bmaltais
bmaltais
Don't think it is broken. Try training this config part of the repo: /test/config/dreambooth-AdamW8bit.json It is a test config I use to validate I have not broken things... and it...
Are you talking about this file: https://github.com/bmaltais/kohya_ss/blob/master/docs/LoRA/options.md
Can you share the .toml file for the training? look like the trainer is warning about radient_accumulation_steps being set to 3 and warning it will cause issues
Not sure I totally get what changes you are asking for... there is no requirements_nvidia.txt file... so what actual requirements do you want to see in it?
Which requirements file need the upgrade to tensorflow 2.16? How do you install the GUI? Do you use special parameters to specify the requirements file? I don't use linux so...
But how will you call this? Is the setup.sh going to handle this as is? I think the best would be if you create a pull request to propose all...
Something in the toml that is outputted in the outputs folder is causing the issue... can you share the toml?
 That file path is causing the issue... Not sure how to fix this... I think it is a bug in sd-scripts... nothing I can do... or maybe this is...
I pushed an updated version that write the toml as utf-8... hope this help: ``` git fetch origin git checkout dev git pull ``` Let me know how it goes.
can you share the toml file from the outputs folder? I think there is a character in there kohya's sd-scripts doe not like.