{"id":13442639,"url":"https://github.com/Sense-GVT/Fast-BEV","last_synced_at":"2025-03-20T14:31:52.055Z","repository":{"id":65400992,"uuid":"554232718","full_name":"Sense-GVT/Fast-BEV","owner":"Sense-GVT","description":"Fast-BEV: A Fast and Strong Bird’s-Eye View Perception Baseline","archived":false,"fork":false,"pushed_at":"2023-09-06T17:52:42.000Z","size":33769,"stargazers_count":620,"open_issues_count":59,"forks_count":92,"subscribers_count":14,"default_branch":"dev","last_synced_at":"2024-10-28T05:59:30.566Z","etag":null,"topics":["2d-to-3d","3d","autonomous","autonomous-driving","bird-eye-view","detection","multi-camera"],"latest_commit_sha":null,"homepage":"","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/Sense-GVT.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,"governance":null,"roadmap":null,"authors":null,"dei":null,"publiccode":null,"codemeta":null}},"created_at":"2022-10-19T13:21:52.000Z","updated_at":"2024-10-28T01:56:58.000Z","dependencies_parsed_at":"2024-10-28T04:00:25.015Z","dependency_job_id":"f065be39-7561-4603-807d-e78625493e9b","html_url":"https://github.com/Sense-GVT/Fast-BEV","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/Sense-GVT%2FFast-BEV","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/Sense-GVT%2FFast-BEV/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/Sense-GVT%2FFast-BEV/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/Sense-GVT%2FFast-BEV/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/Sense-GVT","download_url":"https://codeload.github.com/Sense-GVT/Fast-BEV/tar.gz/refs/heads/dev","host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":244630295,"owners_count":20484349,"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":["2d-to-3d","3d","autonomous","autonomous-driving","bird-eye-view","detection","multi-camera"],"created_at":"2024-07-31T03:01:48.433Z","updated_at":"2025-03-20T14:31:52.048Z","avatar_url":"https://github.com/Sense-GVT.png","language":"Python","funding_links":[],"categories":["Python"],"sub_categories":[],"readme":"# Fast-BEV\n\n[Fast-BEV: A Fast and Strong Bird’s-Eye View Perception Baseline](https://arxiv.org/abs/2301.12511)\n![image](https://github.com/Sense-GVT/Fast-BEV/blob/main/fast-bev++.png)\n![image](https://github.com/Sense-GVT/Fast-BEV/blob/main/benchmark_setting.png)\n![image](https://github.com/Sense-GVT/Fast-BEV/blob/main/benchmark.png)\n\n## Better Inference Implementation\nThanks to the repository [CUDA-FastBEV](https://github.com/Mandylove1993/CUDA-FastBEV) inference using CUDA \u0026 TensorRT. And provide PTQ and QAT int8 quantization code.\nYou can refer to it to get faster speed.\n\n## Usage\n\n[usage](https://github.com/Sense-GVT/Fast-BEV/blob/dev/tools/fastbev_run.sh)\n\n### Installation\n\n* CUDA\u003e=9.2\n* GCC\u003e=5.4\n* Python\u003e=3.6\n* Pytorch\u003e=1.8.1\n* Torchvision\u003e=0.9.1\n* MMCV-full==1.4.0\n* MMDetection==2.14.0\n* MMSegmentation==0.14.1\n\n### Dataset preparation\n\n```\n  .\n  ├── data\n  │   └── nuscenes\n  │       ├── maps\n  │       ├── maps_bev_seg_gt_2class\n  │       ├── nuscenes_infos_test_4d_interval3_max60.pkl\n  │       ├── nuscenes_infos_train_4d_interval3_max60.pkl\n  │       ├── nuscenes_infos_val_4d_interval3_max60.pkl\n  │       ├── v1.0-test\n  │       └── v1.0-trainval\n```\n\n[download](https://drive.google.com/drive/folders/10KyLm0xW3QiLhAefxBbXR-Hw_7nel_tm?usp=sharing)\n\n### Pretraining\n\n```\n  .\n  ├── pretrained_models\n  │   ├── cascade_mask_rcnn_r18_fpn_coco-mstrain_3x_20e_nuim_bbox_mAP_0.5110_segm_mAP_0.4070.pth\n  │   ├── cascade_mask_rcnn_r34_fpn_coco-mstrain_3x_20e_nuim_bbox_mAP_0.5190_segm_mAP_0.4140.pth\n  │   └── cascade_mask_rcnn_r50_fpn_coco-mstrain_3x_20e_nuim_bbox_mAP_0.5400_segm_mAP_0.4300.pth\n```\n\n[download](https://drive.google.com/drive/folders/19BD4totDHtwnHtOqTdn0xYJh7stwYd9l?usp=sharing)\n\n### Training\n\n```\n  .\n  ├── work_dirs\n    └── fastbev\n      └── exp\n          └── paper\n              └── fastbev_m0_r18_s256x704_v200x200x4_c192_d2_f4\n              │   ├── epoch_20.pth\n              │   ├── latest.pth -\u003e epoch_20.pth\n              │   ├── log.eval.fastbev_m0_r18_s256x704_v200x200x4_c192_d2_f4.02062323.txt\n              │   └── log.test.fastbev_m0_r18_s256x704_v200x200x4_c192_d2_f4.02062309.txt\n              ├── fastbev_m1_r18_s320x880_v200x200x4_c192_d2_f4\n              │   ├── epoch_20.pth\n              │   ├── latest.pth -\u003e epoch_20.pth\n              │   ├── log.eval.fastbev_m1_r18_s320x880_v200x200x4_c192_d2_f4.02080000.txt\n              │   └── log.test.fastbev_m1_r18_s320x880_v200x200x4_c192_d2_f4.02072346.txt\n              ├── fastbev_m2_r34_s256x704_v200x200x4_c224_d4_f4\n              │   ├── epoch_20.pth\n              │   ├── latest.pth -\u003e epoch_20.pth\n              │   ├── log.eval.fastbev_m2_r34_s256x704_v200x200x4_c224_d4_f4.02080021.txt\n              │   └── log.test.fastbev_m2_r34_s256x704_v200x200x4_c224_d4_f4.02080005.txt\n              ├── fastbev_m4_r50_s320x880_v250x250x6_c256_d6_f4\n              │   ├── epoch_20.pth\n              │   ├── latest.pth -\u003e epoch_20.pth\n              │   ├── log.eval.fastbev_m4_r50_s320x880_v250x250x6_c256_d6_f4.02080021.txt\n              │   └── log.test.fastbev_m4_r50_s320x880_v250x250x6_c256_d6_f4.02080005.txt\n              └── fastbev_m5_r50_s512x1408_v250x250x6_c256_d6_f4\n                  ├── epoch_20.pth\n                  ├── latest.pth -\u003e epoch_20.pth\n                  ├── log.eval.fastbev_m5_r50_s512x1408_v250x250x6_c256_d6_f4.02080021.txt\n                  └── log.test.fastbev_m5_r50_s512x1408_v250x250x6_c256_d6_f4.02080001.txt\n```\n\n[download](https://drive.google.com/drive/folders/1Ja9mqOE0iGPysVxmLSrZyUoCEBYu5fMH?usp=sharing)\n\n### Deployment\nTODO\n\n## View Transformation Latency on device\n[2D-to-3D on CUDA \u0026 CPU](https://github.com/Sense-GVT/Fast-BEV/tree/dev/script/view_tranform_cuda)\n\n## Citation\n```\n@article{li2023fast,\n  title={Fast-BEV: A Fast and Strong Bird's-Eye View Perception Baseline},\n  author={Li, Yangguang and Huang, Bin and Chen, Zeren and Cui, Yufeng and Liang, Feng and Shen, Mingzhu and Liu, Fenggang and Xie, Enze and Sheng, Lu and Ouyang, Wanli and others},\n  journal={arXiv preprint arXiv:2301.12511},\n  year={2023}\n}\n```\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2FSense-GVT%2FFast-BEV","html_url":"https://awesome.ecosyste.ms/projects/github.com%2FSense-GVT%2FFast-BEV","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2FSense-GVT%2FFast-BEV/lists"}