Gradient Surgery

Reported on 4 benchmarks across 1 task · 1 paper · 1 SOTA

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

Methodology4 results

Metric LearningonIn-Shop
R@1· 2022-01-27
92.21
best: 96.7 (Unicom+ViT-L@336px)
SOTA
Dissecting the impact of different loss functions with gradient surgery arXiv:2201.11307
Metric LearningonCARS196
R@1· 2022-01-27
86.5
best: 98.2 (Unicom+ViT-L@336px)
Dissecting the impact of different loss functions with gradient surgery arXiv:2201.11307
Metric Learningon CUB-200-2011
R@1· 2022-01-27
63.8
best: 90.1 (Unicom+ViT-L@336px)
Dissecting the impact of different loss functions with gradient surgery arXiv:2201.11307
Metric LearningonStanford Online Products
R@1· 2022-01-27
82.3
best: 91.2 (Unicom+ViT-L@336px)
Dissecting the impact of different loss functions with gradient surgery arXiv:2201.11307