XlabsSampler
Node inputs: model: get FLUX diffusion model (from unet loader) conditioning & negconditioning: input prompts after T5 and clip models (clip only allowed, but you should know, that you will not utilize about 40% of flux power, so use dual text node) latentimage: latent input for flux, may be empty latent or encoded with FLUX AE (VAE Encode) image (for image-to-image using) controlnetcondition: input for XLabs-AI ControlNet conditioning Output: latent: FLUX latent image, should be decoded with VAE Decoder to get image Parameters: noiseseed, controlaftergenerate: controls random generator steps: how many denoise steps will diffusion go timesteptostartcfg: how many steps diffusion will do before start to use negative sampling and cfg truegs: true cfg scale, will be used after first “timesteptostartcfg” steps imagetoimagestrength: how much will the original image affect to the output denoisestrength: how many noise will remain
XLabsNodes · 5 Eingaben · 6 Parameter · 1 Ausgabe
