Alt4OneFLUX
This version is built on from my Alt4OneXL and Alt4OnePony models datasets.
Mileage may vary...
I'm still testing the different epochs. But adjusting the LoRa wight between 1 and 2.5 give a variety of different results, 3 is cursed. Adjusting the Distilled CFG Scale between 3.5 and 7 give a good variety of results too. So Play around and see what works best for you.
Epoch 5 seems like a good middle ground to work with. Epoch 7 Also gives some good results too.
Tested Models:
flux1-dev-bnb-nf4-v2
flux1-salto-v1.3-bnb-nf4
fluxunchainedNF4_unchained11NF4Unet



















