一作NeurIPS 2026Poster
AdaKVQ: Adaptive Mixed-Precision KV Cache Quantization For Efficient Reasoning Models
Tao Zhang · Yun Dai · Ziqian Zeng
首次获取该候选人的完整档案及 StartLux 评估共 20 积分；每位用户单独解锁，之后查看免费。主动刷新 20 积分。
实时查找电子邮箱，费用不超过20积分；未找到邮箱不扣费。
Tao Zhang · Yun Dai · Ziqian Zeng
Tao Zhang · Ziqian Zeng · Hao Peng · Huiping Zhuang · Cen Chen
Long Chain-of-Thought (CoT) reasoning has significantly advanced the capabilities of Large Language Models (LLMs), but this progress is accompanied by substantial memory and latency overhead from the extensive Key-Value (KV) cache. Although KV cache quantization is a promising compression techniq...