AI Paper Atlas

Research map / Benchmarks / CIFAR-100 clean

CIFAR-100 clean leaderboard

17 papers at ICML, NeurIPS, ICLR, CVPR, AAAI, ICCV and ECCV, and the journals TPAMI, IJCV and JMLR report results on CIFAR-100 clean. Best accuracy: 81.55 (AdAutoMix, ICLR 2024).

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.

CIFAR-100 clean · accuracy

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

#accuracy ↑MethodPaperTopic
181.55AdAutoMix
ResNet18
Adversarial AutoMixup
ICLR 2024 · Robustness table
Recognition and long-tail
280.9Noisy Feature Mixup (α=1.0)
Wide ResNet-18
Noisy Feature Mixup
ICLR 2022 · Table: Robustness of Wide-ResN
Noisy labels
380.29Distilled KF
ResNet18
Forget Me Not: Fighting Local Overfitting With Knowledge Fusion and Distillation
TPAMI 2025 · Table 2
Knowledge distillation
479.9SNGP
Wide ResNet-28-10
Simple and Principled Uncertainty Estimation with Deterministic Deep Learning via Distance Awareness
NeurIPS 2020 · CIFAR-100 results table
Uncertainty and Gaussian processes
576.86Gabor-layered
WideResNet
Gabor Layers Enhance Network Robustness
ECCV 2020 · Table 1
Certified and empirical robustness
673.9Proposed (feature scattering)Defense Against Adversarial Attacks Using Feature Scattering-based Adversarial Training
NeurIPS 2019 · Table 2b
Attacks on text and graphs
772.19CE+SRLearning With Noisy Labels via Sparse Regularization
ICCV 2021 · Table 1
Noisy labels
870.83CEε+MAE$\epsilon$-Softmax: Approximating One-Hot Vectors for Mitigating Label Noise
NeurIPS 2024 · Table: CIFAR-10/100 symmetric
Noisy labels
970.3ProSelfLC (B=10)
ResNet-44
ProSelfLC: Progressive Self Label Correction for Training Robust Deep Neural Networks
CVPR 2021 · Table: standard setting
Noisy labels
1067.7PC loss with logit constraints
VGG-13
Improving Adversarial Robustness via Probabilistically Compact Loss with Logit Constraints
AAAI 2021 · Table: CIFAR-100 and Tiny Imag
Black-box and transfer attacks
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.