{"id":13443682,"url":"https://github.com/lzccccc/SMOKE","last_synced_at":"2025-03-20T17:31:01.780Z","repository":{"id":37543045,"uuid":"233560751","full_name":"lzccccc/SMOKE","owner":"lzccccc","description":"SMOKE: Single-Stage Monocular 3D Object Detection via Keypoint Estimation","archived":false,"fork":false,"pushed_at":"2021-09-16T15:25:21.000Z","size":10133,"stargazers_count":704,"open_issues_count":60,"forks_count":177,"subscribers_count":21,"default_branch":"master","last_synced_at":"2024-10-28T07:39:40.572Z","etag":null,"topics":["3d-object-detection","autonomous-driving"],"latest_commit_sha":null,"homepage":"","language":"Python","has_issues":true,"has_wiki":null,"has_pages":null,"mirror_url":null,"source_name":null,"license":"mit","status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/lzccccc.png","metadata":{"files":{"readme":"README.md","changelog":null,"contributing":null,"funding":null,"license":"LICENSE","code_of_conduct":null,"threat_model":null,"audit":null,"citation":null,"codeowners":null,"security":null,"support":null}},"created_at":"2020-01-13T09:40:19.000Z","updated_at":"2024-10-24T05:36:33.000Z","dependencies_parsed_at":"2022-07-19T18:04:33.593Z","dependency_job_id":null,"html_url":"https://github.com/lzccccc/SMOKE","commit_stats":null,"previous_names":[],"tags_count":0,"template":false,"template_full_name":null,"repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/lzccccc%2FSMOKE","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/lzccccc%2FSMOKE/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/lzccccc%2FSMOKE/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/lzccccc%2FSMOKE/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/lzccccc","download_url":"https://codeload.github.com/lzccccc/SMOKE/tar.gz/refs/heads/master","host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":244660268,"owners_count":20489310,"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":["3d-object-detection","autonomous-driving"],"created_at":"2024-07-31T03:02:07.007Z","updated_at":"2025-03-20T17:30:58.531Z","avatar_url":"https://github.com/lzccccc.png","language":"Python","funding_links":[],"categories":["Python"],"sub_categories":[],"readme":"# SMOKE: Single-Stage Monocular 3D Object Detection via Keypoint Estimation\n\n\u003cimg align=\"center\" src=\"figures/animation.gif\" width=\"750\"\u003e\n\n[Video](https://www.youtube.com/watch?v=pvM_bASOQmo)\n\nThis repository is the official implementation of our paper [SMOKE: Single-Stage Monocular 3D Object Detection via Keypoint Estimation](https://arxiv.org/pdf/2002.10111.pdf).\nFor more details, please see our paper.\n\n## Introduction\nSMOKE is a **real-time** monocular 3D object detector for autonomous driving. \nThe runtime on a single NVIDIA TITAN XP GPU is **~30ms**. \nPart of the code comes from [CenterNet](https://github.com/xingyizhou/CenterNet), \n[maskrcnn-benchmark](https://github.com/facebookresearch/maskrcnn-benchmark),\nand [Detectron2](https://github.com/facebookresearch/detectron2).\n\nThe performance on KITTI 3D detection (3D/BEV) is as follows:\n\n|             |     Easy      |    Moderate    |     Hard     |\n|-------------|:-------------:|:--------------:|:------------:|\n| Car         | 14.17 / 21.08 | 9.88 / 15.13   | 8.63 / 12.91 | \n| Pedestrian  | 5.16  / 6.22  | 3.24 / 4.05    | 2.53 / 3.38  | \n| Cyclist     | 1.11  / 1.62  | 0.60 / 0.98    | 0.47 / 0.74  |\n\nThe pretrained weights can be downloaded [here](https://drive.google.com/open?id=11VK8_HfR7t0wm-6dCNP5KS3Vh-Qm686-).\n\n## Requirements\nAll codes are tested under the following environment:\n*   Ubuntu 16.04\n*   Python 3.7\n*   Pytorch 1.3.1\n*   CUDA 10.0\n\n## Dataset\nWe train and test our model on official [KITTI 3D Object Dataset](http://www.cvlibs.net/datasets/kitti/eval_object.php?obj_benchmark=3d). \nPlease first download the dataset and organize it as following structure:\n```\nkitti\n│──training\n│    ├──calib \n│    ├──label_2 \n│    ├──image_2\n│    └──ImageSets\n└──testing\n     ├──calib \n     ├──image_2\n     └──ImageSets\n```  \n\n## Setup\n1. We use `conda` to manage the environment:\n```\nconda create -n SMOKE python=3.7\n```\n\n2. Clone this repo:\n```\ngit clone https://github.com/lzccccc/SMOKE\n```\n\n3. Build codes:\n```\npython setup.py build develop\n```\n\n4. Link to dataset directory:\n```\nmkdir datasets\nln -s /path_to_kitti_dataset datasets/kitti\n```\n\n## Getting started\nFirst check the config file under `configs/`. \n\nWe train the model on 4 GPUs with 32 batch size:\n```\npython tools/plain_train_net.py --num-gpus 4 --config-file \"configs/smoke_gn_vector.yaml\"\n```\n\nFor single GPU training, simply run:\n```\npython tools/plain_train_net.py --config-file \"configs/smoke_gn_vector.yaml\"\n```\n\nWe currently only support single GPU testing:\n```\npython tools/plain_train_net.py --eval-only --config-file \"configs/smoke_gn_vector.yaml\"\n```\n\n## Acknowledgement\n[CenterNet](https://github.com/xingyizhou/CenterNet)\n\n[maskrcnn-benchmark](https://github.com/facebookresearch/maskrcnn-benchmark)\n\n[Detectron2](https://github.com/facebookresearch/detectron2)\n\n\n## Citations\nPlease cite our paper if you find SMOKE is helpful for your research.\n```\n@article{liu2020SMOKE,\n  title={{SMOKE}: Single-Stage Monocular 3D Object Detection via Keypoint Estimation},\n  author={Zechen Liu and Zizhang Wu and Roland T\\'oth},\n  journal={arXiv preprint arXiv:2002.10111},\n  year={2020}\n}\n```\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Flzccccc%2FSMOKE","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Flzccccc%2FSMOKE","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Flzccccc%2FSMOKE/lists"}