NLP-CIC-WFU_Clinical_Cases_NER_Paragraph_Tokenized_mBERT_cased_fine_tuned
This model is a fine-tuned version of bert-base-multilingual-cased on the None dataset. It achieves the following results on the evaluation set:
- Loss: 0.0537
- Precision: 0.8585
- Recall: 0.7101
- F1: 0.7773
- Accuracy: 0.9893
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: 5e-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: 7
Training results
| Training Loss | Epoch | Step | Validation Loss | Precision | Recall | F1 | Accuracy |
|---|---|---|---|---|---|---|---|
| 0.0693 | 1.0 | 514 | 0.0416 | 0.9485 | 0.6492 | 0.7708 | 0.9884 |
| 0.0367 | 2.0 | 1028 | 0.0396 | 0.9391 | 0.6710 | 0.7827 | 0.9892 |
| 0.0283 | 3.0 | 1542 | 0.0385 | 0.9388 | 0.6889 | 0.7947 | 0.9899 |
| 0.0222 | 4.0 | 2056 | 0.0422 | 0.9456 | 0.6790 | 0.7904 | 0.9898 |
| 0.0182 | 5.0 | 2570 | 0.0457 | 0.9349 | 0.6925 | 0.7956 | 0.9901 |
| 0.013 | 6.0 | 3084 | 0.0484 | 0.8947 | 0.7062 | 0.7894 | 0.9899 |
| 0.0084 | 7.0 | 3598 | 0.0537 | 0.8585 | 0.7101 | 0.7773 | 0.9893 |
Framework versions
- Transformers 4.19.2
- Pytorch 1.11.0+cu113
- Datasets 2.2.2
- Tokenizers 0.12.1