another one...trying some stuff with the LoRA training process to see what i can do and what I can't.just see this as an experiment.
checkpoint-e18_s342.safetensors