This model is a fine-tuned version of facebook/bart-base on the arxiv_summarization_dataset dataset. It achieves the following results on the evaluation set:
More information needed
More information needed
More information needed
The following hyperparameters were used during training:
| Training Loss | Epoch | Step | Validation Loss | Rouge1 | Rouge2 | Rougel | Rougelsum |
|---|---|---|---|---|---|---|---|
| 3.433 | 1.0 | 403 | 3.0405 | 12.9591 | 4.8638 | 10.377 | 11.94 |
| 3.1155 | 2.0 | 806 | 2.9598 | 12.5926 | 4.7648 | 10.1432 | 11.6224 |
| 2.9888 | 3.0 | 1209 | 2.9456 | 12.5891 | 4.7185 | 10.1631 | 11.6233 |
pytorch_model.bin 558 MBtraining_args.bin 3.7 kB