{"id":13737257,"url":"https://github.com/kunhe/FastAP-metric-learning","last_synced_at":"2025-05-08T13:33:32.804Z","repository":{"id":96035569,"uuid":"190536631","full_name":"kunhe/FastAP-metric-learning","owner":"kunhe","description":"Code for CVPR 2019 paper \"Deep Metric Learning to Rank\"","archived":false,"fork":false,"pushed_at":"2021-09-21T01:48:46.000Z","size":104,"stargazers_count":95,"open_issues_count":2,"forks_count":16,"subscribers_count":6,"default_branch":"master","last_synced_at":"2024-08-04T03:08:30.229Z","etag":null,"topics":["cvpr","deep-metric-learning","learning-to-rank","matconvnet","metric-learning","pytorch"],"latest_commit_sha":null,"homepage":"","language":"MATLAB","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/kunhe.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}},"created_at":"2019-06-06T07:36:00.000Z","updated_at":"2024-01-04T16:34:37.000Z","dependencies_parsed_at":null,"dependency_job_id":"6503eec9-6a5b-41ea-a7bd-c7ea4bbf27f1","html_url":"https://github.com/kunhe/FastAP-metric-learning","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/kunhe%2FFastAP-metric-learning","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/kunhe%2FFastAP-metric-learning/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/kunhe%2FFastAP-metric-learning/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/kunhe%2FFastAP-metric-learning/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/kunhe","download_url":"https://codeload.github.com/kunhe/FastAP-metric-learning/tar.gz/refs/heads/master","host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":224737430,"owners_count":17361345,"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":["cvpr","deep-metric-learning","learning-to-rank","matconvnet","metric-learning","pytorch"],"created_at":"2024-08-03T03:01:38.692Z","updated_at":"2025-05-08T13:33:32.792Z","avatar_url":"https://github.com/kunhe.png","language":"MATLAB","funding_links":[],"categories":["MATLAB"],"sub_categories":[],"readme":"# FastAP: Deep Metric Learning to Rank\nThis repository contains implementation of the following paper:\n\n[Deep Metric Learning to Rank](http://openaccess.thecvf.com/content_CVPR_2019/html/Cakir_Deep_Metric_Learning_to_Rank_CVPR_2019_paper.html)\u003cbr\u003e\n[Fatih Cakir](http://cs-people.bu.edu/fcakir/)\\*, [Kun He](http://cs-people.bu.edu/hekun/)\\*, [Xide Xia](https://xidexia.github.io), [Brian Kulis](http://people.bu.edu/bkulis/), and [Stan Sclaroff](http://www.cs.bu.edu/~sclaroff/) (*equal contribution)\u003cbr\u003e\nIEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2019\n\n![](fastap_github.jpg)\n\n## Other Implementations\n[FastAPLoss](https://kevinmusgrave.github.io/pytorch-metric-learning/losses/#fastaploss) from [pytorch-metric-learning](https://github.com/KevinMusgrave/pytorch-metric-learning)\n\n## Usage \n* **Matlab**: see `matlab/README.md`\n* **PyTorch**: see `pytorch/README.md`\n\n## Datasets\n* Stanford Online Products\n  * Can be downloaded [here](http://cvgl.stanford.edu/projects/lifted_struct/)\n* In-Shop Clothes Retrieval\n  * Can be downloaded [here](http://mmlab.ie.cuhk.edu.hk/projects/DeepFashion.html)\n* PKU VehicleID\n  * Please request the dataset from the authors [here](https://pkuml.org/resources/pku-vehicleid.html)\n\n## Reproducibility\n* We provide trained [MatConvNet](https://www.vlfeat.org/matconvnet/quick/) models and experimental logs for the results in the paper. These models were used to achieve the results in the tables.\n* The logs also include parameters settings that enable one to re-train a model if desired. It also includes evaluation results with model checkpoints at certain epochs.\n    * Table 1: Stanford Online Products\n        * FastAP, ResNet-18, M=256, Dim=512: [[model @ epoch 20](https://drive.google.com/file/d/1sPCG34rV4Bqf0aWF7GrFIDUK7DGjcaB5/view?usp=sharing), [log](https://drive.google.com/open?id=14m3fHgeZu8MIAePFHXe141R60KRwH1d8)]\n        * FastAP, ResNet-50, M=96, Dim=128: [[model @ epoch 30](https://drive.google.com/open?id=1yGUVTskdERdLeF85GP-lLRnwS0KhpvvL), [log](https://drive.google.com/open?id=1A0G1aUBS7URotInbT7eBbys4xfARvCCe)]\n        * FastAP, ResNet-50, M=96, Dim=512: [[model @ epoch 28](https://drive.google.com/file/d/14yEyAYhGzNygBBn8r2RcZut_Ye14mJoK/view?usp=sharing), [log](https://drive.google.com/open?id=19mpLn1OqA2nqpMZtvZ3GvFk_VppOOPTc)]\n        * FastAP, ResNet-50, M=256, Dim=512: [[model @ epoch 12](https://drive.google.com/open?id=1WfV1ArXHG4oksHGE8DZRDwsxupoO60sD), [log](https://drive.google.com/open?id=1shvC5qB8O0l6vH1qa2SG1oxX8_jM1gdi)]\n    * Table 2: In-Shop Clothes\n        * FastAP, ResNet-18, M=256, Dim=512: [[model @ epoch 50](https://drive.google.com/open?id=1ZZ-Fpx9uPkRL-QXL-8-RcROjQVOLcbr5), [log](https://drive.google.com/file/d/1osxoHsMy11v-kvUNTuRG3luhsxMMn78B/view?usp=sharing)]\n        * FastAP, ResNet-50, M=96, Dim=512: [[model @ epoch 40](https://drive.google.com/open?id=1PyiHog7fJp_InvqdAO0dzyJDRMNvAXxm), [log](https://drive.google.com/open?id=14IPgDfkbKo9PnrgMFFRDSIBRW1xwRYs5)]\n        * FastAP, ResNet-50, M=256, Dim=512: [[model @ epoch 35](https://drive.google.com/open?id=1T5IynM63YqnGslnMGppJsmtJdHIWamJv), [log](https://drive.google.com/open?id=1oud9i87FTJE7Ei636bjxqgBXpahysRKK)]\n    * Table 3: PKU VehicleID\n        * FastAP, ResNet-18, M=256, Dim=512: [[model @ epoch 50](https://drive.google.com/open?id=1KsUF2SzkhvBOkHzbrXKj7H5KtN6Z3hRJ), [log](https://drive.google.com/open?id=155Ce-FmI6dmMgJnXdESVHx08unU3jWX2)]\n        * FastAP, ResNet-50, M=96, Dim=512: [[model @ epoch 40](https://drive.google.com/open?id=1AblJelRHStBfWwmZeoRM8iEpNRNdOobn), [log](https://drive.google.com/open?id=1twswLE-j9kLxUsk5Ku7vWqBp0Sml65sG)]\n        * FastAP, ResNet-50, M=256, dim=512: [[model @ epoch 30](https://drive.google.com/open?id=1MAimhKEyEfq2LDYnUaDburFH2YsUhrpA), [log](https://drive.google.com/open?id=1CtNk-wxSZToO703OvfK8ndFQzFnpOvVS)]\n        \n        (M=mini-batch size)\n * PyTorch code is a direct port from our MATLAB implementation. We haven't tried reproducing the paper results with our PyTorch code. **For reproducibility use the MATLAB version**. \n * Note that the mini-batch sampling strategy must also be used alongside the FastAP loss for good results.\n\n## Contact\nFor questions and comments, feel free to contact: kunhe@fb.com or fcakirs@gmail.com\n\n## License\nMIT\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fkunhe%2FFastAP-metric-learning","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Fkunhe%2FFastAP-metric-learning","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fkunhe%2FFastAP-metric-learning/lists"}