KR-FinBert-finetuned-ner
This model is a fine-tuned version of snunlp/KR-FinBert on the klue dataset. It achieves the following results on the evaluation set:
- Loss: 0.1634
- Precision: 0.7082
- Recall: 0.7610
- F1: 0.7337
- Accuracy: 0.9504
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: 16
- eval_batch_size: 16
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 3
Training results
Training Loss | Epoch | Step | Validation Loss | Precision | Recall | F1 | Accuracy |
---|---|---|---|---|---|---|---|
0.2028 | 1.0 | 1313 | 0.1852 | 0.6650 | 0.7060 | 0.6849 | 0.9406 |
0.1232 | 2.0 | 2626 | 0.1627 | 0.7028 | 0.7459 | 0.7237 | 0.9487 |
0.0942 | 3.0 | 3939 | 0.1634 | 0.7082 | 0.7610 | 0.7337 | 0.9504 |
Framework versions
- Transformers 4.24.0
- Pytorch 1.12.1+cu113
- Datasets 2.7.1
- Tokenizers 0.13.2
- Downloads last month
- 8
Inference API (serverless) is not available, repository is disabled.
Dataset used to train mepi/KR-FinBert-finetuned-ner
Evaluation results
- Precision on klueself-reported0.708
- Recall on klueself-reported0.761
- F1 on klueself-reported0.734
- Accuracy on klueself-reported0.950