Communication Efficient Split Learning of ViTs with Attention-based Double Compression

Open in new window