I want to duplicate model ‘Wan2.1 I2V 14B’ to paid spaces hardware. In order to get the same speed as on zero gpu (which is like generating 5sec clip in 50sec) what hardware do i need to choose. Being a novice I’m assuming paid spaces are dedicated so will it work on lower end Nvidia GPUs. Please advice which GPU is recommended. Thanks
1 Like
Zero GPU is fast because it shares H200 between users. If you want a GPU with the same speed, this might be it…
Nvidia H100 24 vCPU 250 GB 80 GB 3000 GB $10.00
Thanks John, I’ll have a look at it..
1 Like
If I deploy it at inference endpoint.. its not there in their catalog. And if I’m deploying it through Hugging Space it gives an error: Warning: deploying this model will probably fail because no “handler.py” file was found in the repository. Try selecting a different model or creating a custom handler.
H200 is also there for $5 but again it says 0 quota.
With paid hardware spaces H200 is not listed and if i choose any other again it gives this warning: Your duplicated Space may not work if you switch to a different hardware than the suggested one.
I dont want to for ZeroGpu with restricted usage.
HF is just so confusing…I think i’ll have to deploy it on some other pod.
1 Like
If you want to deploy via API, I think the Diffusers version of them are more convenient.
1 Like