Baseline for the preprint Unlimiformer: Long-Range Transformers with Unlimited Length Input.

This model was finetuned from a BART-base model as a baseline. It was finetuned on the dataset BookSum (full-book setting).