Microsoft's MoGe-3 3D geometry model lands in ComfyUI with point, depth and normal outputs
petewoodbridge · x · 2026-09-19
MoGe-3, a monocular geometry estimation model, is now repackaged by Comfy Org on Hugging Face for direct use in ComfyUI, already logging 92k+ monthly downloads.
- Recovers metric point/depth/normal maps plus FOV in a single pass from open-domain images, targeting fine-detail 3D reconstruction
- The repo bundles MoGe-1/2/3 checkpoints (vitl and vitg) under an MIT license
- Install: drop the safetensors files into ComfyUI/models/geometryestimation/
- Original code lives at microsoft/moge with weights from Ruicheng on Hugging Face
More from Multimodal
- Free course covers the 4 hard problems of production generative media agents — Al_Grigor · 2026-11-04
- PixelBench: testing multimodal models pixel-by-pixel instead of with QA — bdsqlsz · 2026-11-04
- azed_ai shows 2D character art turned into matching 3D model — azed_ai · 2026-11-04
- Kitten-TTS 2 offers free voice cloning on laptops, no GPU needed — solyarisoftware · 2026-11-04
- "We'll Take Your RAM": a fully local-hardware cyberpunk meme from 2030 San Francisco — Wonderful-Yak-3348 · 2026-11-04
- Openfilm: open-source AI video agent turns coding agents into video makers — solyarisoftware · 2026-11-04