bart-base-finetuned-xlsum-10-epoch
This model is a fine-tuned version of facebook/bart-base on the xlsum dataset. It achieves the following results on the evaluation set:
- Loss: 1.7506
- Rouge1: 38.5509
- Rouge2: 17.1804
- Rougel: 31.6297
- Rougelsum: 31.6993
- Gen Len: 19.6701
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 |
|---|---|---|---|---|---|---|---|---|
| 2.0865 | 1.0 | 19158 | 1.8344 | 36.709 | 15.4309 | 29.8542 | 29.9152 | 19.5716 |
| 1.9828 | 2.0 | 38316 | 1.7894 | 37.689 | 16.2171 | 30.7743 | 30.8346 | 19.6636 |
| 1.8778 | 3.0 | 57474 | 1.7727 | 37.5849 | 16.3555 | 30.8276 | 30.8936 | 19.5898 |
| 1.785 | 4.0 | 76632 | 1.7546 | 38.3036 | 16.911 | 31.4077 | 31.4608 | 19.5976 |
| 1.7246 | 5.0 | 95790 | 1.7505 | 38.2107 | 16.929 | 31.3889 | 31.4485 | 19.6316 |
| 1.6883 | 6.0 | 114948 | 1.7467 | 38.2416 | 17.0113 | 31.4098 | 31.4639 | 19.6048 |
| 1.6301 | 7.0 | 134106 | 1.7475 | 38.4083 | 17.1098 | 31.5605 | 31.6322 | 19.6438 |
| 1.6034 | 8.0 | 153264 | 1.7493 | 38.4812 | 17.1939 | 31.6284 | 31.6959 | 19.6574 |
| 1.562 | 9.0 | 172422 | 1.7481 | 38.5011 | 17.1622 | 31.6031 | 31.6808 | 19.7056 |
| 1.5496 | 10.0 | 191580 | 1.7506 | 38.5509 | 17.1804 | 31.6297 | 31.6993 | 19.6701 |
Framework versions
- Transformers 4.13.0
- Pytorch 1.13.1+cpu
- Datasets 2.8.0
- Tokenizers 0.10.3