GTP4All

[marcusquinn]

• https://github.com/nomic-ai/gpt4all
• https://s3.amazonaws.com/static.nomic.ai/gpt4all/2023_GPT4All_Technical_Report.pdf

Demo, data, and code to train an assistant-style large language model with ~800k GPT-3.5-Turbo Generations based on LLaMa

For use with: https://forum.cloudron.io/topic/8984/gpt4all-ui

[robi]

This is now available for multiple platforms so you can run it on your own PC & Mac.

There is also the unfiltered model available which has all the rejection messages removed for censored topics.

Then choose one of the UIs for it or enjoy it in CLI.

[marcusquinn]

Now available with native app installers: https://gpt4all.io

[LoudLemur]

Haha! I found this thread again. There is also gpt4all-j

https://github.com/marella/gpt4all-j
https://pypi.org/project/gpt4all-j/

Which I mention because this project, PrivateGPT, enables you to feed pdf/epub/etc to your local llama and train it:

https://github.com/imartinez/privateGPT

PrivateGPT requires the language models you use to be compatible with gpt4all-j

https://huggingface.co/nomic-ai/gpt4all-j

PrivateGPT seems to have quite a lot of issues and I haven't been able to get it to run, but maybe some brilliant person here might be able to do so.

Would gpt4all-j be a better candidate for Cloudron than gpt4all?

[marcusquinn]

@LoudLemur Don't worry, give it a while and someone will make a GPT refrigerator.