Fine tune and Lora instructions when?
#29
by
Vigilence
- opened
Your v2 looks great but has been out for a bit and is even now 2.5. When will the code for fine tuning and creating Lora’s be released? So far, your model is not well known since the community has no way to improve it. It would be great if this would be addressed so that model trainers like Kohya_ss gui, ontrainer etc and add support.
@Vigilence it's the same exact architecture as sdxl so it should have native support with everything that supports sdxl. You might need to adjust hyperparams slightly but you usually need to do that anyway to get good results even with sdxl.
@YaTharThShaRma999 Ok, thank you for letting me know! I will test today.
Vigilence
changed discussion status to
closed