bart-base-finetuned-multinews
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: 2.4152
- Rouge1: 14.6798
- Rouge2: 5.2044
- Rougel: 11.2346
- Rougelsum: 12.9794
- Gen Len: 20.0
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: 5
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Rouge1 | Rouge2 | Rougel | Rougelsum | Gen Len |
|---|---|---|---|---|---|---|---|---|
| 2.8162 | 1.0 | 506 | 2.4807 | 14.5888 | 4.9839 | 11.0896 | 12.9 | 20.0 |
| 2.6122 | 2.0 | 1012 | 2.4371 | 14.9075 | 5.3211 | 11.2711 | 13.1998 | 20.0 |
| 2.518 | 3.0 | 1518 | 2.4141 | 14.8607 | 5.2903 | 11.332 | 13.1363 | 20.0 |
| 2.4585 | 4.0 | 2024 | 2.4246 | 14.7346 | 5.2263 | 11.2281 | 13.0277 | 20.0 |
| 2.4206 | 5.0 | 2530 | 2.4152 | 14.6798 | 5.2044 | 11.2346 | 12.9794 | 20.0 |
Framework versions
- Transformers 4.40.1
- Pytorch 1.13.1+cu117
- Datasets 2.19.0
- Tokenizers 0.19.1