Hi! I was wondering if anyone could point me to papers, blog posts, etc that explain how to ensemble previously trained models for abstractive text summarization (if possible).
Moreover, is anything like this already implemented in Huggingface?
What do you mean exactly by “ensembling” models?
There are some models available on HuggingFace, for example the BART summarization model fine-tuned on the CNN/DailyMail dataset. You can take a look at the model card and how to use it here. The implementation is very straightforward.