<!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. -->
bert-base-chinese-wikiann-zh-ner
This model is a fine-tuned version of bert-base-chinese on the wikiann dataset. It achieves the following results on the evaluation set:
- Loss: 0.2092
- Precision: 0.7891
- Recall: 0.8061
- F1: 0.7975
- Accuracy: 0.9432
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: 2
- mixed_precision_training: Native AMP
Training results
Training Loss | Epoch | Step | Validation Loss | Precision | Recall | F1 | Accuracy |
---|---|---|---|---|---|---|---|
0.842 | 0.16 | 400 | 0.3530 | 0.5535 | 0.6872 | 0.6131 | 0.8927 |
0.32 | 0.32 | 800 | 0.2800 | 0.6929 | 0.6749 | 0.6838 | 0.9190 |
0.2928 | 0.48 | 1200 | 0.2438 | 0.7031 | 0.7661 | 0.7333 | 0.9301 |
0.245 | 0.64 | 1600 | 0.2525 | 0.6959 | 0.7919 | 0.7408 | 0.9280 |
0.2236 | 0.8 | 2000 | 0.2315 | 0.7441 | 0.7503 | 0.7472 | 0.9342 |
0.2444 | 0.96 | 2400 | 0.2119 | 0.7719 | 0.7675 | 0.7697 | 0.9379 |
0.1899 | 1.12 | 2800 | 0.2267 | 0.7531 | 0.8062 | 0.7788 | 0.9387 |
0.1649 | 1.28 | 3200 | 0.2249 | 0.7519 | 0.8202 | 0.7846 | 0.9395 |
0.1521 | 1.44 | 3600 | 0.2220 | 0.7778 | 0.8032 | 0.7903 | 0.9413 |
0.1787 | 1.6 | 4000 | 0.2185 | 0.7879 | 0.7860 | 0.7869 | 0.9417 |
0.146 | 1.76 | 4400 | 0.2134 | 0.7721 | 0.8128 | 0.7919 | 0.9416 |
0.1557 | 1.92 | 4800 | 0.2111 | 0.7857 | 0.8101 | 0.7977 | 0.9429 |
Framework versions
- Transformers 4.27.4
- Pytorch 2.0.0+cu118
- Datasets 2.11.0
- Tokenizers 0.13.3