The experiment can be considered a failure.
I am leaving the model for review. I am not deleting it so that no one else posts it again and wastes time and credits on it.
---------------------------------------------------
Flux1 Dev2Pro specically for training Lora. On the face, style or subject.
Inspired by the technology described here:
https://medium.com/@zhiwangshi28/why-flux-lora-so-hard-to-train-and-how-to-overcome-it-a0c70bc59eaf
For objectivity, write your full-fledged reviews in the comments. Does Lora work better than on the standard model Flux1.Dev?
Reprint from the reprint - the author is unknown