "I-Gene": {
282
+ "precision": 0.8725490196078431,
283
+ "recall": 0.9541910331384016,
284
+ "f1-score": 0.9115456238361266,
285
+ "support": 1026.0
286
+ },
287
+ "B-OtherMutation": {
288
+ "precision": 0.6060606060606061,
289
+ "recall": 0.7407407407407407,
290
+ "f1-score": 0.6666666666666666,
291
+ "support": 54.0
292
+ },
293
+ "I-OtherMutation": {
294
+ "precision": 0.6,
295
+ "recall": 0.6026200873362445,
296
+ "f1-score": 0.6013071895424836,
297
+ "support": 229.0
298
+ },
299
+ "B-ProteinAllele": {
300
+ "precision": 0.4166666666666667,
301
+ "recall": 0.7142857142857143,
302
+ "f1-score": 0.5263157894736842,
303
+ "support": 7.0
304
+ },
305
+ "I-ProteinAllele": {
306
+ "precision": 0.2777777777777778,
307
+ "recall": 0.7692307692307693,
308
+ "f1-score": 0.40816326530612246,
309
+ "support": 13.0
310
+ },
311
+ "B-ProteinMutation": {
312
+ "precision": 0.9166666666666666,
313
+ "recall": 0.9211822660098522,
314
+ "f1-score": 0.918918918918919,
315
+ "support": 203.0
316
+ },
317
+ "I-ProteinMutation": {
318
+ "precision": 0.9349470499243571,
319
+ "recall": 0.9182763744427934,
320
+ "f1-score": 0.9265367316341829,
321
+ "support": 673.0
322
+ },
323
+ "B-SNP": {
324
+ "precision": 1.0,
325
+ "recall": 1.0,
326
+ "f1-score": 1.0,
327
+ "support": 17.0
328
+ },
329
+ "I-SNP": {
330
+ "precision": 0.9622641509433962,
331
+ "recall": 1.0,
332
+ "f1-score": 0.9807692307692307,
333
+ "support": 51.0
334
+ },
335
+ "B-Species": {
336
+ "precision": 0.96875,
337
+ "recall": 0.9855491329479769,
338
+ "f1-score": 0.9770773638968482,
339
+ "support": 346.0
340
+ },
341
+ "I-Species": {
342
+ "precision": 0.9375,
343
+ "recall": 0.9090909090909091,
344
+ "f1-score": 0.9230769230769231,
345
+ "support": 33.0
346
+ },
347
+ "accuracy": 0.976421550329725,
348
+ "macro avg": {
349
+ "precision": 0.7925941282025739,
350
+ "recall": 0.8662779355061373,
351
+ "f1-score": 0.8134122644816508,
352
+ "support": 32148.0
353
+ },
354
+ "weighted avg": {
355
+ "precision": 0.9778360580996203,
356
+ "recall": 0.976421550329725,
357
+ "f1-score": 0.9768605358437904,
358
+ "support": 32148.0
359
+ }
360
+ },
361
+ "span_level": {
362
+ "AcidChange": {
363
+ "precision": 0.6551724137931034,
364
+ "recall": 0.76,
365
+ "f1-score": 0.7037037037037037,
366
+ "support": 25
367
+ },
368
+ "CellLine": {
369
+ "precision": 0.42105263157894735,
370
+ "recall": 1.0,
371
+ "f1-score": 0.5925925925925926,
372
+ "support": 8
373
+ },
374
+ "DNAAllele": {
375
+ "precision": 0.9411764705882353,
376
+ "recall": 0.8421052631578947,
377
+ "f1-score": 0.8888888888888888,
378
+ "support": 19
379
+ },
380
+ "DNAMutation": {
381
+ "precision": 0.8442622950819673,
382
+ "recall": 0.8728813559322034,
383
+ "f1-score": 0.8583333333333334,
384
+ "support": 118
385
+ },
386
+ "Gene": {
387
+ "precision": 0.8688147295742232,
388
+ "recall": 0.9332509270704573,
389
+ "f1-score": 0.899880810488677,
390
+ "support": 809
391
+ },
392
+ "OtherMutation": {
393
+ "precision": 0.4931506849315068,
394
+ "recall": 0.6666666666666666,
395
+ "f1-score": 0.5669291338582677,
396
+ "support": 54
397
+ },
398
+ "ProteinAllele": {
399
+ "precision": 0.26666666666666666,
400
+ "recall": 0.5714285714285714,
401
+ "f1-score": 0.36363636363636365,
402
+ "support": 7
403
+ },
404
+ "ProteinMutation": {
405
+ "precision": 0.84688995215311,
406
+ "recall": 0.8676470588235294,
407
+ "f1-score": 0.8571428571428571,
408
+ "support": 204
409
+ },
410
+ "SNP": {
411
+ "precision": 1.0,
412
+ "recall": 1.0,
413
+ "f1-score": 1.0,
414
+ "support": 17
415
+ },
416
+ "Species": {
417
+ "precision": 0.9715909090909091,
418
+ "recall": 0.9771428571428571,
419
+ "f1-score": 0.9743589743589745,
420
+ "support": 350
421
+ },
422
+ "macro avg": {
423
+ "precision": 0.730877675345867,
424
+ "recall": 0.849112270022218,
425
+ "f1-score": 0.7705466658003658,
426
+ "support": 1611
427
+ },
428
+ "weighted avg": {
429
+ "precision": 0.8680591703649133,
430
+ "recall": 0.9168218497827436,
431
+ "f1-score": 0.8904727143994291,
432
+ "support": 1611
433
+ }
434
+ }
435
+ },
436
+ "test": {
437
+ "token_level": {
438
+ "O": {
439
+ "precision": 0.992110521601323,
440
+ "recall": 0.991290877796902,
441
+ "f1-score": 0.9917005303395551,
442
+ "support": 29050.0
443
+ },
444
+ "B-AcidChange": {
445
+ "precision": 0.6153846153846154,
446
+ "recall": 0.7272727272727273,
447
+ "f1-score": 0.6666666666666666,
448
+ "support": 11.0
449
+ },
450
+ "I-AcidChange": {
451
+ "precision": 0.6170212765957447,
452
+ "recall": 0.8529411764705882,
453
+ "f1-score": 0.7160493827160493,
454
+ "support": 34.0
455
+ },
456
+ "B-CellLine": {
457
+ "precision": 0.8888888888888888,
458
+ "recall": 1.0,
459
+ "f1-score": 0.9411764705882353,
460
+ "support": 16.0
461
+ },
462
+ "I-CellLine": {
463
+ "precision": 0.8888888888888888,
464
+ "recall": 0.6666666666666666,
465
+ "f1-score": 0.7619047619047619,
466
+ "support": 36.0
467
+ },
468
+ "B-DNAAllele": {
469
+ "precision": 1.0,
470
+ "recall": 0.8,
471
+ "f1-score": 0.8888888888888888,
472
+ "support": 15.0
473
+ },
474
+ "I-DNAAllele": {
475
+ "precision": 0.926829268292683,
476
+ "recall": 0.7755102040816326,
477
+ "f1-score": 0.8444444444444444,
478
+ "support": 49.0
479
+ },
480
+ "B-DNAMutation": {
481
+ "precision": 0.8818897637795275,
482
+ "recall": 0.896,
483
+ "f1-score": 0.8888888888888888,
484
+ "support": 125.0
485
+ },
486
+ "I-DNAMutation": {
487
+ "precision": 0.943089430894309,
488
+ "recall": 0.8592592592592593,
489
+ "f1-score": 0.8992248062015504,
490
+ "support": 675.0
491
+ },
492
+ "B-Gene": {
493
+ "precision": 0.9209183673469388,
494
+ "recall": 0.9425587467362925,
495
+ "f1-score": 0.9316129032258065,
496
+ "support": 766.0
497
+ },
498
+ "I-Gene": {
499
+ "precision": 0.9148936170212766,
500
+ "recall": 0.9265426052889324,
501
+ "f1-score": 0.9206812652068127,
502
+ "support": 1021.0
503
+ },
504
+ "B-OtherMutation": {
505
+ "precision": 0.5487804878048781,
506
+ "recall": 0.7894736842105263,
507
+ "f1-score": 0.6474820143884892,
508
+ "support": 57.0
509
+ },
510
+ "I-OtherMutation": {
511
+ "precision": 0.6742081447963801,
512
+ "recall": 0.764102564102564,
513
+ "f1-score": 0.7163461538461539,
514
+ "support": 195.0
515
+ },
516
+ "B-ProteinAllele": {
517
+ "precision": 1.0,
518
+ "recall": 0.7058823529411765,
519
+ "f1-score": 0.8275862068965517,
520
+ "support": 17.0
521
+ },
522
+ "I-ProteinAllele": {
523
+ "precision": 0.8571428571428571,
524
+ "recall": 0.75,
525
+ "f1-score": 0.8,
526
+ "support": 32.0
527
+ },
528
+ "B-ProteinMutation": {
529
+ "precision": 0.9482758620689655,
530
+ "recall": 0.9090909090909091,
531
+ "f1-score": 0.9282700421940928,
532
+ "support": 121.0
533
+ },
534
+ "I-ProteinMutation": {
535
+ "precision": 0.9340659340659341,
536
+ "recall": 0.96045197740113,
537
+ "f1-score": 0.947075208913649,
538
+ "support": 354.0
539
+ },
540
+ "B-SNP": {
541
+ "precision": 0.9166666666666666,
542
+ "recall": 1.0,
543
+ "f1-score": 0.9565217391304348,
544
+ "support": 22.0
545
+ },
546
+ "I-SNP": {
547
+ "precision": 0.8266666666666667,
548
+ "recall": 1.0,
549
+ "f1-score": 0.9051094890510949,
550
+ "support": 62.0
551
+ },
552
+ "B-Species": {
553
+ "precision": 0.9723926380368099,
554
+ "recall": 0.9519519519519519,
555
+ "f1-score": 0.9620637329286799,
556
+ "support": 333.0
557
+ },
558
+ "I-Species": {
559
+ "precision": 0.8409090909090909,
560
+ "recall": 0.8222222222222222,
561
+ "f1-score": 0.8314606741573034,
562
+ "support": 45.0
563
+ },
564
+ "accuracy": 0.9808088146264681,
565
+ "macro avg": {
566
+ "precision": 0.8623344279453543,
567
+ "recall": 0.8614865678806418,
568
+ "f1-score": 0.8558644890751479,
569
+ "support": 33036.0
570
+ },
571
+ "weighted avg": {
572
+ "precision": 0.9815714652726921,
573
+ "recall": 0.9808088146264681,
574
+ "f1-score": 0.9809991613726176,
575
+ "support": 33036.0
576
+ }
577
+ },
578
+ "span_level": {
579
+ "AcidChange": {
580
+ "precision": 0.5625,
581
+ "recall": 0.8181818181818182,
582
+ "f1-score": 0.6666666666666666,
583
+ "support": 11
584
+ },
585
+ "CellLine": {
586
+ "precision": 0.55,
587
+ "recall": 0.6875,
588
+ "f1-score": 0.6111111111111112,
589
+ "support": 16
590
+ },
591
+ "DNAAllele": {
592
+ "precision": 0.7857142857142857,
593
+ "recall": 0.7333333333333333,
594
+ "f1-score": 0.7586206896551724,
595
+ "support": 15
596
+ },
597
+ "DNAMutation": {
598
+ "precision": 0.8244274809160306,
599
+ "recall": 0.864,
600
+ "f1-score": 0.84375,
601
+ "support": 125
602
+ },
603
+ "Gene": {
604
+ "precision": 0.8864795918367347,
605
+ "recall": 0.9073107049608355,
606
+ "f1-score": 0.896774193548387,
607
+ "support": 766
608
+ },
609
+ "OtherMutation": {
610
+ "precision": 0.41025641025641024,
611
+ "recall": 0.5614035087719298,
612
+ "f1-score": 0.47407407407407404,
613
+ "support": 57
614
+ },
615
+ "ProteinAllele": {
616
+ "precision": 0.8571428571428571,
617
+ "recall": 0.7058823529411765,
618
+ "f1-score": 0.7741935483870968,
619
+ "support": 17
620
+ },
621
+ "ProteinMutation": {
622
+ "precision": 0.9035087719298246,
623
+ "recall": 0.8512396694214877,
624
+ "f1-score": 0.8765957446808511,
625
+ "support": 121
626
+ },
627
+ "SNP": {
628
+ "precision": 0.8148148148148148,
629
+ "recall": 1.0,
630
+ "f1-score": 0.8979591836734693,
631
+ "support": 22
632
+ },
633
+ "Species": {
634
+ "precision": 0.9546827794561934,
635
+ "recall": 0.948948948948949,
636
+ "f1-score": 0.9518072289156627,
637
+ "support": 333
638
+ },
639
+ "macro avg": {
640
+ "precision": 0.754952699206715,
641
+ "recall": 0.8077800336559531,
642
+ "f1-score": 0.7751552440712491,
643
+ "support": 1483
644
+ },
645
+ "weighted avg": {
646
+ "precision": 0.8711975142222171,
647
+ "recall": 0.8894133513149022,
648
+ "f1-score": 0.8791953544402704,
649
+ "support": 1483
650
+ }
651
+ }
652
+ }
653
+ }
performance_report.md ADDED
@@ -0,0 +1,144 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # Performance on Training Set
2
+
3
+ ## Span Level
4
+
5
+ | Label | Precision | Recall | F1-score | Support |
6
+ | --- | --- | --- | --- | --- |
7
+ | AcidChange | 1.000 | 1.000 | 1.000 | 57 |
8
+ | CellLine | 1.000 | 1.000 | 1.000 | 23 |
9
+ | DNAAllele | 1.000 | 1.000 | 1.000 | 35 |
10
+ | DNAMutation | 0.997 | 0.997 | 0.997 | 391 |
11
+ | Gene | 0.996 | 0.986 | 0.991 | 2484 |
12
+ | OtherMutation | 0.972 | 0.932 | 0.951 | 147 |
13
+ | ProteinAllele | 1.000 | 1.000 | 1.000 | 43 |
14
+ | ProteinMutation | 0.997 | 0.977 | 0.987 | 394 |
15
+ | SNP | 1.000 | 1.000 | 1.000 | 96 |
16
+ | Species | 0.999 | 0.984 | 0.992 | 1073 |
17
+ | macro avg | 0.996 | 0.988 | 0.992 | 4743 |
18
+ | weighted avg | 0.996 | 0.985 | 0.991 | 4743 |
19
+
20
+ ## Token Level
21
+
22
+ | Label | Precision | Recall | F1-score | Support |
23
+ | --- | --- | --- | --- | --- |
24
+ | O | 0.999 | 1.000 | 1.000 | 86375 |
25
+ | B-AcidChange | 1.000 | 1.000 | 1.000 | 57 |
26
+ | I-AcidChange | 1.000 | 1.000 | 1.000 | 170 |
27
+ | B-CellLine | 1.000 | 1.000 | 1.000 | 23 |
28
+ | I-CellLine | 1.000 | 1.000 | 1.000 | 33 |
29
+ | B-DNAAllele | 1.000 | 1.000 | 1.000 | 35 |
30
+ | I-DNAAllele | 1.000 | 1.000 | 1.000 | 92 |
31
+ | B-DNAMutation | 1.000 | 1.000 | 1.000 | 391 |
32
+ | I-DNAMutation | 1.000 | 1.000 | 1.000 | 2067 |
33
+ | B-Gene | 1.000 | 0.998 | 0.999 | 2469 |
34
+ | I-Gene | 0.999 | 0.989 | 0.994 | 3414 |
35
+ | B-OtherMutation | 1.000 | 1.000 | 1.000 | 147 |
36
+ | I-OtherMutation | 1.000 | 1.000 | 1.000 | 508 |
37
+ | B-ProteinAllele | 1.000 | 1.000 | 1.000 | 43 |
38
+ | I-ProteinAllele | 1.000 | 1.000 | 1.000 | 92 |
39
+ | B-ProteinMutation | 1.000 | 1.000 | 1.000 | 386 |
40
+ | I-ProteinMutation | 1.000 | 0.999 | 1.000 | 1352 |
41
+ | B-SNP | 1.000 | 1.000 | 1.000 | 96 |
42
+ | I-SNP | 1.000 | 1.000 | 1.000 | 279 |
43
+ | B-Species | 0.999 | 0.998 | 0.999 | 1066 |
44
+ | I-Species | 1.000 | 0.986 | 0.993 | 220 |
45
+ | macro avg | 1.000 | 0.999 | 0.999 | 99315 |
46
+ | weighted avg | 1.000 | 1.000 | 1.000 | 99315 |
47
+
48
+
49
+ # Performance on Validation Set
50
+
51
+ ## Span Level
52
+
53
+ | Label | Precision | Recall | F1-score | Support |
54
+ | --- | --- | --- | --- | --- |
55
+ | AcidChange | 0.655 | 0.760 | 0.704 | 25 |
56
+ | CellLine | 0.421 | 1.000 | 0.593 | 8 |
57
+ | DNAAllele | 0.941 | 0.842 | 0.889 | 19 |
58
+ | DNAMutation | 0.844 | 0.873 | 0.858 | 118 |
59
+ | Gene | 0.869 | 0.933 | 0.900 | 809 |
60
+ | OtherMutation | 0.493 | 0.667 | 0.567 | 54 |
61
+ | ProteinAllele | 0.267 | 0.571 | 0.364 | 7 |
62
+ | ProteinMutation | 0.847 | 0.868 | 0.857 | 204 |
63
+ | SNP | 1.000 | 1.000 | 1.000 | 17 |
64
+ | Species | 0.972 | 0.977 | 0.974 | 350 |
65
+ | macro avg | 0.731 | 0.849 | 0.771 | 1611 |
66
+ | weighted avg | 0.868 | 0.917 | 0.890 | 1611 |
67
+
68
+ ## Token Level
69
+
70
+ | Label | Precision | Recall | F1-score | Support |
71
+ | --- | --- | --- | --- | --- |
72
+ | O | 0.993 | 0.987 | 0.990 | 27722 |
73
+ | B-AcidChange | 0.760 | 0.760 | 0.760 | 25 |
74
+ | I-AcidChange | 0.694 | 0.758 | 0.725 | 66 |
75
+ | B-CellLine | 0.421 | 1.000 | 0.593 | 8 |
76
+ | I-CellLine | 0.714 | 1.000 | 0.833 | 10 |
77
+ | B-DNAAllele | 0.929 | 0.722 | 0.812 | 18 |
78
+ | I-DNAAllele | 0.900 | 0.692 | 0.783 | 52 |
79
+ | B-DNAMutation | 0.902 | 0.932 | 0.917 | 118 |
80
+ | I-DNAMutation | 0.935 | 0.856 | 0.894 | 674 |
81
+ | B-Gene | 0.904 | 0.969 | 0.935 | 803 |
82
+ | I-Gene | 0.873 | 0.954 | 0.912 | 1026 |
83
+ | B-OtherMutation | 0.606 | 0.741 | 0.667 | 54 |
84
+ | I-OtherMutation | 0.600 | 0.603 | 0.601 | 229 |
85
+ | B-ProteinAllele | 0.417 | 0.714 | 0.526 | 7 |
86
+ | I-ProteinAllele | 0.278 | 0.769 | 0.408 | 13 |
87
+ | B-ProteinMutation | 0.917 | 0.921 | 0.919 | 203 |
88
+ | I-ProteinMutation | 0.935 | 0.918 | 0.927 | 673 |
89
+ | B-SNP | 1.000 | 1.000 | 1.000 | 17 |
90
+ | I-SNP | 0.962 | 1.000 | 0.981 | 51 |
91
+ | B-Species | 0.969 | 0.986 | 0.977 | 346 |
92
+ | I-Species | 0.938 | 0.909 | 0.923 | 33 |
93
+ | macro avg | 0.793 | 0.866 | 0.813 | 32148 |
94
+ | weighted avg | 0.978 | 0.976 | 0.977 | 32148 |
95
+
96
+
97
+ # Performance on Testing Set
98
+
99
+ ## Span Level
100
+
101
+ | Label | Precision | Recall | F1-score | Support |
102
+ | --- | --- | --- | --- | --- |
103
+ | AcidChange | 0.562 | 0.818 | 0.667 | 11 |
104
+ | CellLine | 0.550 | 0.688 | 0.611 | 16 |
105
+ | DNAAllele | 0.786 | 0.733 | 0.759 | 15 |
106
+ | DNAMutation | 0.824 | 0.864 | 0.844 | 125 |
107
+ | Gene | 0.886 | 0.907 | 0.897 | 766 |
108
+ | OtherMutation | 0.410 | 0.561 | 0.474 | 57 |
109
+ | ProteinAllele | 0.857 | 0.706 | 0.774 | 17 |
110
+ | ProteinMutation | 0.904 | 0.851 | 0.877 | 121 |
111
+ | SNP | 0.815 | 1.000 | 0.898 | 22 |
112
+ | Species | 0.955 | 0.949 | 0.952 | 333 |
113
+ | macro avg | 0.755 | 0.808 | 0.775 | 1483 |
114
+ | weighted avg | 0.871 | 0.889 | 0.879 | 1483 |
115
+
116
+ ## Token Level
117
+
118
+ | Label | Precision | Recall | F1-score | Support |
119
+ | --- | --- | --- | --- | --- |
120
+ | O | 0.992 | 0.991 | 0.992 | 29050 |
121
+ | B-AcidChange | 0.615 | 0.727 | 0.667 | 11 |
122
+ | I-AcidChange | 0.617 | 0.853 | 0.716 | 34 |
123
+ | B-CellLine | 0.889 | 1.000 | 0.941 | 16 |
124
+ | I-CellLine | 0.889 | 0.667 | 0.762 | 36 |
125
+ | B-DNAAllele | 1.000 | 0.800 | 0.889 | 15 |
126
+ | I-DNAAllele | 0.927 | 0.776 | 0.844 | 49 |
127
+ | B-DNAMutation | 0.882 | 0.896 | 0.889 | 125 |
128
+ | I-DNAMutation | 0.943 | 0.859 | 0.899 | 675 |
129
+ | B-Gene | 0.921 | 0.943 | 0.932 | 766 |
130
+ | I-Gene | 0.915 | 0.927 | 0.921 | 1021 |
131
+ | B-OtherMutation | 0.549 | 0.789 | 0.647 | 57 |
132
+ | I-OtherMutation | 0.674 | 0.764 | 0.716 | 195 |
133
+ | B-ProteinAllele | 1.000 | 0.706 | 0.828 | 17 |
134
+ | I-ProteinAllele | 0.857 | 0.750 | 0.800 | 32 |
135
+ | B-ProteinMutation | 0.948 | 0.909 | 0.928 | 121 |
136
+ | I-ProteinMutation | 0.934 | 0.960 | 0.947 | 354 |
137
+ | B-SNP | 0.917 | 1.000 | 0.957 | 22 |
138
+ | I-SNP | 0.827 | 1.000 | 0.905 | 62 |
139
+ | B-Species | 0.972 | 0.952 | 0.962 | 333 |
140
+ | I-Species | 0.841 | 0.822 | 0.831 | 45 |
141
+ | macro avg | 0.862 | 0.861 | 0.856 | 33036 |
142
+ | weighted avg | 0.982 | 0.981 | 0.981 | 33036 |
143
+
144
+
rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:48277aac9233bb97b5ed847d4070db3bcc75af042bc93363baa226a19f421291
3
+ size 14244
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
+ "1": {
12
+ "content": "[UNK]",
13
+ "lstrip": false,
14
+ "normalized": false,
15
+ "rstrip": false,
16
+ "single_word": false,
17
+ "special": true
18
+ },
19
+ "2": {
20
+ "content": "[CLS]",
21
+ "lstrip": false,
22
+ "normalized": false,
23
+ "rstrip": false,
24
+ "single_word": false,
25
+ "special": true
26
+ },
27
+ "3": {
28
+ "content": "[SEP]",
29
+ "lstrip": false,
30
+ "normalized": false,
31
+ "rstrip": false,
32
+ "single_word": false,
33
+ "special": true
34
+ },
35
+ "4": {
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
+ "extra_special_tokens": {},
49
+ "mask_token": "[MASK]",
50
+ "model_max_length": 512,
51
+ "never_split": null,
52
+ "pad_token": "[PAD]",
53
+ "sep_token": "[SEP]",
54
+ "strip_accents": null,
55
+ "tokenize_chinese_chars": true,
56
+ "tokenizer_class": "BertTokenizer",
57
+ "unk_token": "[UNK]"
58
+ }
trainer_state.json ADDED
@@ -0,0 +1,251 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": 0.8134122644816508,
3
+ "best_model_checkpoint": "tmp_ner_fantastic-bale-13_29/run-67/checkpoint-323",
4
+ "epoch": 17.0,
5
+ "eval_steps": 500,
6
+ "global_step": 323,
7
+ "is_hyper_param_search": true,
8
+ "is_local_process_zero": true,
9
+ "is_world_process_zero": true,
10
+ "log_history": [
11
+ {
12
+ "epoch": 1.0,
13
+ "eval_accuracy": 0.8623242503421675,
14
+ "eval_loss": 0.7986403703689575,
15
+ "eval_macro_f1": 0.04409872183381453,
16
+ "eval_macro_precision": 0.04106305954010322,
17
+ "eval_macro_recall": 0.047619047619047616,
18
+ "eval_runtime": 0.7532,
19
+ "eval_samples_per_second": 132.763,
20
+ "eval_steps_per_second": 17.259,
21
+ "step": 19
22
+ },
23
+ {
24
+ "epoch": 2.0,
25
+ "eval_accuracy": 0.9034465596615653,
26
+ "eval_loss": 0.3603612184524536,
27
+ "eval_macro_f1": 0.12106609239101934,
28
+ "eval_macro_precision": 0.1259515690157799,
29
+ "eval_macro_recall": 0.14408855709243762,
30
+ "eval_runtime": 0.7511,
31
+ "eval_samples_per_second": 133.136,
32
+ "eval_steps_per_second": 17.308,
33
+ "step": 38
34
+ },
35
+ {
36
+ "epoch": 3.0,
37
+ "eval_accuracy": 0.9519410227696902,
38
+ "eval_loss": 0.18390077352523804,
39
+ "eval_macro_f1": 0.3148808447564158,
40
+ "eval_macro_precision": 0.3747061248934609,
41
+ "eval_macro_recall": 0.3212448125969776,
42
+ "eval_runtime": 0.7242,
43
+ "eval_samples_per_second": 138.085,
44
+ "eval_steps_per_second": 17.951,
45
+ "step": 57
46
+ },
47
+ {
48
+ "epoch": 4.0,
49
+ "eval_accuracy": 0.965036705238273,
50
+ "eval_loss": 0.1359422653913498,
51
+ "eval_macro_f1": 0.494219195343988,
52
+ "eval_macro_precision": 0.5137741995898872,
53
+ "eval_macro_recall": 0.5215328224472264,
54
+ "eval_runtime": 0.7408,
55
+ "eval_samples_per_second": 134.981,
56
+ "eval_steps_per_second": 17.547,
57
+ "step": 76
58
+ },
59
+ {
60
+ "epoch": 5.0,
61
+ "eval_accuracy": 0.9700447928331467,
62
+ "eval_loss": 0.11635323613882065,
63
+ "eval_macro_f1": 0.5902868533471133,
64
+ "eval_macro_precision": 0.6804922518993798,
65
+ "eval_macro_recall": 0.6058126722062964,
66
+ "eval_runtime": 0.7171,
67
+ "eval_samples_per_second": 139.455,
68
+ "eval_steps_per_second": 18.129,
69
+ "step": 95
70
+ },
71
+ {
72
+ "epoch": 6.0,
73
+ "eval_accuracy": 0.9751461988304093,
74
+ "eval_loss": 0.11470664292573929,
75
+ "eval_macro_f1": 0.6704539076542243,
76
+ "eval_macro_precision": 0.701121495572558,
77
+ "eval_macro_recall": 0.6900336890446782,
78
+ "eval_runtime": 0.7646,
79
+ "eval_samples_per_second": 130.796,
80
+ "eval_steps_per_second": 17.003,
81
+ "step": 114
82
+ },
83
+ {
84
+ "epoch": 7.0,
85
+ "eval_accuracy": 0.9759549583177802,
86
+ "eval_loss": 0.11508199572563171,
87
+ "eval_macro_f1": 0.7577631140226684,
88
+ "eval_macro_precision": 0.8839369325293773,
89
+ "eval_macro_recall": 0.7206940753172821,
90
+ "eval_runtime": 0.7145,
91
+ "eval_samples_per_second": 139.954,
92
+ "eval_steps_per_second": 18.194,
93
+ "step": 133
94
+ },
95
+ {
96
+ "epoch": 8.0,
97
+ "eval_accuracy": 0.9711024013935549,
98
+ "eval_loss": 0.12795516848564148,
99
+ "eval_macro_f1": 0.7437555649029086,
100
+ "eval_macro_precision": 0.7355731767766348,
101
+ "eval_macro_recall": 0.7848058983150393,
102
+ "eval_runtime": 0.7219,
103
+ "eval_samples_per_second": 138.514,
104
+ "eval_steps_per_second": 18.007,
105
+ "step": 152
106
+ },
107
+ {
108
+ "epoch": 9.0,
109
+ "eval_accuracy": 0.9756750031106134,
110
+ "eval_loss": 0.11990169435739517,
111
+ "eval_macro_f1": 0.7767197416060247,
112
+ "eval_macro_precision": 0.7831564923466648,
113
+ "eval_macro_recall": 0.8015624788490757,
114
+ "eval_runtime": 0.7239,
115
+ "eval_samples_per_second": 138.14,
116
+ "eval_steps_per_second": 17.958,
117
+ "step": 171
118
+ },
119
+ {
120
+ "epoch": 10.0,
121
+ "eval_accuracy": 0.9740885902700013,
122
+ "eval_loss": 0.1274283081293106,
123
+ "eval_macro_f1": 0.7875252140074755,
124
+ "eval_macro_precision": 0.7618302095117122,
125
+ "eval_macro_recall": 0.8457597210340326,
126
+ "eval_runtime": 0.721,
127
+ "eval_samples_per_second": 138.693,
128
+ "eval_steps_per_second": 18.03,
129
+ "step": 190
130
+ },
131
+ {
132
+ "epoch": 11.0,
133
+ "eval_accuracy": 0.975799427647132,
134
+ "eval_loss": 0.12168442457914352,
135
+ "eval_macro_f1": 0.787208882069088,
136
+ "eval_macro_precision": 0.7591686050582892,
137
+ "eval_macro_recall": 0.8584845414175304,
138
+ "eval_runtime": 0.7385,
139
+ "eval_samples_per_second": 135.405,
140
+ "eval_steps_per_second": 17.603,
141
+ "step": 209
142
+ },
143
+ {
144
+ "epoch": 12.0,
145
+ "eval_accuracy": 0.9761415951225582,
146
+ "eval_loss": 0.13082729279994965,
147
+ "eval_macro_f1": 0.7949954841060296,
148
+ "eval_macro_precision": 0.7671102748848504,
149
+ "eval_macro_recall": 0.8572876125725459,
150
+ "eval_runtime": 0.7165,
151
+ "eval_samples_per_second": 139.577,
152
+ "eval_steps_per_second": 18.145,
153
+ "step": 228
154
+ },
155
+ {
156
+ "epoch": 13.0,
157
+ "eval_accuracy": 0.9758927460495209,
158
+ "eval_loss": 0.13127866387367249,
159
+ "eval_macro_f1": 0.7913277905680408,
160
+ "eval_macro_precision": 0.7593907970089686,
161
+ "eval_macro_recall": 0.8538236917477225,
162
+ "eval_runtime": 0.7592,
163
+ "eval_samples_per_second": 131.72,
164
+ "eval_steps_per_second": 17.124,
165
+ "step": 247
166
+ },
167
+ {
168
+ "epoch": 14.0,
169
+ "eval_accuracy": 0.9762038073908175,
170
+ "eval_loss": 0.139426127076149,
171
+ "eval_macro_f1": 0.794994965810495,
172
+ "eval_macro_precision": 0.7684245960923229,
173
+ "eval_macro_recall": 0.8567043677879977,
174
+ "eval_runtime": 0.7554,
175
+ "eval_samples_per_second": 132.38,
176
+ "eval_steps_per_second": 17.209,
177
+ "step": 266
178
+ },
179
+ {
180
+ "epoch": 15.0,
181
+ "eval_accuracy": 0.9767015055368918,
182
+ "eval_loss": 0.13470035791397095,
183
+ "eval_macro_f1": 0.8042100829526746,
184
+ "eval_macro_precision": 0.7791800304712903,
185
+ "eval_macro_recall": 0.8620026752410235,
186
+ "eval_runtime": 0.7309,
187
+ "eval_samples_per_second": 136.825,
188
+ "eval_steps_per_second": 17.787,
189
+ "step": 285
190
+ },
191
+ {
192
+ "epoch": 16.0,
193
+ "eval_accuracy": 0.9766392932686325,
194
+ "eval_loss": 0.14018748700618744,
195
+ "eval_macro_f1": 0.8049061480916583,
196
+ "eval_macro_precision": 0.7872742624263276,
197
+ "eval_macro_recall": 0.8567333662215458,
198
+ "eval_runtime": 0.7278,
199
+ "eval_samples_per_second": 137.407,
200
+ "eval_steps_per_second": 17.863,
201
+ "step": 304
202
+ },
203
+ {
204
+ "epoch": 17.0,
205
+ "eval_accuracy": 0.976421550329725,
206
+ "eval_loss": 0.14483144879341125,
207
+ "eval_macro_f1": 0.8134122644816508,
208
+ "eval_macro_precision": 0.7925941282025739,
209
+ "eval_macro_recall": 0.8662779355061373,
210
+ "eval_runtime": 0.7265,
211
+ "eval_samples_per_second": 137.647,
212
+ "eval_steps_per_second": 17.894,
213
+ "step": 323
214
+ }
215
+ ],
216
+ "logging_steps": 500,
217
+ "max_steps": 608,
218
+ "num_input_tokens_seen": 0,
219
+ "num_train_epochs": 32,
220
+ "save_steps": 500,
221
+ "stateful_callbacks": {
222
+ "EarlyStoppingCallback": {
223
+ "args": {
224
+ "early_stopping_patience": 3,
225
+ "early_stopping_threshold": 0.001
226
+ },
227
+ "attributes": {
228
+ "early_stopping_patience_counter": 0
229
+ }
230
+ },
231
+ "TrainerControl": {
232
+ "args": {
233
+ "should_epoch_stop": false,
234
+ "should_evaluate": false,
235
+ "should_log": false,
236
+ "should_save": true,
237
+ "should_training_stop": false
238
+ },
239
+ "attributes": {}
240
+ }
241
+ },
242
+ "total_flos": 0,
243
+ "train_batch_size": 16,
244
+ "trial_name": null,
245
+ "trial_params": {
246
+ "learning_rate": 9.821730986128684e-05,
247
+ "per_device_train_batch_size": 16,
248
+ "warmup_ratio": 0.10342300416492177,
249
+ "weight_decay": 0.2819767106649835
250
+ }
251
+ }
vocab.txt ADDED
The diff for this file is too large to render. See raw diff