SETH_2e-05_250
This model is a fine-tuned version of microsoft/BiomedNLP-PubMedBERT-base-uncased-abstract-fulltext on the None dataset. It achieves the following results on the evaluation set:
- Loss: 0.0676
- Precision: 0.7820
- Recall: 0.7891
- F1: 0.7855
- Accuracy: 0.9837
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
- training_steps: 500
Training results
Training Loss | Epoch | Step | Validation Loss | Precision | Recall | F1 | Accuracy |
---|---|---|---|---|---|---|---|
0.4635 | 0.76 | 25 | 0.1662 | 0.0 | 0.0 | 0.0 | 0.9625 |
0.0991 | 1.52 | 50 | 0.0805 | 0.7425 | 0.6291 | 0.6811 | 0.9770 |
0.0585 | 2.27 | 75 | 0.0616 | 0.6952 | 0.7836 | 0.7368 | 0.9801 |
0.0495 | 3.03 | 100 | 0.0564 | 0.7129 | 0.7945 | 0.7515 | 0.9819 |
0.0413 | 3.79 | 125 | 0.0531 | 0.7188 | 0.8273 | 0.7692 | 0.9824 |
0.0393 | 4.55 | 150 | 0.0512 | 0.7350 | 0.8218 | 0.7760 | 0.9827 |
0.0317 | 5.3 | 175 | 0.0490 | 0.7543 | 0.7927 | 0.7730 | 0.9832 |
0.0283 | 6.06 | 200 | 0.0546 | 0.7780 | 0.7836 | 0.7808 | 0.9833 |
0.0255 | 6.82 | 225 | 0.0524 | 0.7504 | 0.7818 | 0.7658 | 0.9829 |
0.022 | 7.58 | 250 | 0.0567 | 0.7613 | 0.7945 | 0.7776 | 0.9835 |
0.0183 | 8.33 | 275 | 0.0566 | 0.7730 | 0.7927 | 0.7828 | 0.9842 |
0.0179 | 9.09 | 300 | 0.0592 | 0.7668 | 0.7655 | 0.7662 | 0.9830 |
0.016 | 9.85 | 325 | 0.0648 | 0.7855 | 0.7855 | 0.7855 | 0.9841 |
0.0135 | 10.61 | 350 | 0.0639 | 0.7732 | 0.7873 | 0.7802 | 0.9832 |
0.0121 | 11.36 | 375 | 0.0676 | 0.7820 | 0.7891 | 0.7855 | 0.9837 |
Framework versions
- Transformers 4.27.4
- Pytorch 1.13.1+cu116
- Datasets 2.11.0
- Tokenizers 0.13.2
- Downloads last month
- 4
This model does not have enough activity to be deployed to Inference API (serverless) yet. Increase its social
visibility and check back later, or deploy to Inference Endpoints (dedicated)
instead.