Head-wise Shareable Attention for Large Language Models

Re-assign community