ModelSamplingFlux - ComfyUI Built-in Node Documentation - ComfyUI

The ModelSamplingFlux node applies Flux model sampling to a given model by calculating a shift parameter based on image dimensions. It creates a specialized sampling configuration that adjusts the model’s behavior according to the specified width, height, and shift parameters, then returns the modified model with the new sampling settings applied.

Inputs

Parameter	Description	Data Type	Required	Range
`model`	The model to apply Flux sampling to	MODEL	Yes	-
`max_shift`	Maximum shift value for sampling calculation (default: 1.15)	FLOAT	Yes	0.0 - 100.0 (step 0.01)
`base_shift`	Base shift value for sampling calculation (default: 0.5)	FLOAT	Yes	0.0 - 100.0 (step 0.01)
`width`	Width of the target image in pixels (default: 1024)	INT	Yes	16 - MAX_RESOLUTION (step 8)
`height`	Height of the target image in pixels (default: 1024)	INT	Yes	16 - MAX_RESOLUTION (step 8)

max_shift and base_shift are advanced parameters. The shift applied to the sampling configuration is calculated from the image dimensions: the latent resolution is computed as width × height / 256, and the shift value is interpolated between base_shift at a latent resolution of 256 and max_shift at a latent resolution of 4096.

Outputs

Output Name	Description	Data Type
`model`	The modified model with Flux sampling configuration applied	MODEL

This documentation was AI-generated. If you find any errors or have suggestions for improvement, please feel free to contribute! Edit on GitHub

Source fingerprint (SHA-256): 04065b54ace30a2b20476ed085df871ea89794650e98ae30c40f750357663834