PubMedBERT-LitCovid-v1.3h
This model is a fine-tuned version of microsoft/BiomedNLP-BiomedBERT-base-uncased-abstract on an unknown dataset. It achieves the following results on the evaluation set:
- Loss: 0.8774
- Hamming loss: 0.0137
- F1 micro: 0.8798
- F1 macro: 0.4661
- F1 weighted: 0.8928
- F1 samples: 0.8920
- Precision micro: 0.8303
- Precision macro: 0.3861
- Precision weighted: 0.8575
- Precision samples: 0.8800
- Recall micro: 0.9356
- Recall macro: 0.7033
- Recall weighted: 0.9356
- Recall samples: 0.9451
- Roc Auc: 0.9624
- Accuracy: 0.7154
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: 16
- eval_batch_size: 16
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_ratio: 0.157
- num_epochs: 5
Training results
Training Loss | Epoch | Step | Validation Loss | Hamming loss | F1 micro | F1 macro | F1 weighted | F1 samples | Precision micro | Precision macro | Precision weighted | Precision samples | Recall micro | Recall macro | Recall weighted | Recall samples | Roc Auc | Accuracy |
---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
1.2354 | 1.0 | 2272 | 0.4702 | 0.0473 | 0.6733 | 0.2437 | 0.8136 | 0.7898 | 0.5349 | 0.1961 | 0.7600 | 0.7551 | 0.9084 | 0.7924 | 0.9084 | 0.9263 | 0.9318 | 0.5030 |
0.9475 | 2.0 | 4544 | 0.5277 | 0.0214 | 0.8208 | 0.3449 | 0.8664 | 0.8541 | 0.7441 | 0.2778 | 0.8345 | 0.8362 | 0.9151 | 0.7371 | 0.9151 | 0.9304 | 0.9487 | 0.6209 |
0.7806 | 3.0 | 6816 | 0.5645 | 0.0187 | 0.8432 | 0.3871 | 0.8765 | 0.8632 | 0.7670 | 0.3096 | 0.8337 | 0.8350 | 0.9362 | 0.7431 | 0.9362 | 0.9472 | 0.9601 | 0.6337 |
0.5358 | 4.0 | 9088 | 0.7518 | 0.0145 | 0.8738 | 0.4462 | 0.8893 | 0.8873 | 0.8211 | 0.3653 | 0.8537 | 0.8742 | 0.9337 | 0.6984 | 0.9337 | 0.9443 | 0.9611 | 0.7024 |
0.1808 | 5.0 | 11360 | 0.8774 | 0.0137 | 0.8798 | 0.4661 | 0.8928 | 0.8920 | 0.8303 | 0.3861 | 0.8575 | 0.8800 | 0.9356 | 0.7033 | 0.9356 | 0.9451 | 0.9624 | 0.7154 |
Framework versions
- Transformers 4.28.0
- Pytorch 2.0.0
- Datasets 2.1.0
- Tokenizers 0.13.3
- Downloads last month
- 4