Hacker News Reader

I wonder if they will have cluster support, i have a few old fat servers with a few RTX spec GPU's laying around, my own fast ChatGPT running in the garage would be cool.

View on HN

This plus the the vicuna models is probably as close as you can get in the near term.

https://github.com/oobabooga/text-generation-webui

I’d say try Open Assistant imo it’s better than gpt4all interms of response quality.

Reply on news.ycombinator.com