search

JudiniLabs / code-gpt-docs

Docusaurus page
https://code-gpt-docs.vercel.app
MIT License
582 stars 63 forks source link

Autocomplete feature with OpenAI Key #354

Open lgcCerti opened 3 months ago

lgcCerti commented 3 months ago

Hello,

I'm interested in the autocomplete feature but I'd like to use gpt-4o mini which I have a key for. There only seems to be support for chat and some chat shortcuts.

Will autocomplete support this in the future? Currently I can only use ollama, mistral, and the CodeGPT.

Adel242 commented 2 months ago

Indeed @lgcCerti , as you've pointed out, those are the only technologies currently operational in our autocomplete feature. However, we are certainly open to the possibility of incorporating GPT-4 in the future. We greatly appreciate your feedback and encourage you to reach out to us with any further queries.

lgcCerti commented 2 months ago

Indeed @lgcCerti , as you've pointed out, those are the only technologies currently operational in our autocomplete feature. However, we are certainly open to the possibility of incorporating GPT-4 in the future. We greatly appreciate your feedback and encourage you to reach out to us with any further queries.

Hi Adel, thanks for returning to me

I see now that autocomplete actually makes little sense for API keys, since they cost money per token, and can be used with smaller local models running on CPU so this request was unfounded and not necessary for code-gpt.

Recently I've been introduced to the continue.dev extension and it had some of the co-pilot features I looked for (enter a prompt on the text editor and get the code in the IDE). Its onboarding process made it a little more obvious what models were being used for what (1 for chat, 1 for autocomplete, 1 for embedding) and that turned out to be main thing I was missing but codegpt didn't make too clear.

Perhaps codegpt already had all this and more but unfortunately I was mislead by the UI changes.