-
Notifications
You must be signed in to change notification settings - Fork 31
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Add GitHub Models support #113
Conversation
2af31f7
to
95029aa
Compare
Some changes to see what CI tells me. EDIT: Nice, but Ellama still tells me that the request was timed out. Here's how I configured it: (use-package ellama
:bind ("C-z ," . ellama-transient-main-menu)
:demand t
:ensure t
:init
(setopt ellama-language "English")
(require 'llm-github)
(setopt ellama-provider
(make-llm-github
:chat-model "meta-llama-3.1-405b-instruct"
:key "GITHUB_TOKEN"
:embedding-model "text-embedding-3-large"
:url "models.inference.ai.azure.com/"))
(setopt ellama-summarization-provider
(make-llm-github
:chat-model "meta-llama-3.1-405b-instruct"
:key "GITHUB_TOKEN"
:embedding-model "text-embedding-3-large"
:url "models.inference.ai.azure.com/"))
(setopt ellama-coding-provider
(make-llm-github
:chat-model "meta-llama-3.1-405b-instruct"
:key "GITHUB_TOKEN"
:embedding-model "text-embedding-3-large"
:url "models.inference.ai.azure.com/"))
(setopt ellama-naming-provider
(make-llm-github
:chat-model "meta-llama-3.1-405b-instruct"
:key "GITHUB_TOKEN"
:embedding-model "text-embedding-3-large"
:url "models.inference.ai.azure.com/"))
(setopt ellama-naming-scheme 'ellama-generate-name-by-llm)
(setopt ellama-translation-provider
(make-llm-github
:chat-model "meta-llama-3.1-405b-instruct"
:key "GITHUB_TOKEN"
:embedding-model "text-embedding-3-large"
:url "models.inference.ai.azure.com/"))) I still need to figure out how to set |
Thanks for this PR! Let me know when you get your FSF paperwork finished, and I can merge this in. If you are having any issues with it, please inform me so I can ask the maintainers. |
95029aa
to
1e3a468
Compare
On this, does it take long? I sent my request to [email protected] last saturday (2024-11-23). |
Have they replied? It usually is fast for them to reply, but from the time for them to send and receive your signed paperwork can take weeks. |
No, not yet. But I sent it during the weekend so I think it's fair of them to take their time. I'll wait and if there's no response this saturday I think I'll try to contact them again. |
Hey, just getting back on this. I just got my paper signed! If it applies beyond Emacs core source code (I need to check on ELPA for this), then this PR can be merged. |
This is fantastic, thanks for going through that! With that in mind, I've finally looked at your change and left a few comments. |
1e3a468
to
bea8036
Compare
bea8036
to
dc4073a
Compare
I'm going to merge this in now - there's still improvements I'd like to see made, but I can experiment with this and see if I can refine it further before it is release. Thank you for the contribution! |
Thanks for merging this! One thing though, apparently, the new endpoint is But I'm not that sure about this, as I've been using the configuration I mentioned in this PR with |
Github Models
Some points I'd like to address here:
With your experience I think you could easily spot the issue here, but I'll take a look at the other providers to see if I can fix it myself.
My gptel configuration, which works with this provider:
gptel-api-key
gets the key from.authinfo.gpg
, formated as:machine models.inference.ai.azure.com login apiKey password GITHUB_TOKEN