gwhitez / Lora-Trainer-XL

Modified XL Lora trainer from https://colab.research.google.com/github/hollowstrawberry/kohya-colab/blob/main/Lora_Trainer_XL.ipynb#scrollTo=OglZzI_ujZq-
1 stars 0 forks source link

New problem #1

Open eniiaflannery opened 4 hours ago

eniiaflannery commented 4 hours ago

Ok, this fix lora trainer XL for pony v6 gave me errors at the beginning

I will put the errors that I have given and the solution that has given me gemini, ia of google collab if someone helps the information

ModuleNotFoundError: No module named 'diffusers'

!pip install diffusers==0.18.2

AttributeError: module 'jax.random' has no attribute 'KeyArray'

!pip install jax==0.4.23 jaxlib==0.4.23

ModuleNotFoundError: No module named 'voluptuous'

!pip install voluptuous==0.13.1

ImportError: StableDiffusionXLPipeline requires the invisible-watermark library but it was not found in your environment. You can install it with pip: pip install invisible-watermark>=2.0

!pip install invisible-watermark==0.2.0

The following error occurred when I used the prodigy optimizer

ModuleNotFoundError: No module named 'prodigyopt'

!pip install prodigyopt==1.0

Right now, it is working fine, I will come back when I have tested the trained lora model.

after try the model works fine

gwhitez commented 3 hours ago

everything works perfect with prodigy, maybe it could be a colab error, sometimes it happens.

https://github.com/user-attachments/assets/13257ab7-828c-4bb2-80af-88d06f0524de