HyperAIHyperAI초신경
홈플랫폼문서뉴스연구 논문튜토리얼데이터셋백과사전SOTALLM 모델GPU 랭킹컨퍼런스
전체 검색
소개
HyperAIHyperAI초신경

Command Palette

Search for a command to run...

Image Classification On Webvision 1000

평가 지표

ImageNet Top-1 Accuracy
ImageNet Top-5 Accuracy
Top-1 Accuracy
Top-5 Accuracy

평가 결과

이 벤치마크에서 각 모델의 성능 결과

Paper TitleRepository
MAM (ViT-B/16)--83.6-Improving Image Recognition by Retrieving from Web-Scale Image-Text Data-
CurriculumNet (InceptionResNet-v2)--79.3%93.6%CurriculumNet: Weakly Supervised Learning from Large-Scale Web Images
ERM + CMW-Net-SL (InceptionResNet-V2)--77.9%92.6%CMW-Net: Learning a Class-Aware Sample Weighting Mapping for Robust Deep Learning
NCR+Mixup+DA (ResNet-50)--76.8-Learning with Neighbor Consistency for Noisy Labels
Heteroscedastic (InceptionResNet-v2)--76.6%92.1%Correlated Input-Dependent Label Noise in Large-Scale Image Classification-
MILe (ResNet-50-D)68.786.476.5%90.9%Multi-label Iterated Learning for Image Classification with Label Ambiguity-
SCC (ResNet50-D)70.66%88.46%75.78%91.07%Webly Supervised Image Classification with Self-Contained Confidence
NCR (ResNet-50)--75.7%-Learning with Neighbor Consistency for Noisy Labels
VSGraph-LC (ResNet50)69.42%87.29%75.48%90.15%Webly Supervised Image Classification with Metadata: Automatic Noisy Label Correction via Visual-Semantic Graph
MILe (ResNet-50)67.185.675.2%90.3%Multi-label Iterated Learning for Image Classification with Label Ambiguity-
HAR (InceptionResNet-v2)67.1%86.7%75.0%90.6%Heteroskedastic and Imbalanced Deep Learning with Adaptive Regularization
MentorMix (InceptionResNet-V2)67.5%87.2%74.3%90.5%Beyond Synthetic Noise: Deep Learning on Controlled Noisy Labels
MoPro (ResNet-50)67.8%87.0%73.9%90.0%MoPro: Webly Supervised Learning with Momentum Prototypes
SOMNet (ResNet-50)65.0%85.1%72.2%89.5%Learning from Web Data with Self-Organizing Memory Module-
CurriculumNet (Inception-v2）64.8%84.9%72.1%89.2%CurriculumNet: Weakly Supervised Learning from Large-Scale Web Images
MentorNet (InceptionResNet-V2)62.5%83.0%70.8%88.0%MentorNet: Learning Data-Driven Curriculum for Very Deep Neural Networks on Corrupted Labels
0 of 16 row(s) selected.