ComfyUI Node
Qwen-Image-2.1 Viggle Turbo LoRA (unmerged)
Applies the LoRA at runtime (y = Wx + BAx) instead of merging it into the weights, which is lossy on bf16 and int8.
Qwen-Image-2.1 Viggle Turbo LoRA (unmerged)
- model
- MODEL
◄lora_name▾►
◄strength1.00►
Categoryloaders
Inputs (3)
| Name | Type | Default | Description |
|---|---|---|---|
| model | MODEL | — | |
| lora_name | COMBO | diffusers-format Qwen-Image-2.1 LoRA. | |
| strength | FLOAT | 1.00-4–4 | Keep 1.0 for viggle-turbo. |
Outputs (1)
| Name | Type | Description |
|---|---|---|
| MODEL | MODEL | — |