bart-base-finetuned-xlsum-with-multi-news-10-epoch
This model is a fine-tuned version of facebook/bart-base on the None dataset. It achieves the following results on the evaluation set:
- Loss: 1.7450
- Rouge1: 38.4651
- Rouge2: 17.181
- Rougel: 31.5511
- Rougelsum: 31.6061
- Gen Len: 19.6752
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
- num_epochs: 10
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Rouge1 | Rouge2 | Rougel | Rougelsum | Gen Len |
|---|---|---|---|---|---|---|---|---|
| 2.2151 | 1.0 | 20543 | 1.8364 | 36.8242 | 15.4678 | 29.9179 | 29.9756 | 19.5911 |
| 2.0765 | 2.0 | 41086 | 1.7958 | 37.2066 | 15.9599 | 30.3947 | 30.4615 | 19.5761 |
| 1.9964 | 3.0 | 61629 | 1.7740 | 37.5972 | 16.3675 | 30.7002 | 30.7514 | 19.6134 |
| 1.9353 | 4.0 | 82172 | 1.7585 | 37.9193 | 16.6538 | 31.0422 | 31.0914 | 19.5524 |
| 1.8762 | 5.0 | 102715 | 1.7495 | 38.1038 | 16.8408 | 31.2435 | 31.2985 | 19.628 |
| 1.8289 | 6.0 | 123258 | 1.7471 | 38.0699 | 16.8839 | 31.2437 | 31.2997 | 19.6085 |
| 1.7968 | 7.0 | 143801 | 1.7422 | 38.2211 | 17.0202 | 31.3489 | 31.4019 | 19.7123 |
| 1.7284 | 8.0 | 164344 | 1.7443 | 38.384 | 17.0734 | 31.4734 | 31.5214 | 19.6688 |
| 1.7131 | 9.0 | 184887 | 1.7445 | 38.3076 | 17.0543 | 31.4168 | 31.4559 | 19.6562 |
| 1.6869 | 10.0 | 205430 | 1.7450 | 38.4651 | 17.181 | 31.5511 | 31.6061 | 19.6752 |
Framework versions
- Transformers 4.13.0
- Pytorch 1.13.1+cpu
- Datasets 2.8.0
- Tokenizers 0.10.3