ModelMergeKrea2 - ComfyUI Built-in Node Documentation

This node merges two models by blending their internal components at a fine-grained level, allowing you to control how much of each model’s specific parts influence the final result. It works by taking two input models and applying separate blend ratios to different sections of their architecture.

Inputs

Parameter	Description	Data Type	Required	Range
`model1`	The first model to merge	MODEL	Yes	-
`model2`	The second model to merge	MODEL	Yes	-
`first.`	Blend ratio for the first layer block (default: 1.0)	FLOAT	Yes	0.0 to 1.0 (step: 0.01)
`tmlp.`	Blend ratio for the time MLP block (default: 1.0)	FLOAT	Yes	0.0 to 1.0 (step: 0.01)
`txtmlp.`	Blend ratio for the text MLP block (default: 1.0)	FLOAT	Yes	0.0 to 1.0 (step: 0.01)
`tproj.`	Blend ratio for the time projection block (default: 1.0)	FLOAT	Yes	0.0 to 1.0 (step: 0.01)
`txtfusion.layerwise_blocks.0.`	Blend ratio for the first text fusion layerwise block (default: 1.0)	FLOAT	Yes	0.0 to 1.0 (step: 0.01)
`txtfusion.layerwise_blocks.1.`	Blend ratio for the second text fusion layerwise block (default: 1.0)	FLOAT	Yes	0.0 to 1.0 (step: 0.01)
`txtfusion.projector.`	Blend ratio for the text fusion projector block (default: 1.0)	FLOAT	Yes	0.0 to 1.0 (step: 0.01)
`txtfusion.refiner_blocks.0.`	Blend ratio for the first text fusion refiner block (default: 1.0)	FLOAT	Yes	0.0 to 1.0 (step: 0.01)
`txtfusion.refiner_blocks.1.`	Blend ratio for the second text fusion refiner block (default: 1.0)	FLOAT	Yes	0.0 to 1.0 (step: 0.01)
`blocks.0.` through `blocks.27.`	Blend ratio for each of the 28 main blocks (default: 1.0)	FLOAT	Yes	0.0 to 1.0 (step: 0.01)
`last.`	Blend ratio for the last block (default: 1.0)	FLOAT	Yes	0.0 to 1.0 (step: 0.01)

Each blend ratio controls how much of model2 is used for that specific component, where 0.0 means use only model1, 1.0 means use only model2, and values in between create a weighted blend.

Outputs

Output Name	Description	Data Type
`MODEL`	The merged model resulting from blending the two input models according to the specified ratios	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): ece35524f77fd906dc3109a0818d4d7d3986ec9debb518fd04893048843d7e88

​Inputs

​Outputs

Inputs

Outputs