FDSML: Formula-Driven Supervised Metric Learning with Parameter-Aware Triplet Loss

Open paper
Benchmark Model Rank Results
image-classification-on-cifar-10FDSML (ViT-Base, each) (extra data)#63Percentage correct: 98.23
image-classification-on-flowers-102FDSML (ViT-Tiny, each) (extra data)#64Accuracy: 74.21
image-classification-on-cifar-100FDSML (ViT-Base, each) (extra data)#115Percentage correct: 83.3
image-classification-on-imagenetFDSML (ViT-Tiny, diff) (extra data)#1025Top 1 Accuracy: 73.97%

Rank counts only results with a code link.