{"id":13472380,"url":"https://github.com/researchmm/LightTrack","last_synced_at":"2025-03-26T15:32:19.984Z","repository":{"id":37633000,"uuid":"352858185","full_name":"researchmm/LightTrack","owner":"researchmm","description":"[CVPR21] LightTrack: Finding Lightweight Neural Network for Object Tracking via One-Shot Architecture Search","archived":false,"fork":false,"pushed_at":"2021-12-29T08:19:42.000Z","size":10227,"stargazers_count":393,"open_issues_count":23,"forks_count":59,"subscribers_count":16,"default_branch":"main","last_synced_at":"2024-08-01T16:29:55.779Z","etag":null,"topics":["cvpr2021","efficient","nas"],"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/researchmm.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":"2021-03-30T03:26:30.000Z","updated_at":"2024-07-28T06:16:38.000Z","dependencies_parsed_at":"2022-07-14T21:46:51.509Z","dependency_job_id":null,"html_url":"https://github.com/researchmm/LightTrack","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/researchmm%2FLightTrack","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/researchmm%2FLightTrack/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/researchmm%2FLightTrack/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/researchmm%2FLightTrack/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/researchmm","download_url":"https://codeload.github.com/researchmm/LightTrack/tar.gz/refs/heads/main","host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":222151630,"owners_count":16939423,"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":["cvpr2021","efficient","nas"],"created_at":"2024-07-31T16:00:54.219Z","updated_at":"2024-10-30T02:32:06.279Z","avatar_url":"https://github.com/researchmm.png","language":"Python","funding_links":[],"categories":["Python"],"sub_categories":[],"readme":"# LightTrack: Finding Lightweight Neural Networks for Object Tracking via One-Shot Architecture Search\n\n\u003cdiv align=\"center\"\u003e\n  \u003cimg src=\"Archs.gif\" width=\"800px\" /\u003e\n\u003c/div\u003e\n\nThe official implementation of the paper \n\n[**LightTrack: Finding Lightweight Neural Networks for Object Tracking via One-Shot Architecture Search**](https://arxiv.org/abs/2104.14545)\n\nHiring research interns for visual transformer projects: houwen.peng@microsoft.com\n## News\n- We have uploaded the **pre-trained weights of the SuperNets**(for both ImageNet classification and object tracking) to [Google Drive](https://drive.google.com/drive/folders/1HXhdJO3yhQYw3O7nGUOXHu2S20Bs8CfI). Users can use them as initialization for future research on **efficient object tracking**.\n## Abstract\n\nWe present LightTrack, which uses neural architecture search (NAS) to design more lightweight and efficient object trackers. Comprehensive experiments show that our LightTrack is effective. It can find trackers that achieve superior performance compared to handcrafted SOTA trackers, such as SiamRPN++ and Ocean, while using much fewer model Flops and parameters. Moreover, when deployed on resource-constrained mobile chipsets, the discovered trackers run much faster. For example, on Snapdragon 845 Adreno GPU, LightTrack runs 12× faster than Ocean, while using 13× fewer parameters and 38× fewer Flops. Such improvements might narrow the gap between academic models and industrial deployments in object tracking task.\n\n\u003cdiv align=\"center\"\u003e\n  \u003cimg src=\"LightTrack_Fig1.PNG\" width=\"500px\" /\u003e\n\u003c/div\u003e\n\n## Environment Installation\n```\ncd lighttrack\nconda create -n lighttrack python=3.6\nconda activate lighttrack\nbash install.sh\n```\n## Data Preparation\n- Tracking Benchmarks\n\nPlease put VOT2019 dataset under `$LightTrack/dataset`. The prepared data should look like:\n```\n$LighTrack/dataset/VOT2019.json\n$LighTrack/dataset/VOT2019/agility\n$LighTrack/dataset/VOT2019/ants1\n...\n$LighTrack/dataset/VOT2019/list.txt\n```\n## Test and evaluation\nTest LightTrack-Mobile on VOT2019\n```\nbash tracking/reproduce_vot2019.sh\n```\n## Flops, Params, and Speed\nCompute the flops and params of our LightTrack-Mobile. The flops counter we use is [pytorch-OpCounter](https://github.com/Lyken17/pytorch-OpCounter)\n```\npython tracking/FLOPs_Params.py\n```\nTest the running speed of our LightTrack-Mobile\n```\npython tracking/Speed.py\n```\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fresearchmm%2FLightTrack","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Fresearchmm%2FLightTrack","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fresearchmm%2FLightTrack/lists"}