Y our representations are in the network: composable and parallel adaptation for large scale models

On the ViT -L/16 architecture, our experiments show that a single adapter, 1.3% of the full model, is able to reach full fine-tuning accuracy on average across 11 challenging downstream classification tasks. Compared with other forms of parameter-efficient adaptation, the isolated nature of the InCA adaptation is computationally desirable for large-scale models. For instance, we adapt ViT -G/14 (1.8B+ parameters) quickly with 20+ adapters in parallel on a single V100 GPU (76% GPU memory reduction) and exhaustively identify its