Fetching the paper…

APTQ: Attention-aware Post-Training Mixed-Precision Quantization for Large Language Models · Around