{"id":13644257,"url":"https://github.com/buxihuo/OW-YOLO","last_synced_at":"2025-04-21T07:30:42.856Z","repository":{"id":40495098,"uuid":"474900381","full_name":"buxihuo/OW-YOLO","owner":"buxihuo","description":"Detect known and unknown objects in the open world（具有区分已知与未知能力的全新检测器））","archived":false,"fork":false,"pushed_at":"2023-02-26T04:13:53.000Z","size":6418,"stargazers_count":91,"open_issues_count":14,"forks_count":7,"subscribers_count":5,"default_branch":"ow-yolo-det","last_synced_at":"2024-11-09T16:44:10.267Z","etag":null,"topics":["autolabel","deep-learning","object-detection","open-world","open-world-detection","pytorch","unknown","yolo","yolov5","yolov8"],"latest_commit_sha":null,"homepage":"","language":"Python","has_issues":true,"has_wiki":null,"has_pages":null,"mirror_url":null,"source_name":null,"license":null,"status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/buxihuo.png","metadata":{"files":{"readme":"README.md","changelog":null,"contributing":null,"funding":".github/FUNDING.yml","license":null,"code_of_conduct":null,"threat_model":null,"audit":null,"citation":null,"codeowners":null,"security":null,"support":null,"governance":null,"roadmap":null,"authors":null},"funding":{"github":null,"patreon":null,"open_collective":null,"ko_fi":null,"tidelift":null,"community_bridge":null,"liberapay":null,"issuehunt":null,"otechie":null,"lfx_crowdfunding":null,"custom":"https://github.com/buxihuo/OW-YOLO/blob/ow-yolo-det/.github/weixin.md"}},"created_at":"2022-03-28T07:53:25.000Z","updated_at":"2024-10-07T20:58:35.000Z","dependencies_parsed_at":"2024-01-14T12:18:23.920Z","dependency_job_id":"c361712d-750a-4bcb-8594-876f61ec5580","html_url":"https://github.com/buxihuo/OW-YOLO","commit_stats":null,"previous_names":[],"tags_count":4,"template":false,"template_full_name":null,"repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/buxihuo%2FOW-YOLO","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/buxihuo%2FOW-YOLO/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/buxihuo%2FOW-YOLO/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/buxihuo%2FOW-YOLO/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/buxihuo","download_url":"https://codeload.github.com/buxihuo/OW-YOLO/tar.gz/refs/heads/ow-yolo-det","host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":250014551,"owners_count":21360970,"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":["autolabel","deep-learning","object-detection","open-world","open-world-detection","pytorch","unknown","yolo","yolov5","yolov8"],"created_at":"2024-08-02T01:01:59.864Z","updated_at":"2025-04-21T07:30:38.536Z","avatar_url":"https://github.com/buxihuo.png","language":"Python","funding_links":["https://github.com/buxihuo/OW-YOLO/blob/ow-yolo-det/.github/weixin.md"],"categories":["Object Detection Applications","Frameworks"],"sub_categories":[],"readme":"![015](https://user-images.githubusercontent.com/84908793/221391769-2503d415-60c3-4c4b-a432-a4f0fdfa186c.png)\n\n# 快速开始\n\n## 1. 安装\n  同yolov5\n\n## 2. 推理\n\n\u003cdetails\u003e \n\u003csummary\u003e 推理示例\u003c/summary\u003e\n\n```bash\n$ python detect.py --source data/images --weights m-obj365.pt --unknownconf 0.45 --conf 0.25 \n$ python detect.py --source data/images --weights s-coco.pt --unknownconf 0.25 --conf 0.25\n'''\n参数解读\nunknownconf: 当网络预测的“不知道”分数大于此阈值时预测为不知道，否则输出已知分类。\n1）与已知类精度关系：当已知类精度越高时，“不知道”在已知类上发生的情况将越少，预测未知类时可以设定更低的unknownconf而不影响已知类性能。\n2）与训练集大小关系：训练集越丰富，预测未知类的能力越强\n注：可根据需求调节此参数，需要注意的是由于资源问题，在objects365数据集下训练的模型m-obj365.pt仅在小模型下训练了30轮，精度较低，建议采用较高阈值。\n其他参数：\n1）非极大值抑制：默认类内NMS（非极大值抑制）iou阈值为0.45，参数为iou；同时进行类间NMS，iou阈值为0.75，后续将提供参数接口。\n2）不知道的物体类名：可在detect文件中修改，后续将提供接口。             \n'''\n\n```\n\u003c/details\u003e\n\n\u003cdetails\u003e \n\u003csummary\u003e 视频展示\u003c/summary\u003e\n\n1. [demo1](https://b23.tv/MfpEmAm)\n2. [demo2](https://www.bilibili.com/video/BV1Nm4y1P7UW/?share_source=copy_web\u0026vd_source=4f63c00122ad06d30c832c5c6f903637)\n\n\u003c/details\u003e\n\n## 3. 预训练模型\n\n[s-coco.pt](https://github.com/buxihuo/OW-YOLO/releases/download/0.1/s-coco.pt)\u003cbr\u003e\n[m-obj365.pt](https://github.com/buxihuo/OW-YOLO/releases/download/0.1/m-obj365.pt)\u003cbr\u003e\n### coco数据集性能\n|Model                       |size\u003cbr\u003e\u003csup\u003e(pixels)|mAP\u003csup\u003eval\u003cbr\u003e0.5:0.95 |mAP\u003csup\u003eval\u003cbr\u003e0.5 |\n|---                         |---                  |---                     |--- \n|yolov8-n             |640                  |37.3                        |52.6\n|OW-yolov8-n          |640                  |37.9 (only known 38)                        |53.7 (only known 53.9)\n|OW-yolov8-n-la       |640                  |                        |\n```bash\nla : label attention\n\n```\n\n## 4. 后续功能\n图像分类、实例分割\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fbuxihuo%2FOW-YOLO","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Fbuxihuo%2FOW-YOLO","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fbuxihuo%2FOW-YOLO/lists"}