ComfyUI Node
Eric Qwen-Image Apply LoRA
A ComfyUI node in Eric Qwen-Image with 7 inputs and 1 output.
Eric Qwen-Image Apply LoRA
- pipeline
- pipeline
◄lora_name▾►
◄weight_stage10.00►
◄weight_stage20.00►
◄weight_stage30.00►
◄lora_path_override►
◄keep_loadedtrue►
CategoryEric Qwen-Image
Inputs (7)
| Name | Type | Default | Description |
|---|---|---|---|
| pipeline | QWEN_IMAGE_PIPELINE | — | |
| lora_name | COMBO | Select LoRA from ComfyUI/models/loras/ | |
| weight_stage1 | FLOAT | 0.00-2–2 | Stage 1 (draft/composition) weight. Also used as the weight for non-UltraGen nodes. 0 = disabled. |
| weight_stage2 | FLOAT | 0.00-2–2 | Stage 2 (refinement) weight. 0 = disabled. |
| weight_stage3 | FLOAT | 0.00-2–2 | Stage 3 (final polish) weight. 0 = disabled. |
| lora_path_overrideopt | STRING | Optional: Override with custom path (leave empty to use dropdown) | |
| keep_loadedopt | BOOLEAN | true | Keep the LoRA adapter loaded between runs. • True (default): load once, reuse on subsequent runs (faster; useful for final render settings). • False: unload ALL adapters before loading this LoRA on every run. Ensures a clean state each generation; useful when testing or swapping LoRAs frequently. |
Outputs (1)
| Name | Type | Description |
|---|---|---|
| pipeline | QWEN_IMAGE_PIPELINE | — |