lm-sys/FastChat

support for 4bit quantization from transfomer library.

Open

#1.798 aberto em 27 de jun. de 2023

Ver no GitHub
 (7 comments) (2 reactions) (0 assignees)Python (4.736 forks)batch import
enhancementgood first issue

Métricas do repositório

Stars
 (38.959 stars)
Métricas de merge de PR
 (Nenhuma PRs mesclada em 30d)

Description

Loading a vicuna13B using 4bit quantization from the transformers library is possible load_in_4bit. How difficult could be for Fastach to support it?

Guia do colaborador