{"id":26695150,"url":"https://github.com/rmurai0610/MASt3R-SLAM","last_synced_at":"2025-03-26T19:01:42.636Z","repository":{"id":279467439,"uuid":"938231495","full_name":"rmurai0610/MASt3R-SLAM","owner":"rmurai0610","description":"[CVPR 2025] MASt3R-SLAM: Real-Time Dense SLAM with 3D Reconstruction Priors","archived":false,"fork":false,"pushed_at":"2025-03-11T12:23:34.000Z","size":48505,"stargazers_count":1583,"open_issues_count":13,"forks_count":131,"subscribers_count":20,"default_branch":"main","last_synced_at":"2025-03-24T12:02:07.935Z","etag":null,"topics":["computer-vision","cvpr2025","robotics","slam"],"latest_commit_sha":null,"homepage":"https://edexheim.github.io/mast3r-slam/","language":"Python","has_issues":true,"has_wiki":null,"has_pages":null,"mirror_url":null,"source_name":null,"license":"other","status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/rmurai0610.png","metadata":{"files":{"readme":"README.md","changelog":null,"contributing":null,"funding":null,"license":"LICENSE.md","code_of_conduct":null,"threat_model":null,"audit":null,"citation":null,"codeowners":null,"security":null,"support":null,"governance":null,"roadmap":null,"authors":null,"dei":null,"publiccode":null,"codemeta":null}},"created_at":"2025-02-24T16:18:37.000Z","updated_at":"2025-03-24T11:57:31.000Z","dependencies_parsed_at":"2025-03-24T12:00:32.964Z","dependency_job_id":"5d9e26ee-37f8-40d8-ba75-7105d47e95c4","html_url":"https://github.com/rmurai0610/MASt3R-SLAM","commit_stats":null,"previous_names":["rmurai0610/mast3r-slam"],"tags_count":0,"template":false,"template_full_name":null,"repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/rmurai0610%2FMASt3R-SLAM","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/rmurai0610%2FMASt3R-SLAM/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/rmurai0610%2FMASt3R-SLAM/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/rmurai0610%2FMASt3R-SLAM/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/rmurai0610","download_url":"https://codeload.github.com/rmurai0610/MASt3R-SLAM/tar.gz/refs/heads/main","host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":245718740,"owners_count":20661160,"icon_url":"https://github.com/github.png","version":null,"created_at":"2022-05-30T11:31:42.601Z","updated_at":"2022-07-04T15:15:14.044Z","host_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub","repositories_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories","repository_names_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repository_names","owners_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners"}},"keywords":["computer-vision","cvpr2025","robotics","slam"],"created_at":"2025-03-26T19:01:41.706Z","updated_at":"2025-03-26T19:01:42.620Z","avatar_url":"https://github.com/rmurai0610.png","language":"Python","funding_links":[],"categories":["Python","novel","Visual Odometry \u0026 SLAM"],"sub_categories":["Visual \u0026 Visual-Inertial Odometry"],"readme":"[comment]: \u003c\u003e (# MASt3R-SLAM: Real-Time Dense SLAM with 3D Reconstruction Priors)\n\n\u003cp align=\"center\"\u003e\n  \u003ch1 align=\"center\"\u003eMASt3R-SLAM: Real-Time Dense SLAM with 3D Reconstruction Priors\u003c/h1\u003e\n  \u003cp align=\"center\"\u003e\n    \u003ca href=\"https://rmurai.co.uk/\"\u003e\u003cstrong\u003eRiku Murai*\u003c/strong\u003e\u003c/a\u003e\n    ·\n    \u003ca href=\"https://edexheim.github.io/\"\u003e\u003cstrong\u003eEric Dexheimer*\u003c/strong\u003e\u003c/a\u003e\n    ·\n    \u003ca href=\"https://www.doc.ic.ac.uk/~ajd/\"\u003e\u003cstrong\u003eAndrew J. Davison\u003c/strong\u003e\u003c/a\u003e\n  \u003c/p\u003e\n  \u003cp align=\"center\"\u003e(* Equal Contribution)\u003c/p\u003e\n\n[comment]: \u003c\u003e (  \u003ch2 align=\"center\"\u003ePAPER\u003c/h2\u003e)\n  \u003ch3 align=\"center\"\u003e\u003ca href=\"https://arxiv.org/abs/2412.12392\"\u003ePaper\u003c/a\u003e | \u003ca href=\"https://youtu.be/wozt71NBFTQ\"\u003eVideo\u003c/a\u003e | \u003ca href=\"https://edexheim.github.io/mast3r-slam/\"\u003eProject Page\u003c/a\u003e\u003c/h3\u003e\n  \u003cdiv align=\"center\"\u003e\u003c/div\u003e\n\n\u003cp align=\"center\"\u003e\n    \u003cimg src=\"./media/teaser.gif\" alt=\"teaser\" width=\"100%\"\u003e\n\u003c/p\u003e\n\u003cbr\u003e\n\n# Getting Started\n## Installation\n```\nconda create -n mast3r-slam python=3.11\nconda activate mast3r-slam\n```\nCheck the system's CUDA version with nvcc\n```\nnvcc --version\n```\nInstall pytorch with **matching** CUDA version following:\n```\n# CUDA 11.8\nconda install pytorch==2.5.1 torchvision==0.20.1 torchaudio==2.5.1  pytorch-cuda=11.8 -c pytorch -c nvidia\n# CUDA 12.1\nconda install pytorch==2.5.1 torchvision==0.20.1 torchaudio==2.5.1 pytorch-cuda=12.1 -c pytorch -c nvidia\n# CUDA 12.4\nconda install pytorch==2.5.1 torchvision==0.20.1 torchaudio==2.5.1 pytorch-cuda=12.4 -c pytorch -c nvidia\n```\n\nClone the repo and install the dependencies.\n```\ngit clone https://github.com/rmurai0610/MASt3R-SLAM.git --recursive\ncd MASt3R-SLAM/\n\n# if you've clone the repo without --recursive run\n# git submodule update --init --recursive\n\npip install -e thirdparty/mast3r\npip install -e thirdparty/in3d\npip install --no-build-isolation -e .\n \n\n# Optionally install torchcodec for faster mp4 loading\npip install torchcodec==0.1\n```\n\nSetup the checkpoints for MASt3R and retrieval.  The license for the checkpoints and more information on the datasets used is written [here](https://github.com/naver/mast3r/blob/mast3r_sfm/CHECKPOINTS_NOTICE).\n```\nmkdir -p checkpoints/\nwget https://download.europe.naverlabs.com/ComputerVision/MASt3R/MASt3R_ViTLarge_BaseDecoder_512_catmlpdpt_metric.pth -P checkpoints/\nwget https://download.europe.naverlabs.com/ComputerVision/MASt3R/MASt3R_ViTLarge_BaseDecoder_512_catmlpdpt_metric_retrieval_trainingfree.pth -P checkpoints/\nwget https://download.europe.naverlabs.com/ComputerVision/MASt3R/MASt3R_ViTLarge_BaseDecoder_512_catmlpdpt_metric_retrieval_codebook.pkl -P checkpoints/\n```\n\n## WSL Users\nWe have primarily tested on Ubuntu.  If you are using WSL, please checkout to the windows branch and follow the above installation.\n```\ngit checkout windows\n```\nThis disables multiprocessing which causes an issue with shared memory as discussed [here](https://github.com/rmurai0610/MASt3R-SLAM/issues/21).\n\n## Examples\n```\nbash ./scripts/download_tum.sh\npython main.py --dataset datasets/tum/rgbd_dataset_freiburg1_room/ --config config/calib.yaml\n```\n## Live Demo\nConnect a realsense camera to the PC and run\n```\npython main.py --dataset realsense --config config/base.yaml\n```\n## Running on a video\nOur system can process either MP4 videos or folders containing RGB images.\n```\npython main.py --dataset \u003cpath/to/video\u003e.mp4 --config config/base.yaml\npython main.py --dataset \u003cpath/to/folder\u003e --config config/base.yaml\n```\nIf the calibration parameters are known, you can specify them in intrinsics.yaml\n```\npython main.py --dataset \u003cpath/to/video\u003e.mp4 --config config/base.yaml --calib config/intrinsics.yaml\npython main.py --dataset \u003cpath/to/folder\u003e --config config/base.yaml --calib config/intrinsics.yaml\n```\n\n## Downloading Dataset\n### TUM-RGBD Dataset\n```\nbash ./scripts/download_tum.sh\n```\n\n### 7-Scenes Dataset\n```\nbash ./scripts/download_7_scenes.sh\n```\n\n### EuRoC Dataset\n```\nbash ./scripts/download_euroc.sh\n```\n### ETH3D SLAM Dataset\n```\nbash ./scripts/download_eth3d.sh\n```\n\n## Running Evaluations\nAll evaluation script will run our system in a single-threaded, headless mode.\nWe can run evaluations with/without calibration:\n### TUM-RGBD Dataset\n```\nbash ./scripts/eval_tum.sh \nbash ./scripts/eval_tum.sh --no-calib\n```\n\n### 7-Scenes Dataset\n```\nbash ./scripts/eval_7_scenes.sh \nbash ./scripts/eval_7_scenes.sh --no-calib\n```\n\n### EuRoC Dataset\n```\nbash ./scripts/eval_euroc.sh \nbash ./scripts/eval_euroc.sh --no-calib\n```\n### ETH3D SLAM Dataset\n```\nbash ./scripts/eval_eth3d.sh \n```\n\n## Reproducibility\nThere might be minor differences between the released version and the results in the paper after developing this multi-processing version. \nWe run all our experiments on an RTX 4090, and the performance may differ when running with a different GPU.\n\n## Acknowledgement\nWe sincerely thank the developers and contributors of the many open-source projects that our code is built upon.\n- [MASt3R](https://github.com/naver/mast3r)\n- [MASt3R-SfM](https://github.com/naver/mast3r/tree/mast3r_sfm)\n- [DROID-SLAM](https://github.com/princeton-vl/DROID-SLAM)\n- [ModernGL](https://github.com/moderngl/moderngl)\n\n# Citation\nIf you found this code/work to be useful in your own research, please considering citing the following:\n\n```bibtex\n@article{murai2024_mast3rslam,\n    title={{MASt3R-SLAM}: Real-Time Dense {SLAM} with {3D} Reconstruction Priors},\n    author={Murai, Riku and Dexheimer, Eric and Davison, Andrew J.},\n    journal={arXiv preprint},\n    year={2024},\n}      \n```\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Frmurai0610%2FMASt3R-SLAM","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Frmurai0610%2FMASt3R-SLAM","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Frmurai0610%2FMASt3R-SLAM/lists"}