search

cycneuramus / signal-aichat

An AI chatbot for Signal powered by Google Bard, Bing Chat, ChatGPT, HuggingChat, and llama.cpp
GNU Affero General Public License v3.0
80 stars 15 forks source link

Allow specifying the gpt model #15

Closed PaulWoitaschek closed 1 year ago

PaulWoitaschek commented 1 year ago

I'd like to use gpt4. Currently, gpt3 is hardcoded https://github.com/cycneuramus/signal-aichat/blob/cf05eaf8dee951d68d1f5509d39568b2b343f0bd/ai.py#L132

It would be great to have an option to select which model to use.

cycneuramus commented 1 year ago

Should be fixed in 0.6.0

PaulWoitaschek commented 1 year ago

😢 That could have been my first python PR ^^

thanks!

PaulWoitaschek commented 1 year ago

@cycneuramus

Does it deploy automatically? A compose pull isn't fetching anything and it seems nothing is deployed in ghrc https://github.com/cycneuramus/signal-aichat/pkgs/container/signal-aichat/versions?filters%5Bversion_type%5D=tagged

cycneuramus commented 1 year ago

It should deploy automatically, but it seems broken ATM. I'll take a look as soon as I have time.

cycneuramus commented 1 year ago

The new version should be available now on the master tag (docker pull ghcr.io/cycneuramus/signal-aichat:master)

PaulWoitaschek commented 1 year ago

Thanks. Should we update the compose file to point to that? Right now with what's recommended in the compose file it errors with

Head "https://registry.gitlab.com/v2/signald/signald/manifests/latest": received unexpected HTTP status: 503 Service Unavailable