ComfyUI Node
Lotus
A ComfyUI node in 🧪AILab/Geometry with 7 inputs and 1 output.
Lotus
- image
- IMAGE
◄modedepth►
◄resolution1024►
◄normalizationmin_max►
◄colormapgray►
◄weight_dtypefp16►
◄unload_modelfalse►
Category🧪AILab/Geometry
Inputs (7)
| Name | Type | Default | Description |
|---|---|---|---|
| image | IMAGE | Input video frames or single image. Connect from Load Image or Load Video. | |
| mode | COMBO | depth | Geometry estimation mode: - depth: SOTA Disparity Depth map - normal: SOTA Surface Normal vector map |
| resolution | INT | 1024256–2048 | Model internal processing resolution. Automatically resized back to input resolution. Set to 768 or 512 for 2x~3x faster video processing! |
| normalization | COMBO | min_max | Depth range normalization mode: - min_max: Standard 0 to 1 range (near=white 1.0, far=black 0.0) - raw: Unscaled numerical depth values |
| colormap | COMBO | gray | Output color format for depth: - gray: Grayscale (Required for ControlNet and Depth-to-Video) - spectral: Lotus official thermal colormap - turbo: Rainbow colormap - inferno: High-contrast thermal heatmap |
| weight_dtype | COMBO | fp16 | Computation precision: fp16 (recommended for speed & VRAM), bf16, or fp32. |
| unload_model | BOOLEAN | false | Unload model from VRAM immediately after execution to free maximum GPU memory. |
Outputs (1)
| Name | Type | Description |
|---|---|---|
| IMAGE | IMAGE | — |