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

Blurry images since today #238

Open
lhovav opened this issue Dec 18, 2024 · 3 comments
Open

Blurry images since today #238

lhovav opened this issue Dec 18, 2024 · 3 comments

Comments

@lhovav
Copy link

lhovav commented Dec 18, 2024

Hi,

I'm getting blurry sample images every 250 images,

started today:
What I did that I think it's a bug:

  1. tried different seed - still blurry
  2. tried two different datasets from yesterday (which were perfect) - still blurry,
  3. deployed another pod (I'm running RunPod) - still blurry

also, the linkess is meh, compared to yesterday (the .safetensor doesn't produce blurry images on local ComfyUI)
blurry

Everything is the same, I'm using the same config.yaml file, the same file structure, just changing the lora's name.

does anyone experience this issue too?

Thanks

@steveant
Copy link

steveant commented Dec 28, 2024

I've had the same issue. Besides what you mentioned I also:

  • reinstalled locally on windows and on linux with latest (followed README.md process)
  • I noticed the samples generated have different levels of fogginess and some are consistently clear. (odd).
  • Next I'm going to try to revert to an an older changeset hoping it's an issue in the main repo and not the submodules...

FWIW - the first round of latents that get saved to output//samples before any training are foggy.

@duchamps0305
Copy link

yes here too.

@julien-blanchon
Copy link

I also have blurry image but it's mainly on white background which is a know Flux problem

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

4 participants