README.md

metadata

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.