currently testing my first Flux Merge. Diffusion model -> goes in [diffusion_models] folder

comfyUI workflow example: https://civarchive.com/articles/20235

Credits:

Forest Labs

Chroma team

Flux Krea team

Models versions merged that was sourced from the internet and fine tuned.

Makes Good images with 20 steps or less.

I taken care to insure this model is [SafeForWork], but it might give soft nudes if you prompt for nudity. this model is focused on creative realism and to move away from the generic flux plastic look.

Highly recommend using this lora, Its a game changer! https://civarchive.com/models/686704/flux-dev-to-schnell-4-step-lora

The following settings and worked well for when testing this model:

Simple image: try 12 steps with config set at 4

Personally I use 20 steps with config set at 3

30 steps with config set at 2.8

40 steps with config set at 2.5

60 steps with config set at 2

120 steps with config set at 1.8

180 steps with config set at 1.5

Sampler i used for testing :

ddim with scheduler ddim_uniform

or

Euler with scheduler Beta

If you like this model, please consider uploading your generated image's here or tipping some buzz, it helps me make new content: Thank you.