Closed Yigit20 closed 1 year ago
You are not going to be able to train on a GPU with only 4GB of VRAM. You can train a LoRA with about 8GB of VRAM, but Dreambooth itself requires around 24GB of VRAM (and even more if you use the Adan D-Adaptation optimizer).
Is there an existing issue for this?
What happened?
I am using SD for a while with 1650 Super, i know that it's not enough for the Sd. Although, i did create lots of images with it. But when i try to Train with Dreambooth that error happens. I tried a lot of things but couldn't help it. I don't think this is about the GPU because ı read some people manage to train with 1650 super.
Steps to reproduce the problem
I did create a model then i set the setting and dataset directory. And after all class images were created this error showed up.
Commit and libraries
[+] xformers version 0.0.17 installed. [+] torch version 2.0.1+cu118 installed. [+] torchvision version 0.15.2+cu118 installed. [+] accelerate version 0.19.0 installed. [+] diffusers version 0.16.1 installed. [+] transformers version 4.29.2 installed. [+] bitsandbytes version 0.35.4 installed.
Command Line Arguments
Console logs
Additional information
No response