
CBQ introduces a cross-block reconstruction method for large language models, achieving superior low-bit quantization and outperforming existing methods.


coremsg

cbq-cross-block-quantization-for-large-language-models


CBQ: Cross-Block Quantization for Large Language Models


title_rewrite


An accurate and efficient low-bitwidth post-training quantization method, QLLM, is proposed to address the challenge of activation outliers in quantizing large language models.


accurate-and-efficient-low-bitwidth-quantization-for-large-language-models


Accurate and Efficient Low-Bitwidth Quantization for Large Language Models
