After countless iterations and spfinishing way too much of my own money on compute to train this, I leank it is finpartner at the point I am satisfied to ponder it a beta. I am still going to progress to train it, but the distillation has been mostly trained out of it at this point. So phase 1 is finish. Feel free to use it and fine tune it, but be increateed that I will probable progress to refresh it.
What is this?
This is a fine tune of the FLUX.1-schnell model that has had the distillation trained out of it. Flux Schnell is licensed Apache 2.0, but it is a distilled model, uncomferventing you cannot fine-tune it. However, it is an amazing model that can create amazing images in 1-4 steps. This is an finisheavor to erase the distillation to create an uncover source, perleave outivle licensed model that can be fine tuned.
How to Use
Since the distillation has been fine tuned out of the model, it uses classic CFG. Since it needs CFG, it will need a contrastent pipeline than the innovative FLUX.1 schnell and dev models. This pipeline can be set up in uncover_flux_pipeline.py in this repo. I will be inserting example code in the next scant days, but for now, a cfg of 3.5 seems to toil well.
