Partial FC: Training 10 Million Identities on a Single Machine

[2010.05222] Partial FC: Training 10 Million Identities on a Single Machine Skip to main content Search Submit Donate Log in Search arXiv Press Enter to search · Advanced search Computer Science > Computer Vision and Pattern Recognition arXiv:2010.05222 (cs) [Submitted on 11 Oct 2020 ( v1 ), last revised 15 Aug 2026 (this version, v4)] Title: Partial FC: Training 10 Million Identities on a Single Machine Authors: Xiang An , Xuhan Zhu , Yang Xiao , Lan Wu , Ming Zhang , Yuan Gao , Bin Qin , Debing Zhang , Ying Fu , Jiankang Deng View a PDF of the paper titled Partial FC: Training 10 Million Identities on a Single Machine, by Xiang An and 9 other authors View PDF Abstract: Training face recognition models with millions of identities is challenging because classifier storage, logit memory, and computation grow linearly with the number of classes, eventually making full softmax impractical even when the backbone itself fits comfortably in memory. We present Partial FC (PFC), a scalable approximation to large-class softmax that preserves every positive class center while activating only a sampled subset of negative centers in each mini-batch. This asymmetric treatment retains every target term while avoiding exhaustive interaction with millions of mostly uninformative negatives. Our distributed implementation partitions the classifier across GPUs and samples within each owned shard, so sampling reduces local matrix multiplication and logit storage while sharding eliminates class-gradient synchronization across workers. Together, these properties reduce GPU-resident classifier memory, computation, and class-dependent communication without feature-based hard-negative retrieval. End-to-end system benchmarks demonstrate efficient scaling to massive class spaces, including 64 million classes on a single eight-GPU machine. Across large-scale face-recognition datasets, moderate sampling rates maintain competitive recognition accuracy while substantially improving training efficiency. Our best PFC configurations achieve 97.2\% TAR on IJB-C at FAR $=10^{-4}$ and 94.0\% TAR on ICCV21-MFR at FAR $=10^{-6}$. Beyond clean training data, PFC is robust to inter-class conflicts, label noise, and long-tailed identity distributions: under 40\% label noise, PFC with conflict filtering raises ICCV21-MFR TAR from 43.9\% to 80.2\%, while on long-tailed data PFC improves TAR from 87.4\% to 92.0\%. These results establish positive-preserving negative sampling as an effective foundation for scalable, accurate, and robust identity classification. Comments: 8 pages, 9 figures Subjects: Computer Vision and Pattern Recognition (cs.CV) ; Distributed, Parallel, and Cluster Computing (cs.DC) Cite as: arXiv:2010.05222 [cs.CV] (or arXiv:2010.05222v4 [cs.CV] for this version) https://doi.org/10.48550/arXiv.2010.05222 Focus to learn more arXiv-issued DOI via DataCite Submission history From: Xiang An [ view email ] [v1] Sun, 11 Oct 2020 11:15:26 UTC (1,058 KB) [v2] Sat, 23 Jan 2021 05:25:06 UTC (1,058 KB) [v3] Mon, 3 Aug 2026 07:53:31 UTC (419 KB) [v4] Sat, 15 Aug 2026 05:41:40 UTC (604 KB) Full-text links: Access Paper: View a PDF of the paper titled Partial FC: Training 10 Million Identities on a Single Machine, by Xiang An and 9 other authors View PDF TeX Source view license Current browse context: cs.CV < prev | next > new | recent | 2020-10 Change to browse by: cs cs.DC References & Citations NASA ADS Google Scholar Semantic Scholar DBLP - CS Bibliography listing | bibtex Yang Xiao Lan Wu Ming Zhang Yuan Gao Ying Fu export BibTeX citation Loading... BibTeX formatted citation × loading... Data provided by: Bookmark Bibliographic Tools Bibliographic and Citation Tools Bibliographic Explorer Toggle Bibliographic Explorer ( What is the Explorer? ) Connected Papers Toggle Connected Papers ( What is Connected Papers? ) Litmaps Toggle Litmaps ( What is Litmaps? ) scite.ai Toggle scite Smart Citations ( What are Smart Citations? ) Code, Data, Media Code, Data and Media Associated with this Article alphaXiv Toggle alphaXiv ( What is alphaXiv? ) Links to Code Toggle CatalyzeX Code Finder for Papers ( What is CatalyzeX? ) DagsHub Toggle DagsHub ( What is DagsHub? ) GotitPub Toggle Gotit.pub ( What is GotitPub? ) Huggingface Toggle Hugging Face ( What is Huggingface? ) ScienceCast Toggle ScienceCast ( What is ScienceCast? ) Demos Demos Replicate Toggle Replicate ( What is Replicate? ) Spaces Toggle Hugging Face Spaces ( What is Spaces? ) Spaces Toggle TXYZ.AI ( What is TXYZ.AI? ) Related Papers Recommenders and Search Tools Link to Influence Flower Influence Flower ( What are Influence Flowers? ) Core recommender toggle CORE Recommender ( What is CORE? ) Author Venue Institution Topic About arXivLabs arXivLabs: experimental projects with community collaborators arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website. Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them. Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs . Which authors of this paper are endorsers? | Disable MathJax ( What is MathJax? ) We gratefully acknowledge support from our major funders , member institutions , , and all contributors. About · Help · Contact · Subscribe · Copyright · Privacy · Accessibility · Operational Status (opens in new tab) Major funding support from

Related documents

• Quantifying Memorization and Privacy Risks in Genomic Language Models#629935
• Quantifying Memorization and Privacy Risks in Genomic Language Models#628073
• Machine Unlearning for the XGBoost Model with Network Intrusion Datasets#287131
• ICOR: improving codon optimization with recurrent neural networks#19744
• A Harmonized, Global, Cross-Script Corpus of 1.05 Million Business Names Labeled for Personal Naming#403477