{"id":13935831,"url":"https://github.com/FateScript/CenterNet-better","last_synced_at":"2025-07-19T21:30:44.822Z","repository":{"id":46525504,"uuid":"242475123","full_name":"FateScript/CenterNet-better","owner":"FateScript","description":"An easy to understand and better performance version of CenterNet","archived":false,"fork":false,"pushed_at":"2021-04-12T03:42:16.000Z","size":190,"stargazers_count":548,"open_issues_count":7,"forks_count":105,"subscribers_count":14,"default_branch":"master","last_synced_at":"2025-05-25T04:07:45.889Z","etag":null,"topics":["computer-vision","deep-learning","object-detection"],"latest_commit_sha":null,"homepage":null,"language":"Python","has_issues":true,"has_wiki":null,"has_pages":null,"mirror_url":null,"source_name":null,"license":"apache-2.0","status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/FateScript.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":"2020-02-23T07:30:28.000Z","updated_at":"2025-03-18T00:31:00.000Z","dependencies_parsed_at":"2022-08-26T03:32:15.785Z","dependency_job_id":null,"html_url":"https://github.com/FateScript/CenterNet-better","commit_stats":null,"previous_names":[],"tags_count":0,"template":false,"template_full_name":null,"purl":"pkg:github/FateScript/CenterNet-better","repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/FateScript%2FCenterNet-better","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/FateScript%2FCenterNet-better/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/FateScript%2FCenterNet-better/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/FateScript%2FCenterNet-better/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/FateScript","download_url":"https://codeload.github.com/FateScript/CenterNet-better/tar.gz/refs/heads/master","sbom_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/FateScript%2FCenterNet-better/sbom","host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":266019657,"owners_count":23864916,"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","deep-learning","object-detection"],"created_at":"2024-08-07T23:02:07.718Z","updated_at":"2025-07-19T21:30:44.233Z","avatar_url":"https://github.com/FateScript.png","language":"Python","funding_links":[],"categories":["Python"],"sub_categories":[],"readme":"# CenterNet\n\nThis repo is implemented based on my dl_lib, some parts of  code in my dl_lib  is based on [detectron2](https://github.com/facebookresearch/detectron2).\n\n## Motivation\n\n[Objects as Points](https://arxiv.org/abs/1904.07850) is one of my favorite paper in object detection area. However, its [code](https://github.com/xingyizhou/CenterNet/blob/master/README.md) is a little difficult to understand. I believe that CenterNet could get higher pts and implemented in a more elegant way, so I write this repo.\n\n## Performance\n\nThis repo use less training time to get a better performance, it nearly spend half training time and get 1~2 pts higher mAP compared with the old repo. Here is the table of performance.\n\n| Backbone     |  mAP    |  FPS    |  V100 FPS |  trained model    |  \n|:------------:|:-------:|:-------:|:---------:|:-----------------:|  \n|ResNet-18     | 29.8    | 92      | 113       | [google drive](https://drive.google.com/open?id=1D3tO95sdlsh9egOjOg0N-2HHmMfqbt5X)   |  \n|ResNet-50     | 34.9    | 57      | 71        | [google drive](https://drive.google.com/open?id=1t5Bw520_fJrn3aeSVxDBYNIgwpNdLR5s)   |  \n|ResNet-101    | 36.8    | 43      | 50        | [google drive](https://drive.google.com/open?id=1762Y93i9QreUTHq-87Ir73R2nNcrHuk0)   |  \n\n## What\\'s New?\n* **treat config as a object.** You could run your config file and check the config value, which is really helpful for debug.\n* **Common training / testing scripts in default.** you just need to invoke `dl_train/test --num-gpus x` in your playground and your projects only need to include all project-specific configs and network modules.\n* **Performance report is dumped automaticly.** After your training is over, we will evaluate your model automatically and generate a markdown file.\n* **Vectorize some operations.** This improves the speed and efficiency.\n\n## What\\'s comming\n  - [ ] Support DLA backbone\n  - [ ] Support Hourglass backbone\n  - [ ] Support KeyPoints dataset\n\n## Get started\n### Requirements\n* Python \u003e= 3.6\n* PyTorch \u003e= 1.3\n* torchvision that matches the PyTorch installation.\n* OpenCV\n* pycocotools\n```shell\npip install cython; pip install 'git+https://github.com/cocodataset/cocoapi.git#subdirectory=PythonAPI'\n```\n* GCC \u003e= 4.9\n```shell\ngcc --version\n```\n\n### Installation\n\nMake sure that your get at least one gpu when compiled. Run:\n```shell\npip install -e .\n```\n\n### Training\nFor example, if you want to train CenterNet with resnet-18 backbone, run:\n```shell\ncd playground/centernet.res18.coco.512size\ndl_train --num-gpus 8\n```\nAfter training process, a README.md file will be generated automatically and this file will report your model\\'s performance.  \n\nNOTE: For ResNet-18 and ResNet-50 backbone, we suppose your machine has over 150GB Memory for training. If your memory is not enough, please change NUM_WORKER (in config.py) to a smaller value.\n\n### Testing and Evaluation\n```shell\ndl_test --num-gpus 8 \n```\ntest downloaded model:\n```shell\ndl_test --num-gpus 8  MODEL.WEIGHTS path/to/your/save_dir/ckpt.pth \n```\n\n## Acknowledgement\n* [detectron2](https://github.com/facebookresearch/detectron2)\n* [CenterNet](https://github.com/xingyizhou/CenterNet)\n\n## Coding style\n\nplease refer to  [google python coding style](https://zh-google-styleguide.readthedocs.io/en/latest/google-python-styleguide/python_style_rules/)\n\n## Citing CenterNet-better\n\nIf you use CenterNet-better in your research or wish to refer to the baseline results published in this repo, please use the following BibTeX entry.\n\n```BibTeX\n@misc{wang2020centernet_better,\n  author =       {Feng Wang},\n  title =        {CenterNet-better},\n  howpublished = {\\url{https://github.com/FateScript/CenterNet-better}},\n  year =         {2020}\n}\n```\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2FFateScript%2FCenterNet-better","html_url":"https://awesome.ecosyste.ms/projects/github.com%2FFateScript%2FCenterNet-better","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2FFateScript%2FCenterNet-better/lists"}