Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Tutorial on best practices for training video loras (hunyuan example) #105

Open
paul-lupu opened this issue Feb 27, 2025 · 1 comment
Open

Comments

@paul-lupu
Copy link

Heyo,

Spoke with someone that did a lot of really really high quality loras and he was gracious enough to summarize our talk in this article: https://civitai.com/articles/11942/training-a-lora-the-right-way

The stuff he does is amazing with no hand issue, artifacting etc so definitely recommend giving it a read;

@Sarania I know you mentioned you're also having issues.

Also one thing i don't think he mentioned: if you want to stack loras, it's easier if they are dim 16 instead of 32.

@Sarania
Copy link

Sarania commented Feb 27, 2025

Hey I appreciate it. Most of that was already stuff I was doing, though I am training at a significantly higher resolution than is suggested there so I might try lowering it. I think my LR has been too high, and I've been experimenting with a lower LR but LoraPlus and that might have improved it. I'm not really having hand issues anymore(lowering LR helped a lot there, also sticking with shift/7 for video training) but issues in fine detail separation.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

2 participants