metadata
license: apache-2.0
base_model: facebook/bart-base
tags:
- generated_from_trainer
model-index:
- name: pubmed-abs-sub-04
results: []
pubmed-abs-sub-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: 0.1795
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 |
|---|---|---|---|
| 0.5322 | 0.11 | 500 | 0.4266 |
| 0.3544 | 0.21 | 1000 | 0.3467 |
| 0.3892 | 0.32 | 1500 | 0.3076 |
| 0.3657 | 0.43 | 2000 | 0.2814 |
| 0.2904 | 0.54 | 2500 | 0.2656 |
| 0.2551 | 0.64 | 3000 | 0.2447 |
| 0.2641 | 0.75 | 3500 | 0.2373 |
| 0.284 | 0.86 | 4000 | 0.2218 |
| 0.2631 | 0.96 | 4500 | 0.2186 |
| 0.2031 | 1.07 | 5000 | 0.2196 |
| 0.1983 | 1.18 | 5500 | 0.2147 |
| 0.1842 | 1.28 | 6000 | 0.2127 |
| 0.1697 | 1.39 | 6500 | 0.2020 |
| 0.2006 | 1.5 | 7000 | 0.2014 |
| 0.1788 | 1.61 | 7500 | 0.1987 |
| 0.1589 | 1.71 | 8000 | 0.1938 |
| 0.2047 | 1.82 | 8500 | 0.1909 |
| 0.1583 | 1.93 | 9000 | 0.1896 |
| 0.1437 | 2.03 | 9500 | 0.1900 |
| 0.1466 | 2.14 | 10000 | 0.1890 |
| 0.1618 | 2.25 | 10500 | 0.1883 |
| 0.1187 | 2.35 | 11000 | 0.1856 |
| 0.1142 | 2.46 | 11500 | 0.1837 |
| 0.1382 | 2.57 | 12000 | 0.1811 |
| 0.1403 | 2.68 | 12500 | 0.1804 |
| 0.1045 | 2.78 | 13000 | 0.1810 |
| 0.1056 | 2.89 | 13500 | 0.1801 |
| 0.12 | 3.0 | 14000 | 0.1795 |
Framework versions
- Transformers 4.34.1
- Pytorch 2.1.0
- Datasets 2.14.6
- Tokenizers 0.14.1