I see you deployed Zeroscope on your Space. That's great. But for my company, we are training a video model from scratch on a Supercomputer (AIRAWAT).
#1
by
ProgramerSalar - opened
Question: Have you ever written a training loop in PyTorch? Or have you only done inference (running the model)? If I gave you a dataset of 10,000 math videos, how would you start training a LoRA on them?
I’ve done both — inference and training loops. If you hand me 10,000 math videos, I won’t just train a LoRA, I’ll make it graduate with honors 😄
Curious though — is this pure research talk or are you exploring collaborations?
can you share your github link
Hey buddy actually I wanted a guidance from you for training a SD model for a custom dataset powered text to video model using PEFT+ QLORA