List your company
ResearchFormerly WizWand

Zero-shot Image Classification on ImageNet 1k (test)

80.11Accuracy (Zero-shot)Base
15.307632.131348.95565.7787Mar 27, 2026
Updated 7mo ago

Evaluation results

MethodLinks
BaseBackbone=ViT-L/14, Pre-training=TeCoA, epsilon_t=2/255, Evaluation Scenario=Clean2026.0380.11
ET3Backbone=ViT-L/14, Pre-training=TeCoA, epsilon_t=2/255, Evaluation Scenario=Clean2026.0379.82
ET3Backbone=ViT-L/14, Pre-training=TeCoA, epsilon_t=4/255, Evaluation Scenario=Clean2026.0375.2
BaseBackbone=ViT-L/14, Pre-training=TeCoA, epsilon_t=4/255, Evaluation Scenario=Clean2026.0374.91
BaseBackbone=ViT-L/14, Pre-training=FARE, epsilon_t=2/255, Evaluation Scenario=Clean2026.0374.48
ET3Backbone=ViT-L/14, Pre-training=FARE, epsilon_t=2/255, Evaluation Scenario=Clean2026.0374.07
ET3Backbone=ViT-L/14, Pre-training=FARE, epsilon_t=4/255, Evaluation Scenario=Clean2026.0370.97
BaseBackbone=ViT-L/14, Pre-training=FARE, epsilon_t=4/255, Evaluation Scenario=Clean2026.0370.78
ET3Backbone=ViT-L/14, Pre-training=TeCoA, epsilon_t=4/255, Evaluation Scenario=Robust (AA)2026.0353
BaseBackbone=ViT-L/14, Pre-training=TeCoA, epsilon_t=4/255, Evaluation Scenario=Robust (AA)2026.0344.5
ET3Backbone=ViT-L/14, Pre-training=TeCoA, epsilon_t=2/255, Evaluation Scenario=Robust (AA)2026.0344.4
ET3Backbone=ViT-L/14, Pre-training=FARE, epsilon_t=4/255, Evaluation Scenario=Robust (AA)2026.0341.2
BaseBackbone=ViT-L/14, Pre-training=TeCoA, epsilon_t=2/255, Evaluation Scenario=Robust (AA)2026.0337
BaseBackbone=ViT-L/14, Pre-training=FARE, epsilon_t=4/255, Evaluation Scenario=Robust (AA)2026.0334.8
ET3Backbone=ViT-L/14, Pre-training=FARE, epsilon_t=2/255, Evaluation Scenario=Robust (AA)2026.0325.2
BaseBackbone=ViT-L/14, Pre-training=FARE, epsilon_t=2/255, Evaluation Scenario=Robust (AA)2026.0317.8