Closed syonfox closed 4 months ago
pdwittig
commented
10 months ago Hey @syonfox. We don't currently support bring your own model, but likely will in the mid-term future. Out of curiosity, what kind of custom models are you using and how does that fit into your product/workflow/use case.
Also, we're in early innings and dont have anything published re infra/performance, but curious what you're looking for there as well.
Lmk. Thanks.
davystrong
commented
8 months ago @pdwittig can you give any indication of when this is likely to be supported?
syonfox
commented
8 months ago I was thinking of the instruct embedding model for my platform as well as hosting domain specific fine tuned lama models.
It would also be usefully to have a yolo based image classifier and general purpose object detection. A little more notch but a segmentation model would also be useful
On Thu., Jan. 4, 2024, 15:39 David Armstrong, @.***> wrote:
@pdwittig https://github.com/pdwittig can you give any indication of when this is likely to be supported?
— Reply to this email directly, view it on GitHub https://github.com/cloudflare/cloudflare-docs/issues/11039#issuecomment-1876831911, or unsubscribe https://github.com/notifications/unsubscribe-auth/ACHJVDLCPUY4W4VC52UGKCDYMZ5V7AVCNFSM6AAAAAA5PP6IYGVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMYTQNZWHAZTCOJRGE . You are receiving this because you were mentioned.Message ID: @.***>
jvel07
commented
5 months ago @pdwittig any dates on when custom models will be supported? I'd like to host a generative app but not sure how to use GPU resources on Cloudfare other than workers AI, which currently, doesn't support custom ML inference pipelines.
kodster28
commented
4 months ago Howdy @syonfox, checked internally and these are on the roadmap... but no firm timeline.
Closing out this issue specifically, but keep an eye on the Workers AI changelog and Changelog RSS feed to see when it hits.
Which Cloudflare product(s) does this pertain to?
Workers
Subject Matter
AI workers
Content Location
https://github.com/cloudflare/cloudflare-docs/blob/production/content/workers-ai/models/embedding.md
Additional information
Hello I'm exited to play around with the new workers AIon Edge GPU
can you point me in the right direction for any blog on the infrastructure and performance of your implementation?
I'm also wondering if there are plans for allowing users to create their own ai models rather than use a list of provided ones.
Ideally, we could define a repository or module and deploy it to Cloudflare before then being able to use it in our workers without needing a backend at all :)
How to deploy a custom model / explicitly state it's not possible yet but planned ;)
Thanks for the awesome platform and all your great work, this is definitely revolutionizing the way we create high-performance scalable apps
Happy coding!