{"id":16505211,"url":"https://github.com/qidiso/mobilefacenet-v2","last_synced_at":"2025-04-06T11:08:55.723Z","repository":{"id":153101194,"uuid":"136049733","full_name":"qidiso/mobilefacenet-V2","owner":"qidiso","description":"🔥improve the accuracy of mobilefacenet(insight face) reached 99.733 in the cfp-ff、 the 99.68+ in lfw,96.71+ in agedb30.🔥","archived":false,"fork":false,"pushed_at":"2019-08-08T09:40:08.000Z","size":481,"stargazers_count":397,"open_issues_count":19,"forks_count":148,"subscribers_count":71,"default_branch":"master","last_synced_at":"2025-03-30T10:07:38.090Z","etag":null,"topics":["face","fast","insightface","mobile","mobilenet"],"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/qidiso.png","metadata":{"files":{"readme":"README.md","changelog":null,"contributing":null,"funding":null,"license":null,"code_of_conduct":null,"threat_model":null,"audit":null,"citation":null,"codeowners":null,"security":null,"support":null,"governance":null}},"created_at":"2018-06-04T16:00:42.000Z","updated_at":"2025-03-17T09:25:12.000Z","dependencies_parsed_at":"2023-06-03T17:45:30.229Z","dependency_job_id":null,"html_url":"https://github.com/qidiso/mobilefacenet-V2","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/qidiso%2Fmobilefacenet-V2","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/qidiso%2Fmobilefacenet-V2/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/qidiso%2Fmobilefacenet-V2/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/qidiso%2Fmobilefacenet-V2/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/qidiso","download_url":"https://codeload.github.com/qidiso/mobilefacenet-V2/tar.gz/refs/heads/master","host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":247471520,"owners_count":20944158,"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":["face","fast","insightface","mobile","mobilenet"],"created_at":"2024-10-11T15:09:53.180Z","updated_at":"2025-04-06T11:08:55.704Z","avatar_url":"https://github.com/qidiso.png","language":"Python","funding_links":[],"categories":[],"sub_categories":[],"readme":"# mobilefacenet-V2\n\nnow we get more higher accuray:\n\n\n[lfw][12000]Accuracy-Flip: 0.99667+-0.00358  \n[agedb_30][12000]Accuracy-Flip: 0.96667+-0.00167 use my modified mobilenet network.\n\nlr-batch-epoch: 0.01 11738 1\ntesting verification..\n(12000, 512)\ninfer time 39.129495\n[lfw][36000]XNorm: 22.729305\n[lfw][36000]Accuracy-Flip: 0.99667+-0.00358\n\n\n\nimprove the accuracy of mobilefacenet \nin paper mobilefacenet论文(https://arxiv.org/abs/1804.07573) \n\nFirst step training (use softmax to pretrain): \ntrain softmax(facenet):\n\n[lfw][62000]XNorm: 23.029881\n[lfw][62000]Accuracy-Flip: 0.99383+-0.00308\ntesting verification..\n(14000, 512)\ninfer time 20.121058\n[cfp_fp][62000]XNorm: 24.043967\n[cfp_fp][62000]Accuracy-Flip: 0.89343+-0.01705\ntesting verification..\n(12000, 512)\ninfer time 16.860138\n[agedb_30][62000]XNorm: 23.566453\n[agedb_30][62000]Accuracy-Flip: 0.93883+-0.01675\nsaving 31\nINFO:root:Saved checkpoint to \"../models/MF/model-y1-softmax12-0031.params\"\n\n\npretrained models:\nhttps://pan.baidu.com/s/1xBq9FoL79z7K892aFWkmFw\n\n\nSecond step:\nCUDA_VISIBLE_DEVICES='0,1' python -u train_softmax.py --network y1 --ckpt 2 --loss-type 4 --margin-s [128] --lr-steps 120000,180000,210000,230000 --emb-size [512] --per-batch-size 150 --data-dir ../data/faces_ms1m_112x112 --pretrained ../models/MobileFaceNet/model-y1-softmax,20 --prefix ../models/MF/model-y1-arcface\n\nThird step:\nCUDA_VISIBLE_DEVICES='0,1' python -u train_softmax.py --network y1 --ckpt 2 --loss-type 4  --lr 0.001 --lr-steps 40000,60000,70000 --wd 0.00004 --fc7-wd-mult 10 --emb-size 512 --per-batch-size 150 --margin-s 64 --data-dir ../data/faces_ms1m_112x112 --pretrained ../models/MF/model-y1-arcface,46 --prefix ../models/MF/model-y1-arcface\n\nUpdate wd=0.00001 , --fc7-wd-mult 10 --emb-size 512\ni get new Accuracy:\n###### Accuracy\n| dbname | accuracy |\n| ----- |:-----:|\n| lfw |0.996233|\n| cfp_fp |0.94300|\n| age_db30 |0.96383|\n\n\n##########first\n#CUDA_VISIBLE_DEVICES='0' python -u train_softmax.py --network y1 --ckpt 2 --loss-type 4   --lr 0.1 --emb-size 512 --per-batch-size 240 --margin-s 64 --wd 0.00001 --fc7-wd-mult 10 --data-dir /Users/sunyimac/faces_emore --pretrained ../models/MobileFaceNet/model-y1-arcfaced,18 --prefix ../models/MobileFaceNet/model-y1-arcface\n\n#CUDA_VISIBLE_DEVICES='0' python -u train_softmax.py --network y1 --ckpt 2 --loss-type 4   --lr 0.01 --emb-size 512 --per-batch-size 240 --margin-s 64 --wd 0.00001 --fc7-wd-mult 10 --data-dir /Users/sunyimac/faces_emore --pretrained ../models/MobileFaceNet/model-y1-arcface,62 --prefix ../models/MobileFaceNet/model-y1-arcfaced\n\nCUDA_VISIBLE_DEVICES='0' python -u train_softmax.py --network y1 --ckpt 2 --loss-type 4   --lr 0.00001 --emb-size 512 --per-batch-size 240  --wd 0.00001 --fc7-wd-mult 10 --data-dir /Users/sunyimac/faces_emore --pretrained ../models/MobileFaceNet/model-y1-arcface,75 --prefix ../models/MobileFaceNet/model-y1-arcfaced\n\nUpdate wd=0.000001 trainning is not end. now is the new Accuracy:\ni get new higher Accuracy:\n###### Accuracy\n| dbname | accuracy |\n| ----- |:-----:|\n| lfw |0.99667|\n| cfp_fp |0.94300|\n| age_db30 |0.96700|\n\nUpdate wd=0.0000001 trainning is not end. now is the new Accuracy:\ni get new higher Accuracy:\n###### Accuracy🔥\n| dbname | accuracy |\n| ----- |:-----:|\n| lfw |0.99683|\n| cfp_ff |0.99733|\n| cfp_fp |0.94500|\n| age_db30 |0.96717|\nyou can visit my log file:\nhttps://github.com/qidiso/mobilefacenet-V2/blob/master/retrain0.001.log\n\n# Now Release the models:\n[models:]https://github.com/aidlearning/AidLearning-FrameWork/tree/master/src/facencnn/models\n(reached 99.733 in the cfp-ff、 the 99.68+ in lfw,96.71+ in agedb30)\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fqidiso%2Fmobilefacenet-v2","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Fqidiso%2Fmobilefacenet-v2","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fqidiso%2Fmobilefacenet-v2/lists"}