-
Notifications
You must be signed in to change notification settings - Fork 570
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
improve: improve support for pyannote.audio #688
Conversation
Hey there! Sorry for the slow response. As of now, the code snippet is based in the tag of a model. So if a model repo has Instead, I would suggest the two code types directly in the same snippet, although my concern would be of the sippet being too long. Maybe we can simplify and just do inference on the whole file. It's not like the transformers code snippets have a full example. WDYT? |
Are you suggesting to only stick with the |
Maybe just to take a step back and understand the usage. Can you use both Once you have a repo with the tag |
No, because a For instance,
Yes, that was the point of this PR :) If that is too much trouble, we could focus on pipelines ( |
Alright, much clearer now. My suggestion is then to do the following
This will allow to have a single filter for all
Do you think this would solve the need? |
Perfect. Will make the changes. |
Done. |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Thank you very much for the PR! There are some unrelated changes, would you be able to remove them?
Afterwards we can go ahead and merge
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
LGTM! Thanks 🚀
Thanks! Any update on the "audio-segmentation" widget? ;) |
Also, out of curiosity, when will this be deployed? |
Hey there! Sorry, we're doing a repo cleanup/split to make things much much much better organized, so this is taking a bit longer to get deployed. I'll update you as soon as we do this. |
The change was also applied in huggingface/hub-docs@e00b7bd, this should be deployed soon. |
@Narsil we'll need a deploy of this change in api-inference updating the framework name, which requires a small internal PR iirc. Could you help us with this? |
pyannote.audio has two main types of "inference":
I think both have their own use and target different users (e.g. models for researchers, pipelines for users who just want to get speaker diarization results).