ComfyUI Node
Lora Smart Combine (QQ)
Smartly combine multiple LoRAs with rank extension using SVD decomposition
Lora Smart Combine (QQ)
◄lora_1_name▾►
◄lora_2_name▾►
◄lora_3_name▾►
◄lora_4_name▾►
◄target_rank16►
◄blend_weight_11.0►
◄blend_weight_21.0►
◄blend_weight_31.0►
◄blend_weight_41.0►
◄alpha_handlingblend_proportional►
◄output_dtypematch_original►
◄verbosetrue►
CategoryKJNodes/lora
Inputs (12)
| Name | Type | Default | Description |
|---|---|---|---|
| lora_1_name | COMBO | First LoRA to combine | |
| lora_2_name | COMBO | Second LoRA to combine | |
| lora_3_name | COMBO | Third LoRA (optional) | |
| lora_4_name | COMBO | Fourth LoRA (optional) | |
| target_rank | INT | 161–4096 | Target rank for the combined LoRA |
| blend_weight_1 | FLOAT | 1.00–10 | Blend weight for first LoRA |
| blend_weight_2 | FLOAT | 1.00–10 | Blend weight for second LoRA |
| blend_weight_3 | FLOAT | 1.00–10 | Blend weight for third LoRA |
| blend_weight_4 | FLOAT | 1.00–10 | Blend weight for fourth LoRA |
| alpha_handling | COMBO | blend_proportional | How to handle alpha values when combining |
| output_dtype | COMBO | match_original | Data type for output LoRA |
| verbose | BOOLEAN | true | Print detailed combination information |
Outputs (0)
No outputs