AI Paper Atlas

Research map / Benchmarks / ImageNet-1K fine-tuning

ImageNet-1K fine-tuning leaderboard

17 papers at ICML, NeurIPS, ICLR, CVPR, AAAI, ICCV and ECCV, and the journals TPAMI, IJCV and JMLR report results on ImageNet-1K fine-tuning. Best accuracy: 88.1 (MSG-MAE, StyleGANv2-ADA-P, CVPR 2023).

Numbers are as each paper reports them, read by AI from its results tables and not checked by hand. Papers can differ in setting (training data, model size, extra data); the topic column shows what each paper is about. Open the paper before you compare or quote a number.

ImageNet-1K fine-tuning · accuracy

17 papers, each with its best reported number for its own method; higher is better.

#accuracy ↑MethodPaperTopic
188.1MSG-MAE, StyleGANv2-ADA-P
ViT-L
Improving Visual Representation Learning Through Perceptual Understanding
CVPR 2023 · Table: Classification performa
Human and embodied modeling
287.8SuperClass ViT-L/16
ViT-L/16
Classification Done Right for Vision-Language Pre-Training
NeurIPS 2024 · Fine-tuning comparison table
CLIP and zero-shot
387.7CapPa L/14
L/14
Image Captioners Are Scalable Vision Learners Too
NeurIPS 2023 · Table 3
Human and embodied modeling
487.7FD-CLIP
ViT-L/14 CLIP, 224
Improving CLIP Fine-tuning Performance
ICCV 2023 · Table 2
CNN features
586.51SCL-RHE (Ours)
BEiT-3 (pre-trained)
Understanding and Mitigating Human-Labelling Errors in Supervised Contrastive Learning
ECCV 2024 · Table 4
Noisy labels
686.3MVP (300 ep)
ViT-L/16
MVP: Multimodality-Guided Visual Pre-training
ECCV 2022 · Table 1
CLIP and zero-shot
786LUT
ViT-L/16
Learning with Unmasked Tokens Drives Stronger Vision Learners
ECCV 2024 · Table 1
Self-supervised and contrastive
885.8MR-MAE (CLIP teacher, 400 ep)
CViT-B
Mimic before Reconstruct: Enhancing Masked Autoencoders with Feature Mimicking
IJCV 2023 · Table 1
Recognition and long-tail
985.6mc-BEiT (800 ep)
ViT-L/16
Mc-BEiT: Multi-Choice Discretization for Image BERT Pre-training
ECCV 2022 · Table 1
CNN features
1085ConvMAE
ConViT-B
MCMAE: Masked Convolution Meets Masked Autoencoders
NeurIPS 2022 · Table 1
CNN features
7 more papers on this leaderboard, the baselines reported on it and a CSV download are free with a scime account.
Sign in free to see all

Not affiliated with any conference, journal or the benchmark's authors. Corrections: contact@latenticlabs.com.