---
language: 
  - en
tags:
- summarization
- diff generation
datasets:
- nonsense corpus
metrics:
- rouge
---

hello! this is the pretrained BART. The dataset used for pretraining is nonsense summary corpus with output as diff.