MoGe-3

MoGe-3 is Microsoft's monocular geometry model. From one RGB image it predicts metric depth in metres, a camera-space point map, the camera intrinsics, a validity mask and surface normals. It is MoGe-2 plus a sparse 3D UNet that refines the predicted depth on a voxel grid. LibreYOLO ships the official ViT-L and ViT-g checkpoints for inference, validation and normal export.

Tasks
depth, surface normals
Sizes
l, g at 840 px
Install
pip install libreyolo
Support tier
Inference only, since v1.8.0. Predict, validate and export only. Training features do not apply.
Upstream
MoGe-3 by Microsoft, MIT. Paper, source
Licenses
Code MIT, weights MIT. Commercial use

Install

MoGe-3 needs no optional extra. The sparse refiner is plain PyTorch and camera recovery uses SciPy, which is in the base install.

bash
pip install libreyolo

Predict

Weights download from the LibreYOLO Hugging Face mirrors on first use.

Python
from libreyolo import LibreYOLO, SAMPLE_IMAGE model = LibreYOLO("LibreMoGe3l-depth.pt")depth = model(SAMPLE_IMAGE).depth_map print(depth.data.shape)        # (H, W) metres, inf outside the valid maskprint(depth.point_map.shape)   # (H, W, 3) camera-frame points in metresprint(depth.intrinsics)        # (3, 3) OpenCV pixel intrinsicsprint(depth.valid_mask.shape)  # (H, W) bool
Normals
from libreyolo import LibreYOLO, SAMPLE_IMAGE model = LibreYOLO("LibreMoGe3l-normal.pt")normal = model(SAMPLE_IMAGE).normal_mapprint(normal.array.shape)      # (H, W, 3) unit vectors, OpenCV frame
CLI
libreyolo predict model=LibreMoGe3l-depth.pt source=https://raw.githubusercontent.com/LibreYOLO/libreyolo/release/libreyolo/assets/guggenheim-bilbao.jpg save=True

With a -depth checkpoint, result.depth_map holds metric depth in metres (encoding="depth", metric=True). The point map, intrinsics and validity mask are attached to it as depth_map.point_map, depth_map.intrinsics and depth_map.valid_mask; pixels outside the mask are inf. Points use OpenCV axes, and the intrinsics put pixel centres at integer coordinates. With a -normal checkpoint, result.normal_map holds unit normals in the OpenCV camera frame.

imgsz sets the ViT token budget, (imgsz // 14) ** 2 tokens laid out to the image aspect ratio. The default 840 is upstream's highest resolution level (3600 tokens); 490 is close to its lowest. Inputs keep their native resolution and every output is interpolated back to it. Inference runs in FP32 and always applies three refinement steps, as upstream does by default. One image runs per forward pass; there is no stacked-batch path.

CPU, CUDA and MPS all work without a compiled extension. The refiner's memory grows with imgsz, not with the input resolution.

Variants

Weight fileBackboneTask
LibreMoGe3l-depth.ptDINOv2 ViT-L/14depth, point map, intrinsics, mask
LibreMoGe3l-normal.ptDINOv2 ViT-L/14normals
LibreMoGe3g-depth.ptDINOv2 ViT-g/14depth, point map, intrinsics, mask
LibreMoGe3g-normal.ptDINOv2 ViT-g/14normals

The -depth and -normal files of one size hold the same tensors; the suffix picks the task. Against upstream on an RTX 3090 in FP32, the single-forward graph matches bit for bit and refined depth matches to floating-point rounding (p99 relative difference 4.7e-6 for l, 3.7e-5 for g).

MoGe-3 is not trainable in LibreYOLO: train() raises. Zero-shot val() runs the shared depth and normal validators.

Export

TaskONNXTorchScriptExecuTorchTensorRTOpenVINOPaddleMNNRKNNAxelerancnnTFLiteCoreMLCore AI
DepthDepth to ONNX: not supportedDepth to TorchScript: not supportedDepth to ExecuTorch: not supportedDepth to TensorRT: not supportedDepth to OpenVINO: not supportedDepth to Paddle: not supportedDepth to MNN: not supportedDepth to RKNN: not supportedDepth to Axelera: not supportedDepth to ncnn: not supportedDepth to TFLite: not supportedDepth to CoreML: not supportedDepth to Core AI: not supported
Surface normalsSurface normals to ONNX: supportedSurface normals to TorchScript: supportedSurface normals to ExecuTorch: not supportedSurface normals to TensorRT: not supportedSurface normals to OpenVINO: not supportedSurface normals to Paddle: not supportedSurface normals to MNN: not supportedSurface normals to RKNN: not supportedSurface normals to Axelera: not supportedSurface normals to ncnn: not supportedSurface normals to TFLite: not supportedSurface normals to CoreML: not supportedSurface normals to Core AI: not supported

Only the normal graph exports, to ONNX and TorchScript, on a fixed square canvas. Depth export is blocked: the refiner builds data-dependent voxel sets and camera recovery is a CPU solve.

Python
from libreyolo import LibreYOLO # Only the normal graph exports; depth export is blocked.model = LibreYOLO("LibreMoGe3l-normal.pt")model.export(format="onnx")model.export(format="torchscript")

Checkpoints

FileInput (px)Weights license
Depth
LibreMoGe3l-depth.ptmit
LibreMoGe3g-depth.ptmit
Surface normals
LibreMoGe3l-normal.pt840mit
LibreMoGe3g-normal.pt840mit

Every file above exists in the LibreYOLO org today and downloads on first use.

Licensing

Check the license on the Hugging Face repository of the specific weights you download. Every checkpoint in the LibreYOLO org carries one, and they are not always the same across a family. That repository is the authoritative source; the summary below describes what applied when this page was last verified.

This is a description of the licenses involved, not legal advice. If the answer matters commercially, read the licenses yourself and take your own counsel.

Original work
MoGe-3, Microsoft
Upstream license
MIT
LibreYOLO code
MIT
Weights
MIT, republished at huggingface.co/LibreYOLO
Interpretation
MIT is a permissive license, so the MoGe-3 code and weights can be used in commercial and closed-source products; the only obligation is to keep the copyright and license notice with copies. LibreYOLO ports the forward pass, sparse refinement and camera recovery from microsoft/MoGe at commit 74fbce054ebed49800de42d0ad0e83495065719a (MIT, Copyright Microsoft Corporation), re-expresses the FlexGEMM sparse kernels in plain PyTorch without vendoring them (FlexGEMM is MIT), and follows MIT utils3d helpers; the DINOv2 backbone keeps its separate Apache-2.0 attribution (libreyolo/models/moge3/NOTICE). The official checkpoints Ruicheng/moge-3-vitl (revision 184008f877d7ad1ad4c2cd2182a9bd1f63d0e5be) and Ruicheng/moge-3-vitg (revision 6ef26c5a4b4148dab5ccaacdb08b72dc66380475) declare MIT; LibreYOLO mirrors them unchanged, plus checkpoint metadata, as LibreYOLO/LibreMoGe3{l,g}-{depth,normal}, not gated.

MoGe-3 is by Microsoft (microsoft/MoGe). Code and weights are MIT, so commercial use is fine. The DINOv2 backbone keeps its Apache-2.0 attribution. LibreYOLO mirrors the official Ruicheng/moge-3-vitl and Ruicheng/moge-3-vitg checkpoints unchanged as LibreYOLO/LibreMoGe3{l,g}-{depth,normal}, not gated.

Citation

@misc{kong2026finedetailmonoculargeometryestimation,
      title={Fine-Detail Monocular Geometry Estimation with Self-Guided Sparse Volumetric Refinement},
      author={Lingyu Kong and Ruicheng Li and Ruicheng Wang and Sicheng Xu and Chengtang Yao and Jianfeng Xiang and Jiaolong Yang},
      year={2026},
      eprint={2607.17967},
      archivePrefix={arXiv},
      primaryClass={cs.CV},
      url={https://arxiv.org/abs/2607.17967},
}

Copied from the authors' citation block at github.com/microsoft/MoGe#-citation.

Verified against LibreYOLO v1.8.0. Support tables, checkpoints and benchmark numbers on this page are generated from the released library and the published weights, not written by hand.