TY  - RPRT
TI  - Cocktail: Chunk-Adaptive Mixed-Precision Quantization for Long-Context LLM Inference
AU  - Wei Tao
AU  - Bin Zhang
AU  - Xiaoyang Qu
AU  - Jiguang Wan
AU  - Jianzong Wang
PY  - 2025
UR  - https://arxiv.org/abs/2503.23294
ID  - 2503.23294
ER  -