MoSA: Mixture of Sparse Adapters for Visual Efficient Tuning
Open paper
Benchmark
Model
Rank
Results
object-counting-on-clevr-count
MoSA ViT-B/16 (VTAB-1k)
#10
Accuracy: 71.5
monocular-depth-estimation-on-clevr-distance
MoSA ViT-B/16 (VTAB-1k)
#11
Accuracy: 58.1
Rank counts only results with a code link.