V-MoE-L/16 (Every-2)

Reported on 5 benchmarks across 2 tasks · 1 paper

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

Computer Vision5 results

Image ClassificationonJFT-300M
prec@1· 2021-06-10
57.65
best: 60.62 (V-MoE-H/14 (Every-2))
Scaling Vision with Sparse Mixture of Experts arXiv:2106.05974
Image ClassificationonImageNet - 5-shot
Top 1 Accuracy· 2021-06-10
77.1
best: 82.78 (ViT-MoE-15B (Every-2))
Scaling Vision with Sparse Mixture of Experts arXiv:2106.05974
Image ClassificationonImageNet - 1-shot
Top 1 Accuracy· 2021-06-10
62.41
best: 68.66 (ViT-MoE-15B (Every-2))
Scaling Vision with Sparse Mixture of Experts arXiv:2106.05974
Few-Shot Image ClassificationonImageNet - 5-shot
Top 1 Accuracy· 2021-06-10
77.1
best: 82.78 (ViT-MoE-15B (Every-2))
Scaling Vision with Sparse Mixture of Experts arXiv:2106.05974
Few-Shot Image ClassificationonImageNet - 1-shot
Top 1 Accuracy· 2021-06-10
62.41
best: 68.66 (ViT-MoE-15B (Every-2))
Scaling Vision with Sparse Mixture of Experts arXiv:2106.05974