bert-finetuned-ner-wikiann
This model is a fine-tuned version of bert-base-cased on the wikiann dataset. It achieves the following results on the evaluation set:
- Loss: 0.3147
- Precision: 0.8174
- Recall: 0.8443
- F1: 0.8307
- Accuracy: 0.9269
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
- num_epochs: 3
Training results
| Training Loss | Epoch | Step | Validation Loss | Precision | Recall | F1 | Accuracy |
|---|---|---|---|---|---|---|---|
| 0.2826 | 1.0 | 2500 | 0.2833 | 0.7952 | 0.8265 | 0.8105 | 0.9205 |
| 0.2052 | 2.0 | 5000 | 0.2620 | 0.8013 | 0.8371 | 0.8188 | 0.9255 |
| 0.1356 | 3.0 | 7500 | 0.3147 | 0.8174 | 0.8443 | 0.8307 | 0.9269 |
Framework versions
- Transformers 4.28.0
- Pytorch 2.0.1+cu118
- Datasets 2.12.0
- Tokenizers 0.13.3