bart-pt-asqa-cb
This model is a fine-tuned version of vblagoje/bart_lfqa on the None dataset. It achieves the following results on the evaluation set:
- Loss: 2.5362
- Rougelsum: 38.9467
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-06
- train_batch_size: 16
- eval_batch_size: 8
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 20
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Rougelsum |
|---|---|---|---|---|
| No log | 1.0 | 273 | 2.5653 | 37.6939 |
| 2.6009 | 2.0 | 546 | 2.5295 | 38.2398 |
| 2.6009 | 3.0 | 819 | 2.5315 | 38.5946 |
| 2.3852 | 4.0 | 1092 | 2.5146 | 38.4771 |
| 2.3852 | 5.0 | 1365 | 2.5240 | 38.5706 |
| 2.2644 | 6.0 | 1638 | 2.5253 | 38.7506 |
| 2.2644 | 7.0 | 1911 | 2.5355 | 38.9004 |
| 2.1703 | 8.0 | 2184 | 2.5309 | 38.9528 |
| 2.1703 | 9.0 | 2457 | 2.5362 | 38.9467 |
Framework versions
- Transformers 4.23.0.dev0
- Pytorch 1.12.1+cu102
- Datasets 2.4.0
- Tokenizers 0.12.1