TY  - RPRT
TI  - Fast and Efficient 2-bit LLM Inference on GPU: 2/4/16-bit in a Weight Matrix with Asynchronous Dequantization
AU  - Jinhao Li
AU  - Jiaming Xu
AU  - Shiyao Li
AU  - Shan Huang
AU  - Jun Liu
AU  - Yaoxiu Lian
AU  - Guohao Dai
PY  - 2024
UR  - https://arxiv.org/abs/2311.16442
ID  - 2311.16442
ER  -