A LoRA extracted from the FluxUnchained model by socalguitarist. Credit goes to him for creating this model.
Why I created this LoRA :
To use on CivitAI generation services, since base models are not supported yet.
To control the strength.
To combine or merge with other LoRAs and base (dev/schnell) models.
Rank128 is chunky but will give you more details on the genitals (Recommended for NSFW and CivitAI).
Rank16 will save space but will provide less detail, which can be mitigated to an extent by increasing the strength of the LoRA (~1.50).
Description
FAQ
Comments (4)
so Flux Unchained is a lora merge, and not a fine tune? in other words: will i get the same results with this lora + flux dev as with the FU checkpoint?
Yes, you should get at least similar results if not exactly the same
Flux Unchained says it's trained, not a LoRA merge. You can extract a LoRA from every checkpoint, no matter how it was created. But depending on the checkpoint and the size of the extracted LoRA, the fidelity might not be perfect.
This ones actually good. Fairly flexible and seems to work well with other loras so far.
This thing works insanely well with consistency v32. It fills in a lot of the gaps.