bert-base-chinese-finetuned-ner_0301_J_DATA
This model is a fine-tuned version of ckiplab/bert-base-chinese-ner on the None dataset. It achieves the following results on the evaluation set:
- Loss: 0.0764
- Precision: 0.9663
- Recall: 0.9708
- F1: 0.9685
- Accuracy: 0.9925
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: 2
- eval_batch_size: 2
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 15
Training results
Training Loss | Epoch | Step | Validation Loss | Precision | Recall | F1 | Accuracy |
---|---|---|---|---|---|---|---|
0.3872 | 1.0 | 705 | 0.1222 | 0.9088 | 0.9311 | 0.9198 | 0.9781 |
0.0732 | 2.0 | 1410 | 0.0642 | 0.9303 | 0.9509 | 0.9405 | 0.9900 |
0.034 | 3.0 | 2115 | 0.0588 | 0.9616 | 0.9661 | 0.9639 | 0.9909 |
0.0267 | 4.0 | 2820 | 0.0631 | 0.9639 | 0.9673 | 0.9656 | 0.9925 |
0.0232 | 5.0 | 3525 | 0.0617 | 0.9630 | 0.9720 | 0.9674 | 0.9924 |
0.017 | 6.0 | 4230 | 0.0652 | 0.9674 | 0.9708 | 0.9691 | 0.9926 |
0.0123 | 7.0 | 4935 | 0.0573 | 0.9618 | 0.9720 | 0.9669 | 0.9923 |
0.009 | 8.0 | 5640 | 0.0667 | 0.9651 | 0.9696 | 0.9674 | 0.9922 |
0.0055 | 9.0 | 6345 | 0.0768 | 0.9640 | 0.9696 | 0.9668 | 0.9925 |
0.0045 | 10.0 | 7050 | 0.0775 | 0.9662 | 0.9696 | 0.9679 | 0.9925 |
0.004 | 11.0 | 7755 | 0.0753 | 0.9606 | 0.9685 | 0.9645 | 0.9923 |
0.0018 | 12.0 | 8460 | 0.0735 | 0.9629 | 0.9696 | 0.9662 | 0.9925 |
0.0019 | 13.0 | 9165 | 0.0754 | 0.9663 | 0.9708 | 0.9685 | 0.9927 |
0.0019 | 14.0 | 9870 | 0.0760 | 0.9651 | 0.9696 | 0.9674 | 0.9925 |
0.0013 | 15.0 | 10575 | 0.0764 | 0.9663 | 0.9708 | 0.9685 | 0.9925 |
Framework versions
- Transformers 4.20.1
- Pytorch 1.13.0+cu117
- Datasets 2.8.0
- Tokenizers 0.12.1
- Downloads last month
- 5
This model does not have enough activity to be deployed to Inference API (serverless) yet. Increase its social
visibility and check back later, or deploy to Inference Endpoints (dedicated)
instead.