<!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. -->
ClinicalBioBERT
This model is a fine-tuned version of emilyalsentzer/Bio_ClinicalBERT on an unknown dataset. It achieves the following results on the evaluation set:
- Loss: 0.9404
- Accuracy: 0.77
- Precision: 0.8333
- Recall: 0.8209
- F1: 0.8271
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: 10
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy | Precision | Recall | F1 |
|---|---|---|---|---|---|---|---|
| 0.693 | 1.0 | 50 | 0.6142 | 0.61 | 0.8182 | 0.5373 | 0.6486 |
| 0.5547 | 2.0 | 100 | 0.5753 | 0.66 | 0.8367 | 0.6119 | 0.7069 |
| 0.3912 | 3.0 | 150 | 0.5167 | 0.8 | 0.8406 | 0.8657 | 0.8529 |
| 0.2618 | 4.0 | 200 | 0.6664 | 0.8 | 0.8133 | 0.9104 | 0.8592 |
| 0.1648 | 5.0 | 250 | 0.5954 | 0.79 | 0.8594 | 0.8209 | 0.8397 |
| 0.1446 | 6.0 | 300 | 0.6131 | 0.81 | 0.8871 | 0.8209 | 0.8527 |
| 0.0841 | 7.0 | 350 | 0.8966 | 0.79 | 0.8194 | 0.8806 | 0.8489 |
| 0.0708 | 8.0 | 400 | 0.9366 | 0.78 | 0.8169 | 0.8657 | 0.8406 |
| 0.049 | 9.0 | 450 | 0.9523 | 0.78 | 0.8358 | 0.8358 | 0.8358 |
| 0.0516 | 10.0 | 500 | 0.9404 | 0.77 | 0.8333 | 0.8209 | 0.8271 |
Framework versions
- Transformers 4.21.2
- Pytorch 1.12.1+cu113
- Datasets 2.4.0
- Tokenizers 0.12.1
