Skip to content

stijn-q42/FastChat

 
 

Repository files navigation

Fastchat

This is a fork of lm-sys/FastChat, a handy tool for locally running Large Language Models. Into the inference I've hacked internet access. This was not intended for real use, but as a tech demo that was slapped together on a wednesday night.

Changes compared to the original repo:

  • Vicuna model system prompt specifies how to use the internet
  • inference.py chat_loop is heavily modified to interupt excecution and fetch internet information
  • find_website.py was added for retrieving information from web.

Start guide and more can be found in FASTCHAT_README.md

About

A fork of fastchat with internet abilities hacked into it.

Resources

License

Stars

0 stars

Watchers

0 watching

Forks

Releases

No releases published

Packages

 
 
 

Contributors

Languages

  • Python 93.4%
  • JavaScript 2.5%
  • HTML 1.9%
  • Shell 1.7%
  • Other 0.5%