search

UW-Madison-DSI / ask-xDD

Retrieval-Augmented Generation (RAG) on 17M full text journal articles.
https://xdd.wisc.edu/
MIT License
2 stars 2 forks source link

Review available open-source model for document labeling/extraction linking. #89

Closed JasonLo closed 9 months ago

JasonLo commented 9 months ago

Reasonable size (7b) open source best LLM. This is useful if we want to apply fact-sheet type chat-bot extraction within a single doc.

Somewhat popular summarizers, only provides generic summarization.

JasonLo commented 9 months ago

More candidates (non-chatbot):

lay summary specifically for scientific aritcles, code

TransformerSum library, with extractive and abstractive summarization

transformer lib

JasonLo commented 9 months ago

Preliminary hand testing shows that summarization models doesn't works as well as chatbots.

JasonLo commented 9 months ago

Let's use OpenChat3.5 for now.