bart-base-finetuned-cnn-news
This model is a fine-tuned version of facebook/bart-base on the cnn_dailymail dataset. It achieves the following results on the evaluation set:
- Loss: 2.8560
- Rouge1: 21.8948
- Rouge2: 9.7157
- Rougel: 17.9348
- Rougelsum: 20.5347
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: 0.00056
- train_batch_size: 8
- eval_batch_size: 8
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 5
Training results
Training Loss | Epoch | Step | Validation Loss | Rouge1 | Rouge2 | Rougel | Rougelsum |
---|---|---|---|---|---|---|---|
3.7005 | 1.0 | 718 | 2.9872 | 21.7279 | 9.0406 | 17.392 | 20.0627 |
2.937 | 2.0 | 1436 | 2.8590 | 21.3056 | 8.5254 | 17.2338 | 20.0403 |
2.2642 | 3.0 | 2154 | 2.6744 | 21.277 | 9.6162 | 17.7775 | 20.1688 |
1.5774 | 4.0 | 2872 | 2.7020 | 21.7458 | 9.846 | 18.1649 | 20.7067 |
1.0174 | 5.0 | 3590 | 2.8560 | 21.8948 | 9.7157 | 17.9348 | 20.5347 |
Framework versions
- Transformers 4.27.2
- Pytorch 1.13.1+cu117
- Datasets 2.11.0
- Tokenizers 0.13.3
- Downloads last month
- 2,288
Inference API (serverless) is not available, repository is disabled.
Dataset used to train hardikJ11/bart-base-finetuned-cnn-news
Evaluation results
- Rouge1 on cnn_dailymailvalidation set self-reported21.895