LeViT-256

Reported on 6 benchmarks across 1 task · 1 paper

Note: results are matched by exact model name. Different papers may use the same name for different model variants.

Computer Vision6 results

Image ClassificationonStanford Cars
Accuracy· 2021-04-02
88.2
best: 96.868 (efficient adaptive ensembling)
LeViT: a Vision Transformer in ConvNet's Clothing for Faster Inference arXiv:2104.01136
Image ClassificationonImageNet V2
Top 1 Accuracy· 2021-04-02
69.9
best: 84.63 (Model soups (BASIC-L))
LeViT: a Vision Transformer in ConvNet's Clothing for Faster Inference arXiv:2104.01136
Image ClassificationonCIFAR-10
Percentage correct· uses extra data· 2021-04-02
98.1
best: 99.5 (ViT-H/14)
LeViT: a Vision Transformer in ConvNet's Clothing for Faster Inference arXiv:2104.01136
Image ClassificationonFlowers-102
Accuracy· uses extra data· 2021-04-02
97.7
best: 99.76 (CCT-14/7x2)
LeViT: a Vision Transformer in ConvNet's Clothing for Faster Inference arXiv:2104.01136
Image ClassificationoniNaturalist 2019
Top-1 Accuracy· 2021-04-02
72.3
best: 88.5 (Hiera-H (448px))
LeViT: a Vision Transformer in ConvNet's Clothing for Faster Inference arXiv:2104.01136
Image ClassificationonImageNet
GFLOPs· 2021-04-02
1.066
best: 1478 (InternImage-H)
LeViT: a Vision Transformer in ConvNet's Clothing for Faster Inference arXiv:2104.01136