I’m running a space on an H100, and I’ve been getting a scheduling failure error. When I try to restart the space, I just get a 500 internal error. I tried switching to a ZeroGPU space and then back to H100, but now the space is stuck as a ZeroGPU space and it won’t turn back into an H100 space.
Are there steps I can take to restore the space on an H100 instance?