HyperAI超神经

首页算力平台文档资讯论文教程数据集百科 SOTA LLM 模型天梯 GPU 天梯顶会

中文

HyperAI超神经

Face Identification On Megaface

评估指标

Accuracy

评测结果

各个模型在此基准测试上的表现结果

		Paper Title	Repository
Cos+UNPG	99.27%	Unified Negative Pair Generation toward Well-discriminative Feature Space for Face Recognition
PartialFC + Glint360K + R100	99.10%	Partial FC: Training 10 Million Identities on a Single Machine
Arc+UNPG	98.82%	Unified Negative Pair Generation toward Well-discriminative Feature Space for Face Recognition
Prodpoly	98.78%	Deep Polynomial Neural Networks
GhostFaceNetV2-1	98.64%	GhostFaceNets: Lightweight Face Recognition Model From Cheap Operations	-
ArcFace + MS1MV2 + R100 + R	98.35%	ArcFace: Additive Angular Margin Loss for Deep Face Recognition
Mag+UNPG	98.03%	Unified Negative Pair Generation toward Well-discriminative Feature Space for Face Recognition
SV-AM-Softmax	97.2%	Support Vector Guided Softmax Loss for Face Recognition
CosFace	82.72%	CosFace: Large Margin Cosine Loss for Deep Face Recognition
SphereFace (3-patch ensemble)	75.766%	SphereFace: Deep Hypersphere Embedding for Face Recognition
Light CNN-29	73.749%	A Light CNN for Deep Face Representation with Noisy Labels
SphereFace (single model)	72.729%	SphereFace: Deep Hypersphere Embedding for Face Recognition
FaceNet	70.49%	FaceNet: A Unified Embedding for Face Recognition and Clustering

0 of 13 row(s) selected.

Face Identification On Megaface | SOTA | HyperAI超神经