本文介绍了FaceX-Zoo的PyTorch工具箱、MixFaceNets的高效网络和PocketNet的超轻量级设计，同时涵盖了LFW等开源数据集和Insightface等重要资源。
----------------------------------------- 开源工程 ----------------------------------------------------
一、FaceX-Zoo：FaceX-Zoo is a PyTorch toolbox for face recognition.
https://github.com/JDAI-CV/FaceX-Zoo
二、MixFaceNets：Extremely Efficient Face Recognition Networks
Paper Arxiv：https://arxiv.org/pdf/2107.13046.pdf
GitHub：https://github.com/fdbtrs/mixfacenets
三、PocketNet: Extreme Lightweight Face Recognition Network using Neural Architecture Search and Multi-Step Knowledge Distillation
Paper Arxiv：https://arxiv.org/pdf/2108.10710.pdf
Github：https://github.com/fdbtrs/PocketNet
------------------------------------------ 开源数据 -----------------------------------------------------------
一、LFW/CFP/AGEDB/IJB-B/IJB-C：https://github.com/IrvingMeng/MagFace
二、Trillionpairs：
http://trillionpairs.deepglint.com/data
该数据集清洗后的标签等：https://github.com/JDAI-CV/FaceX-Zoo/tree/main/training_mode
三、Insightface datasets
差点把insightface的链接给忘了
扫一扫
被折叠的 条评论
为什么被折叠?
到【灌水乐园】发言
