{"id":13443503,"url":"https://github.com/voldemortX/pytorch-auto-drive","last_synced_at":"2025-03-20T16:31:36.848Z","repository":{"id":37386957,"uuid":"227979887","full_name":"voldemortX/pytorch-auto-drive","owner":"voldemortX","description":"PytorchAutoDrive: Segmentation models (ERFNet, ENet, DeepLab, FCN...) and Lane detection models (SCNN, RESA, LSTR, LaneATT, BézierLaneNet...) based on PyTorch with fast training, visualization, benchmarking \u0026 deployment help","archived":false,"fork":false,"pushed_at":"2023-10-04T03:24:58.000Z","size":5088,"stargazers_count":825,"open_issues_count":39,"forks_count":137,"subscribers_count":10,"default_branch":"master","last_synced_at":"2024-08-01T03:43:47.861Z","etag":null,"topics":["apex","cityscapes","culane","deeplab","erfnet","fcn","gtav","lane-detection","lstr","mobilenet","onnx","pascal-voc","pytorch","resa","scnn","semantic-segmentation","tensorboard","tensorrt","tusimple"],"latest_commit_sha":null,"homepage":"","language":"Python","has_issues":true,"has_wiki":null,"has_pages":null,"mirror_url":null,"source_name":null,"license":"bsd-3-clause","status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/voldemortX.png","metadata":{"files":{"readme":"README.md","changelog":null,"contributing":"docs/CONTRIBUTING.md","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}},"created_at":"2019-12-14T06:54:31.000Z","updated_at":"2024-07-28T02:38:55.000Z","dependencies_parsed_at":"2022-07-11T23:33:36.305Z","dependency_job_id":"ef7ed478-a642-4c7f-a540-9d4d5ecaf18d","html_url":"https://github.com/voldemortX/pytorch-auto-drive","commit_stats":null,"previous_names":[],"tags_count":12,"template":false,"template_full_name":null,"repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/voldemortX%2Fpytorch-auto-drive","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/voldemortX%2Fpytorch-auto-drive/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/voldemortX%2Fpytorch-auto-drive/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/voldemortX%2Fpytorch-auto-drive/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/voldemortX","download_url":"https://codeload.github.com/voldemortX/pytorch-auto-drive/tar.gz/refs/heads/master","host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":221780025,"owners_count":16879040,"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":["apex","cityscapes","culane","deeplab","erfnet","fcn","gtav","lane-detection","lstr","mobilenet","onnx","pascal-voc","pytorch","resa","scnn","semantic-segmentation","tensorboard","tensorrt","tusimple"],"created_at":"2024-07-31T03:02:02.321Z","updated_at":"2025-10-14T12:09:56.799Z","avatar_url":"https://github.com/voldemortX.png","language":"Python","funding_links":[],"categories":["Python"],"sub_categories":[],"readme":"# PytorchAutoDrive: Framework for self-driving perception\n\n\n*PytorchAutoDrive* is a **pure Python** framework includes semantic segmentation models, lane detection models based on **PyTorch**. Here we provide full stack supports from research (model training, testing, fair benchmarking by simply writing configs) to application (visualization, model deployment).\n\n**Paper:** [Rethinking Efficient Lane Detection via Curve Modeling](https://arxiv.org/abs/2203.02431) (CVPR 2022)\n\n**Poster:** [PytorchAutoDrive: Toolkit \u0026 Fair Benchmark for Autonomous Driving Research](https://drive.google.com/file/d/14EgcwPnKvAZJ1aWqBv6W9Msm666Wqi5a/view?usp=sharing) (PyTorch Developer Day 2021)\n\n*This repository is under active development, results with models uploaded are stable. For legacy code users, please check [deprecations](https://github.com/voldemortX/pytorch-auto-drive/issues/14) for changes.*\n\n**A demo video from ERFNet:**\n\nhttps://user-images.githubusercontent.com/32259501/148680744-a18793cd-f437-461f-8c3a-b909c9931709.mp4\n\n## Highlights\n\nVarious methods on a wide range of backbones, **config** based implementations, **modulated** and **easily understood** codes, image/keypoint loading, transformations and **visualizations**, **mixed precision training**, tensorboard logging and **deployment support** with ONNX and TensorRT.\n\nModels from this repo are faster to train (**single card trainable**) and often have better performance than other implementations, see [wiki](https://github.com/voldemortX/pytorch-auto-drive/wiki/Notes) for reasons and technical specification of models.\n\n## Supported datasets: \n\n| Task | Dataset |\n| :---: | :---: |\n| semantic segmentation | PASCAL VOC 2012 |\n| semantic segmentation | Cityscapes |\n| semantic segmentation | GTAV* |\n| semantic segmentation | SYNTHIA* |\n| lane detection | CULane |\n| lane detection | TuSimple |\n| lane detection | LLAMAS |\n| lane detection | BDD100K (*In progress*) |\n\n\\* The UDA baseline setup, with Cityscapes *val* set as validation.\n\n## Supported models:\n\n| Task | Backbone | Model/Method |\n| :---: | :---: | :---: |\n| semantic segmentation | ResNet-101 | [FCN](/configs/semantic_segmentation/fcn) |\n| semantic segmentation | ResNet-101 | [DeeplabV2](https://arxiv.org/abs/1606.00915) |\n| semantic segmentation | ResNet-101 | [DeeplabV3](https://arxiv.org/abs/1706.05587) |\n| semantic segmentation | - | [ENet](https://arxiv.org/abs/1606.02147) |\n| semantic segmentation | - | [ERFNet](/configs/semantic_segmentation/erfnet) |\n| lane detection | ENet, ERFNet, VGG16, ResNets (18, 34, 50, 101), MobileNets (V2, V3-Large), RepVGGs (A0, A1, B0, B1g2, B2), Swin (Tiny) | [Baseline](/configs/lane_detection/baseline) |\n| lane detection | ERFNet, VGG16, ResNets (18, 34, 50, 101), RepVGGs (A1) | [SCNN](https://arxiv.org/abs/1712.06080) |\n| lane detection | ResNets (18, 34, 50, 101), MobileNets (V2, V3-Large), ERFNet | [RESA](https://arxiv.org/abs/2008.13719) |\n| lane detection | ERFNet, ENet | [SAD](https://arxiv.org/abs/1908.00821) ([*Postponed*](https://github.com/voldemortX/pytorch-auto-drive/wiki/Notes)) |\n| lane detection | ERFNet | [PRNet](http://www.ecva.net/papers/eccv_2020/papers_ECCV/papers/123630698.pdf) (*In progress*) |\n| lane detection | ResNets (18, 34, 50, 101), ResNet18-reduced | [LSTR](https://arxiv.org/abs/2011.04233) |\n| lane detection | ResNets (18, 34) | [LaneATT](https://arxiv.org/abs/2010.12035) |\n| lane detection | ResNets (18, 34) | [BézierLaneNet](/configs/lane_detection/bezierlanenet) |\n\n## Model Zoo\n\nWe provide solid results (average/best/detailed), training time, shell scripts and trained models available for download in [MODEL_ZOO.md](docs/MODEL_ZOO_J.md).\n\n## Installation\n\nPlease prepare the environment and code with [INSTALL.md](docs/INSTALL.md). Then follow the instructions in [DATASET.md](docs/DATASET.md) to set up datasets. \n\n## Getting Started\n\nGet started with [LANEDETECTION.md](docs/LANEDETECTION.md) for lane detection.\n\nGet started with [SEGMENTATION.md](docs/SEGMENTATION.md) for semantic segmentation.\n\n## Visualization Tools\n\nRefer to [VISUALIZATION.md](docs/VISUALIZATION.md) for a visualization \u0026 inference tutorial, for image and video inputs.\n\n## Benchmark Tools\n\nRefer to [BENCHMARK.md](docs/BENCHMARK.md) for a benchmarking tutorial, including FPS test, FLOPs \u0026 memory count for each supported model.\n\n## Deployment\n\nRefer to [DEPLOY.md](docs/DEPLOY.md) for ONNX and TensorRT deployment supports.\n\n## Advanced Tutorial\n\nCheckout [ADVANCED_TUTORIAL.md](docs/ADVANCED_TUTORIAL.md) for advanced use cases and how to code in PytorchAutoDrive.\n\n## Contributing\n\nRefer to [CONTRIBUTING.md](/docs/CONTRIBUTING.md) for contribution guides.\n\n## Citation\n\nIf you feel this framework substantially helped your research or you want a reference when using our results, please cite the following paper that made the official release of PytorchAutoDrive:\n\n```\n@inproceedings{feng2022rethinking,\n  title={Rethinking efficient lane detection via curve modeling},\n  author={Feng, Zhengyang and Guo, Shaohua and Tan, Xin and Xu, Ke and Wang, Min and Ma, Lizhuang},\n  booktitle={Computer Vision and Pattern Recognition},\n  year={2022}\n}\n```\n\n## Credits:\n\nPytorchAutoDrive is maintained by Zhengyang Feng ([voldemortX](https://github.com/voldemortX)) and Shaohua Guo ([cedricgsh](https://github.com/cedricgsh)).\n\nContributors (GitHub ID): [kalkun](https://github.com/kalkun), [LittleJohnKhan](https://github.com/LittleJohnKhan), [francis0407](https://github.com/francis0407), [PannenetsF](https://github.com/PannenetsF), [bjzhb666](https://github.com/bjzhb666)\n\nPeople who sponsored us (e.g., with hardware): [Lizhuang Ma](https://dmcv.sjtu.edu.cn/people/), [Xin Tan](https://tanxincs.github.io/TAN-Xin.github.io/), Junshu Tang ([junshutang](https://github.com/junshutang)), Fengqi Liu ([FengqiLiu1221](https://github.com/FengqiLiu1221)) \n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2FvoldemortX%2Fpytorch-auto-drive","html_url":"https://awesome.ecosyste.ms/projects/github.com%2FvoldemortX%2Fpytorch-auto-drive","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2FvoldemortX%2Fpytorch-auto-drive/lists"}