-
Notifications
You must be signed in to change notification settings - Fork 4.2k
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
Lora keys not loaded for SD3-Medium Lora #3701
Comments
Can you post a lora file in that format? |
Sure thing! For the record it was trained with this https://github.com/huggingface/diffusers/blob/main/examples/dreambooth/README_sd3.md |
@comfyanonymous added another file with just 100 steps, its the final output from diffusers, rank 128 I think this was, so I guess this file should be "complete" as a I noticed the checkpoints might be missing the TENC |
Not sure what your lora is supposed to do so let me know if this works. |
Its supposed to be an Anime style, and while it is botched, it loads and changes the image, thank you! |
@comfyanonymous I've tried a couple of LoRAs with the new code and they still give me the same error as #3701 (comment) |
Same, I've tried with the new code with a couple of LoRAs and nothing changed. Do I need to restart the UI or recompile anything? |
Actually I found that some lora trainers use different keys, (simpletuner) it's not the same error, and when I fixed those by detecting those keys as well (I've not done a pull request) I now get a tensor size error. |
Here's a portion of my errors, in case it helps:
|
If you get warnings about _unet_ keys it means that lora is for an older SD version not SD3. |
As the title says, Loras trained with Diffusers' scripts can't be loaded with stock Lora Loader
The text was updated successfully, but these errors were encountered: