search

LambdaLabsML / examples

Deep Learning Examples
MIT License
805 stars 103 forks source link

Is there a way to run the stable diffusion fine-tuning on RTX4090? For all configurations, I keep on facing cuda out of memory error. #55

Open garg-aayush opened 1 year ago

garg-aayush commented 1 year ago

I would like to fine tune Stable Diffusion for a custom dataset similar to Lambdalabs post.

However, whatever configuration I use? I keep on running the CUDA out of memory issue.

Has anyone tried to run fine-tuning on 24GB GPU say 3090 or 4090?

youluexx commented 1 year ago

The same question, if there are two 4090 graphics cards, can they be trained, and how do they need to be configured?

jnulzl commented 1 year ago

The same question