LoraReduceRank

KJNodes/loraExperimentalOutput Node

LoraReduceRankKJ

Resize a LoRA model by reducing it's rank. Based on kohya's sd-scripts: https://github.com/kohya-ss/sd-scripts/blob/main/networks/resize_lora.py

Experimental: This node is experimental and its behavior may change without notice.

Example

JSON Example

{
  "class_type": "LoraReduceRankKJ",
  "inputs": {
    "lora_name": "https://civitai.com/api/download/models/123456?type=Model&format=SafeTensor",
    "new_rank": 8,
    "dynamic_method": "disabled",
    "dynamic_param": 0.2,
    "output_dtype": "match_original",
    "verbose": true
  }
}

This example shows required inputs only. Connection values like ["node_id", 0] should reference actual node IDs from your workflow.

Inputs

Name	Type	Status	Constraints	Default
lora_name?	ENUM 0 options URL: LoRA	required	-	-
new_rank?	INT	required	min: 1, max: 4096, step: 1	8
dynamic_method?	ENUM 4 options disabled sv_ratio sv_cumulative sv_fro	required	-	"disabled"
dynamic_param?	FLOAT	required	min: 0, max: 1, step: 0.01	0.2
output_dtype?	ENUM 4 options match_original fp16 bf16 fp32	required	-	"match_original"
verbose	BOOLEAN	required	-	true

Outputs

This node has no outputs.

Package: comfyui-kjnodes