ComfyUI Node
RunningHub SmartPhotoCrafter ModelLoader
Loads SmartPhotoCrafter models from fixed paths under ComfyUI/models and prepares a cached editor object for the ImageEdit node.
RunningHub SmartPhotoCrafter ModelLoader
- editor
◄use_dit_loratrue►
◄use_vlm_loratrue►
◄dtypebfloat16►
◄deviceauto►
◄vram_limit_gb0.0►
CategoryRunningHub/SmartPhotoCrafter
Inputs (5)
| Name | Type | Default | Description |
|---|---|---|---|
| use_dit_lora | BOOLEAN | true | Load the SmartPhotoCrafter DiT LoRA adapter from models/SmartPhotoCrafter/dit_lora when available. |
| use_vlm_lora | BOOLEAN | true | Load the SmartPhotoCrafter VLM LoRA adapter from models/SmartPhotoCrafter/vlm_lora when available. |
| dtype | COMBO | bfloat16 | Model computation dtype. bfloat16 is recommended on CUDA for this Qwen-based pipeline. |
| device | COMBO | auto | Device used for model execution. auto selects CUDA when available, otherwise CPU. |
| vram_limit_gb | FLOAT | 0.00–128 | DiffSynth VRAM management limit in GB. Set 0 to use total CUDA VRAM minus 0.5 GB automatically. |
Outputs (1)
| Name | Type | Description |
|---|---|---|
| editor | RH_SMARTPHOTOCRAFTER_EDITOR | — |