Time- Memory- and Parameter-Efficient Visual Adaptation
Open paper
Benchmark
Model
Rank
Results
action-classification-on-kinetics-400
LoSA (ViViT-e)
–
Acc@1: 86.2
image-classification-on-imagenet
LoSA (ViT-g)
–
Top 1 Accuracy: 89.0%
Rank counts only results with a code link.