COMCAT: Towards Efficient Compression and Customization of Attention-Based Vision Models

Open in new window