ComfyUI nodes to use Depth Anything V3 - the latest depth estimation model from ByteDance. For now monocular depth, camera pose estimation and 3D point clouds/gaussians are supported. Models autodownload from HuggingFace (depth-anything org). This is a first draft, let me know if you have any feedback! :)

ComfyUI-DepthAnythingV3


Enable tiled processing for memory-efficient inference on high-resolution images.

This node configures the model to process images in tiles with overlapping regions,
then blends the results for seamless output.

Parameters:
- tile_size: Size of each tile (should be multiple of 14 for patch alignment)
- overlap: Overlap between adjacent tiles for smooth blending

Use this when:
- Processing 4K+ resolution images
- GPU memory is limited
- Getting out-of-memory errors

Note: Tiled processing may produce slightly different results at tile boundaries,
but the overlap and blending minimize artifacts.


ComfyUI Node: DA3 Enable Tiled Processing

Category

Inputs

Outputs

Extension: ComfyUI-DepthAnythingV3

Run ComfyUI workflows in the Cloud!