metadata
language:
- List of ISO 639-1 code for your language
- zh
widget:
- text: 中央疫情指揮中心臨時記者會宣布全院區為紅區，擴大隔離，但鄭文燦早在七十二小時前就主張，只要是先前在桃園醫院住院、轉院的患者與陪病家屬，都要居家隔離
example_title: 範例ㄧ
- text: 台東地檢署21日指揮警方前往張靜的事務所及黃姓女友所經營的按摩店進行搜索
example_title: 範例二
- text: 各地停電事件頻傳，即便經濟部與台電均否認「台灣缺電」，但也難消國人的疑慮。
example_title: 範例三
license: gpl-3.0 tags: - generated_from_trainer metrics: - precision - recall - f1 - accuracy model-index: - name: albert-base-chinese-0407-ner results: []
albert-base-chinese-0407-ner
This model is a fine-tuned version of ckiplab/albert-base-chinese on an unknown dataset. It achieves the following results on the evaluation set:
- Loss: 0.0948
- Precision: 0.8603
- Recall: 0.8871
- F1: 0.8735
- Accuracy: 0.9704
Model description
More information needed
Intended uses & limitations
More information needed
Training and evaluation data
More information needed
Training procedure
Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 2e-05
- train_batch_size: 8
- eval_batch_size: 8
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_ratio: 0.1
- num_epochs: 3
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Precision | Recall | F1 | Accuracy |
|---|---|---|---|---|---|---|---|
| 1.3484 | 0.05 | 500 | 0.5395 | 0.1841 | 0.1976 | 0.1906 | 0.8465 |
| 0.3948 | 0.09 | 1000 | 0.2910 | 0.6138 | 0.7113 | 0.6590 | 0.9263 |
| 0.2388 | 0.14 | 1500 | 0.2030 | 0.6628 | 0.7797 | 0.7165 | 0.9414 |
| 0.1864 | 0.18 | 2000 | 0.1729 | 0.7490 | 0.7935 | 0.7706 | 0.9498 |
| 0.1754 | 0.23 | 2500 | 0.1641 | 0.7415 | 0.7869 | 0.7635 | 0.9505 |
| 0.1558 | 0.28 | 3000 | 0.1532 | 0.7680 | 0.8002 | 0.7838 | 0.9530 |
| 0.1497 | 0.32 | 3500 | 0.1424 | 0.7865 | 0.8282 | 0.8068 | 0.9555 |
| 0.1488 | 0.37 | 4000 | 0.1373 | 0.7887 | 0.8111 | 0.7997 | 0.9553 |
| 0.1361 | 0.42 | 4500 | 0.1311 | 0.7942 | 0.8382 | 0.8156 | 0.9590 |
| 0.1335 | 0.46 | 5000 | 0.1264 | 0.7948 | 0.8423 | 0.8179 | 0.9596 |
| 0.1296 | 0.51 | 5500 | 0.1242 | 0.8129 | 0.8416 | 0.8270 | 0.9603 |
| 0.1338 | 0.55 | 6000 | 0.1315 | 0.7910 | 0.8588 | 0.8235 | 0.9586 |
| 0.1267 | 0.6 | 6500 | 0.1193 | 0.8092 | 0.8399 | 0.8243 | 0.9609 |
| 0.1207 | 0.65 | 7000 | 0.1205 | 0.8021 | 0.8469 | 0.8239 | 0.9601 |
| 0.1214 | 0.69 | 7500 | 0.1201 | 0.7969 | 0.8489 | 0.8220 | 0.9605 |
| 0.1168 | 0.74 | 8000 | 0.1134 | 0.8087 | 0.8607 | 0.8339 | 0.9620 |
| 0.1162 | 0.78 | 8500 | 0.1127 | 0.8177 | 0.8492 | 0.8331 | 0.9625 |
| 0.1202 | 0.83 | 9000 | 0.1283 | 0.7986 | 0.8550 | 0.8259 | 0.9580 |
| 0.1135 | 0.88 | 9500 | 0.1101 | 0.8213 | 0.8572 | 0.8389 | 0.9638 |
| 0.1121 | 0.92 | 10000 | 0.1097 | 0.8190 | 0.8588 | 0.8384 | 0.9635 |
| 0.1091 | 0.97 | 10500 | 0.1088 | 0.8180 | 0.8521 | 0.8347 | 0.9632 |
| 0.1058 | 1.02 | 11000 | 0.1085 | 0.8136 | 0.8716 | 0.8416 | 0.9630 |
| 0.0919 | 1.06 | 11500 | 0.1079 | 0.8309 | 0.8566 | 0.8436 | 0.9646 |
| 0.0914 | 1.11 | 12000 | 0.1079 | 0.8423 | 0.8542 | 0.8482 | 0.9656 |
| 0.0921 | 1.15 | 12500 | 0.1109 | 0.8312 | 0.8647 | 0.8476 | 0.9646 |
| 0.0926 | 1.2 | 13000 | 0.1240 | 0.8413 | 0.8488 | 0.8451 | 0.9637 |
| 0.0914 | 1.25 | 13500 | 0.1040 | 0.8336 | 0.8666 | 0.8498 | 0.9652 |
| 0.0917 | 1.29 | 14000 | 0.1032 | 0.8352 | 0.8707 | 0.8526 | 0.9662 |
| 0.0928 | 1.34 | 14500 | 0.1052 | 0.8347 | 0.8656 | 0.8498 | 0.9651 |
| 0.0906 | 1.38 | 15000 | 0.1032 | 0.8399 | 0.8619 | 0.8507 | 0.9662 |
| 0.0903 | 1.43 | 15500 | 0.1074 | 0.8180 | 0.8708 | 0.8436 | 0.9651 |
| 0.0889 | 1.48 | 16000 | 0.0990 | 0.8367 | 0.8713 | 0.8537 | 0.9670 |
| 0.0914 | 1.52 | 16500 | 0.1055 | 0.8508 | 0.8506 | 0.8507 | 0.9661 |
| 0.0934 | 1.57 | 17000 | 0.0979 | 0.8326 | 0.8740 | 0.8528 | 0.9669 |
| 0.0898 | 1.62 | 17500 | 0.1022 | 0.8393 | 0.8615 | 0.8502 | 0.9668 |
| 0.0869 | 1.66 | 18000 | 0.0962 | 0.8484 | 0.8762 | 0.8621 | 0.9682 |
| 0.089 | 1.71 | 18500 | 0.1008 | 0.8447 | 0.8714 | 0.8579 | 0.9674 |
| 0.0927 | 1.75 | 19000 | 0.0986 | 0.8379 | 0.8749 | 0.8560 | 0.9673 |
| 0.0883 | 1.8 | 19500 | 0.0965 | 0.8518 | 0.8749 | 0.8632 | 0.9688 |
| 0.0965 | 1.85 | 20000 | 0.0937 | 0.8412 | 0.8766 | 0.8585 | 0.9682 |
| 0.0834 | 1.89 | 20500 | 0.0920 | 0.8451 | 0.8862 | 0.8652 | 0.9687 |
| 0.0817 | 1.94 | 21000 | 0.0943 | 0.8439 | 0.8800 | 0.8616 | 0.9686 |
| 0.088 | 1.99 | 21500 | 0.0927 | 0.8483 | 0.8762 | 0.8620 | 0.9683 |
| 0.0705 | 2.03 | 22000 | 0.0993 | 0.8525 | 0.8783 | 0.8652 | 0.9690 |
| 0.0709 | 2.08 | 22500 | 0.0976 | 0.8610 | 0.8697 | 0.8653 | 0.9689 |
| 0.0655 | 2.12 | 23000 | 0.0997 | 0.8585 | 0.8665 | 0.8625 | 0.9683 |
| 0.0656 | 2.17 | 23500 | 0.0966 | 0.8569 | 0.8822 | 0.8694 | 0.9695 |
| 0.0698 | 2.22 | 24000 | 0.0955 | 0.8604 | 0.8775 | 0.8689 | 0.9696 |
| 0.065 | 2.26 | 24500 | 0.0971 | 0.8614 | 0.8780 | 0.8696 | 0.9697 |
| 0.0653 | 2.31 | 25000 | 0.0959 | 0.8600 | 0.8787 | 0.8692 | 0.9698 |
| 0.0685 | 2.35 | 25500 | 0.1001 | 0.8610 | 0.8710 | 0.8659 | 0.9690 |
| 0.0684 | 2.4 | 26000 | 0.0969 | 0.8490 | 0.8877 | 0.8679 | 0.9690 |
| 0.0657 | 2.45 | 26500 | 0.0954 | 0.8532 | 0.8832 | 0.8680 | 0.9696 |
| 0.0668 | 2.49 | 27000 | 0.0947 | 0.8604 | 0.8793 | 0.8698 | 0.9695 |
| 0.0644 | 2.54 | 27500 | 0.0989 | 0.8527 | 0.8790 | 0.8656 | 0.9696 |
| 0.0685 | 2.59 | 28000 | 0.0955 | 0.8596 | 0.8772 | 0.8683 | 0.9700 |
| 0.0702 | 2.63 | 28500 | 0.0937 | 0.8585 | 0.8837 | 0.8709 | 0.9700 |
| 0.0644 | 2.68 | 29000 | 0.0946 | 0.8605 | 0.8830 | 0.8716 | 0.9702 |
| 0.065 | 2.72 | 29500 | 0.0953 | 0.8617 | 0.8822 | 0.8719 | 0.9701 |
| 0.063 | 2.77 | 30000 | 0.0943 | 0.8597 | 0.8848 | 0.8721 | 0.9701 |
| 0.0638 | 2.82 | 30500 | 0.0941 | 0.8619 | 0.8846 | 0.8731 | 0.9702 |
| 0.066 | 2.86 | 31000 | 0.0942 | 0.8608 | 0.8847 | 0.8726 | 0.9701 |
| 0.0589 | 2.91 | 31500 | 0.0952 | 0.8632 | 0.8836 | 0.8733 | 0.9704 |
| 0.0568 | 2.95 | 32000 | 0.0948 | 0.8603 | 0.8871 | 0.8735 | 0.9704 |
Framework versions
- Transformers 4.18.0
- Pytorch 1.10.0+cu111
- Datasets 2.0.0
- Tokenizers 0.11.6