Spaces:
Running
on
Zero
Apply for community grant: Academic project (gpu)
Hi @nupurkmr9 , your Space is already running on ZeroGPU, so it seems that you don't need a grant.
Hi, yes, I upgraded the space to run on ZeroGPU since the minimum GPU requirement was > 30GB. Thanks a lot!
@nupurkmr9
Can you uncomment this line and upgrade the sdk_version
to the latest 5.17.1
in the README.md?
In the case of ZeroGPU Spaces, ZeroGPU backend moves the model to GPU when needed and offload it to CPU after a certain time of inactivity, but it determines what should be moved to GPU based on whether .to("cuda")
is called, so without it, the inference is run on CPU. As for the gradio version, there was an issue related to handling ZeroGPU quota in some old versions of gradio, so it would be nice to upgrade to the latest version.
Thanks. I was wondering why I am getting issues with ZeroGPU usage :)
Hi
@nupurkmr9
, this is Yuvi from the gradio team here at Huggingface. Congrats on launching the app and securing the GPU grant!
It would be wonderful if you could add the Gradio code to your repository, along with instructions on how to run the app in your project readme. This way, the community can easily launch the SynCD app on their own machines.
Is there some access limit on GPU per hour?