pubmed-abs-noise-04
This model is a fine-tuned version of facebook/bart-base on an unknown dataset. It achieves the following results on the evaluation set:
- Loss: 0.5519
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_steps: 10
- num_epochs: 3
Training results
| Training Loss | Epoch | Step | Validation Loss |
|---|---|---|---|
| 0.9533 | 0.11 | 500 | 0.7930 |
| 0.8667 | 0.21 | 1000 | 0.7234 |
| 0.797 | 0.32 | 1500 | 0.6901 |
| 0.7887 | 0.43 | 2000 | 0.6696 |
| 0.7616 | 0.54 | 2500 | 0.6600 |
| 0.6708 | 0.64 | 3000 | 0.6435 |
| 0.6794 | 0.75 | 3500 | 0.6287 |
| 0.6342 | 0.86 | 4000 | 0.6163 |
| 0.6815 | 0.96 | 4500 | 0.6073 |
| 0.6312 | 1.07 | 5000 | 0.6038 |
| 0.5506 | 1.18 | 5500 | 0.5975 |
| 0.5828 | 1.28 | 6000 | 0.5972 |
| 0.5568 | 1.39 | 6500 | 0.5920 |
| 0.5834 | 1.5 | 7000 | 0.5809 |
| 0.5236 | 1.61 | 7500 | 0.5808 |
| 0.5446 | 1.71 | 8000 | 0.5727 |
| 0.5838 | 1.82 | 8500 | 0.5691 |
| 0.5038 | 1.93 | 9000 | 0.5628 |
| 0.469 | 2.03 | 9500 | 0.5687 |
| 0.4529 | 2.14 | 10000 | 0.5673 |
| 0.4987 | 2.25 | 10500 | 0.5614 |
| 0.4471 | 2.35 | 11000 | 0.5621 |
| 0.4831 | 2.46 | 11500 | 0.5569 |
| 0.4683 | 2.57 | 12000 | 0.5565 |
| 0.4547 | 2.68 | 12500 | 0.5562 |
| 0.4346 | 2.78 | 13000 | 0.5543 |
| 0.47 | 2.89 | 13500 | 0.5534 |
| 0.4144 | 3.0 | 14000 | 0.5519 |
Framework versions
- Transformers 4.34.1
- Pytorch 2.1.0
- Datasets 2.14.7
- Tokenizers 0.14.1