LoadLoraModel - ComfyUI Built-in Node Documentation

Apply a stack of LoRAs to a diffusion model in a single node. Each row of loras holds a LoRA file, its strength and an on/off switch, and the rows are applied from top to bottom so every row patches the result of the row above it. Use this node instead of chaining several single LoRA loaders when a workflow applies a long list of LoRAs to the same model.

Inputs

Parameter	Description	Data Type	Required	Range
`model`	The diffusion model the LoRAs will be applied to.	MODEL	Yes	-
`loras`	Growable group of LoRAs, applied to the model in row order (`loras.0`, `loras.1`, and so on). Add one row per LoRA; each row holds a file, a strength and an on/off switch.	DYNAMIC_GROUP	Yes	1 to 20 rows

`loras` Row Fields

Every row repeats the following fields, and each field is required within a submitted row.

Parameter	Description	Data Type	Required	Range
`lora_name`	The name of the LoRA file to apply.	COMBO	Yes	Multiple options available
`strength`	How strongly to apply this LoRA. `0` turns it off, and a negative value inverts the effect. (default: 1.0)	FLOAT	Yes	-100 to 100 (step 0.01)
`enabled`	Turn off to skip this LoRA without changing its file or strength. (default: true)	BOOLEAN	Yes	false / true

Parameter Constraints

Row count: at least one row must be submitted and at most 20 rows are accepted, so the highest row index is 19.
Skipped rows: a row is skipped when its file is empty, when enabled is off, or when strength is 0. A negative strength is passed through rather than skipped.
Row order: rows are applied in the order they appear, and each row starts from the model returned by the previous row.

Outputs

Output Name	Description	Data Type
`MODEL`	The diffusion model after the LoRA rows that are not skipped have been 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): a656bba0248d2f6d4eb65e15e3a19f2e76edecd4b34710921a02f3ba5c598e1d

​Inputs

​loras Row Fields

​Parameter Constraints

​Outputs

Inputs

`loras` Row Fields

Parameter Constraints

Outputs