SETH_2e-5_29_03
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: 0.0759
- Precision: 0.6596
- Recall: 0.8537
- F1: 0.7442
- Accuracy: 0.9785
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 |
|---|---|---|---|---|---|---|---|
| 0.5486 | 0.96 | 25 | 0.2095 | 0.0 | 0.0 | 0.0 | 0.9583 |
| 0.1393 | 1.92 | 50 | 0.1119 | 0.6094 | 0.2685 | 0.3728 | 0.9627 |
| 0.0842 | 2.88 | 75 | 0.0841 | 0.5249 | 0.7986 | 0.6334 | 0.9737 |
| 0.0627 | 3.85 | 100 | 0.0737 | 0.5532 | 0.8141 | 0.6588 | 0.9753 |
| 0.0602 | 4.81 | 125 | 0.0683 | 0.6 | 0.8726 | 0.7111 | 0.9756 |
| 0.0448 | 5.77 | 150 | 0.0639 | 0.6717 | 0.8451 | 0.7485 | 0.9803 |
| 0.04 | 6.73 | 175 | 0.0655 | 0.6381 | 0.8709 | 0.7365 | 0.9781 |
| 0.0339 | 7.69 | 200 | 0.0621 | 0.6450 | 0.8726 | 0.7418 | 0.9788 |
| 0.0293 | 8.65 | 225 | 0.0639 | 0.6764 | 0.7952 | 0.7310 | 0.9794 |
| 0.0268 | 9.62 | 250 | 0.0648 | 0.6869 | 0.8571 | 0.7626 | 0.9804 |
| 0.0229 | 10.58 | 275 | 0.0710 | 0.6703 | 0.8571 | 0.7523 | 0.9790 |
| 0.0223 | 11.54 | 300 | 0.0668 | 0.7030 | 0.8107 | 0.7530 | 0.9806 |
| 0.0199 | 12.5 | 325 | 0.0726 | 0.7072 | 0.8313 | 0.7642 | 0.9803 |
| 0.018 | 13.46 | 350 | 0.0759 | 0.6596 | 0.8537 | 0.7442 | 0.9785 |
Framework versions
- Transformers 4.27.3
- Pytorch 1.13.1+cu116
- Datasets 2.10.1
- Tokenizers 0.13.2