output
This model is a fine-tuned version of PlanTL-GOB-ES/bsc-bio-ehr-es on the Rodrigo1771/symptemist-ner dataset. It achieves the following results on the evaluation set:
- Loss: 0.3220
- Precision: 0.7028
- Recall: 0.7132
- F1: 0.7080
- Accuracy: 0.9507
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: 4
- eval_batch_size: 8
- seed: 42
- gradient_accumulation_steps: 4
- total_train_batch_size: 16
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 10.0
Training results
Training Loss | Epoch | Step | Validation Loss | Precision | Recall | F1 | Accuracy |
---|---|---|---|---|---|---|---|
0.1924 | 1.0 | 600 | 0.1498 | 0.5610 | 0.7044 | 0.6246 | 0.9450 |
0.1106 | 2.0 | 1200 | 0.1653 | 0.6096 | 0.7154 | 0.6583 | 0.9449 |
0.0712 | 3.0 | 1800 | 0.1909 | 0.6510 | 0.7044 | 0.6767 | 0.9476 |
0.0496 | 4.0 | 2400 | 0.2023 | 0.6571 | 0.7269 | 0.6902 | 0.9493 |
0.0188 | 5.0 | 3000 | 0.2545 | 0.6684 | 0.7050 | 0.6862 | 0.9482 |
0.0106 | 6.0 | 3600 | 0.2608 | 0.6870 | 0.6979 | 0.6924 | 0.9492 |
0.0089 | 7.0 | 4200 | 0.2945 | 0.6834 | 0.7088 | 0.6959 | 0.9497 |
0.0058 | 8.0 | 4800 | 0.3220 | 0.7028 | 0.7132 | 0.7080 | 0.9507 |
0.004 | 9.0 | 5400 | 0.3512 | 0.6830 | 0.7230 | 0.7025 | 0.9498 |
0.0015 | 10.0 | 6000 | 0.3558 | 0.6763 | 0.7274 | 0.7009 | 0.9493 |
Framework versions
- Transformers 4.41.2
- Pytorch 2.3.0+cu121
- Datasets 2.20.0
- Tokenizers 0.19.1
- Downloads last month
- 4
Finetuned from
Dataset used to train Rodrigo1771/bsc-bio-ehr-es-symptemist-ner
Evaluation results
- Precision on Rodrigo1771/symptemist-nervalidation set self-reported0.703
- Recall on Rodrigo1771/symptemist-nervalidation set self-reported0.713
- F1 on Rodrigo1771/symptemist-nervalidation set self-reported0.708
- Accuracy on Rodrigo1771/symptemist-nervalidation set self-reported0.951