I chnaged the security level to "weak" in the config.ini file and still getting the error. I also updated comfyui. anybody knows how to solve this?
I tried this model on l4 in modal and have been able to get around 12s startup and 3s inference time for every cold start after memory snapshot.
But modal's l4 provisioning kinda sucks lately.
Would runpod be a good choice here? How is its provisioning time and post optimisation snapshot speed for this particular model size?
Any kind of help or comment would be really appreciated. Thanks!