AI Paper Atlas

Research map / Benchmarks / AWA2 GZSL

AWA2 GZSL leaderboard

31 papers at ICML, NeurIPS, ICLR, CVPR, AAAI, ICCV and ECCV, and the journals TPAMI, IJCV and JMLR report results on AWA2 GZSL. Best harmonic mean H: 92.3 (ZeroDiff++, TPAMI 2026).

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.

AWA2 GZSL · harmonic mean H

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

#harmonic mean H ↑MethodPaperTopic
192.3ZeroDiff++
ResNet-101
ZeroDiff++: Generative Test-Time Adaptation for Zero-shot Learning
TPAMI 2026 · Table 2
Zero-/few-shot and meta-learning
286.9ZeroMamba + PUFE
ZeroMamba
When Generalized Zero-Shot Learning Meets PU Learning: A Plug-and-Play Framework for Seen-Class Bias Mitigation
ICML 2026 · Table 1
Zero-/few-shot and meta-learning
382.4SAGE
ViT
SAGE: Structured Attribute-Guided Enhancement for GZSL
AAAI 2026 · Table 1
Zero-/few-shot and meta-learning
474.9SVIP
ViT-base
SVIP: Semantically Contextualized Visual Patches for Zero-Shot Learning
ICCV 2025 · Table 1
Zero-/few-shot and meta-learning
572.5TransZero++
ResNet101
TransZero++: Cross Attribute-Guided Transformer for Zero-Shot Learning
TPAMI 2022 · Table 1
Zero-/few-shot and meta-learning
671APN+CVCAttribute Prototype Network for Any-Shot Learning
IJCV 2022 · Table 2
Zero-/few-shot and meta-learning
770.3Ours (gamma=0.95)
f-CLSWGAN unseen expert
A Boundary Based Out-of-Distribution Classifier for Generalized Zero-Shot Learning
ECCV 2020 · Table 3
Zero-/few-shot and meta-learning
870CE-GZSL
ResNet-101 features
Semantic Contrastive Embedding for Generalized Zero-Shot Learning
IJCV 2022 · Table 1
Zero-/few-shot and meta-learning
967.4DE-VAE
ResNet101 features
A Variational Autoencoder with Deep Embedding Model for Generalized Zero-Shot Learning
AAAI 2020 · Table 2
Zero-/few-shot and meta-learning
1067.1FREE
ResNet-101 features
FREE: Feature Refinement for Generalized Zero-Shot Learning
ICCV 2021 · Table 1
Zero-/few-shot and meta-learning
10 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

AWA2 GZSL · H

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

#H ↑MethodPaperTopic
187.4GenZSLGenZSL: Generative Zero-Shot Learning Via Inductive Variational Autoencoder
ICML 2025 · Table: State-of-the-art compar
Zero-/few-shot and meta-learning
277.6I2MVFormer (3-LLM + Wiki)
ViT/B16
I2MVFormer: Large Language Model Generated Multi-View Document Supervision for Zero-Shot Image Classification
CVPR 2023 · Table 1
Zero-/few-shot and meta-learning
376.5ZeroMamba-Small
VMamba-Small
ZeroMamba: Exploring Visual State Space Model for Zero-Shot Learning
AAAI 2025 · Table 1
CLIP and zero-shot
474.9AENet
ViT-Base
Attend and Enrich: Enhanced Visual Prompt for Zero-Shot Learning
AAAI 2025 · Table 1
CLIP and zero-shot
574.2f-VAEGAN + DSP
f-VAEGAN
Evolving Semantic Prototype Improves Generative Zero-Shot Learning
ICML 2023 · Table 1
Zero-/few-shot and meta-learning
670.2TransZero
ResNet101
TransZero: Attribute-Guided Transformer for Zero-Shot Learning
AAAI 2022 · Table 1
CLIP and zero-shot
770CE-GZSLContrastive Embedding for Generalized Zero-Shot Learning
CVPR 2021 · Table 1
Zero-/few-shot and meta-learning
867.7MSDN
ResNet101
MSDN: Mutually Semantic Distillation Network for Zero-Shot Learning
CVPR 2022 · Table 1
Zero-/few-shot and meta-learning
966.4APNet
ResNet-101
Attribute Propagation Network for Graph Zero-Shot Learning
AAAI 2020 · GZSL results table
Zero-/few-shot and meta-learning
1066.1TDCSSNon-Generative Generalized Zero-Shot Learning via Task-Correlated Disentanglement and Controllable Samples Synthesis
CVPR 2022 · Table: Performance comparisons
Zero-/few-shot and meta-learning
1 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.