Alert button
Picture for Xiuying Wei

Xiuying Wei

Alert button

QLLM: Accurate and Efficient Low-Bitwidth Quantization for Large Language Models

Oct 12, 2023
Jing Liu, Ruihao Gong, Xiuying Wei, Zhiwei Dong, Jianfei Cai, Bohan Zhuang

Figure 1 for QLLM: Accurate and Efficient Low-Bitwidth Quantization for Large Language Models
Figure 2 for QLLM: Accurate and Efficient Low-Bitwidth Quantization for Large Language Models
Figure 3 for QLLM: Accurate and Efficient Low-Bitwidth Quantization for Large Language Models
Figure 4 for QLLM: Accurate and Efficient Low-Bitwidth Quantization for Large Language Models
Viaarxiv icon

Lossy and Lossless (L$^2$) Post-training Model Size Compression

Aug 08, 2023
Yumeng Shi, Shihao Bai, Xiuying Wei, Ruihao Gong, Jianlei Yang

Figure 1 for Lossy and Lossless (L$^2$) Post-training Model Size Compression
Figure 2 for Lossy and Lossless (L$^2$) Post-training Model Size Compression
Figure 3 for Lossy and Lossless (L$^2$) Post-training Model Size Compression
Figure 4 for Lossy and Lossless (L$^2$) Post-training Model Size Compression
Viaarxiv icon

Outlier Suppression+: Accurate quantization of large language models by equivalent and optimal shifting and scaling

Apr 18, 2023
Xiuying Wei, Yunchen Zhang, Yuhang Li, Xiangguo Zhang, Ruihao Gong, Jinyang Guo, Xianglong Liu

Figure 1 for Outlier Suppression+: Accurate quantization of large language models by equivalent and optimal shifting and scaling
Figure 2 for Outlier Suppression+: Accurate quantization of large language models by equivalent and optimal shifting and scaling
Figure 3 for Outlier Suppression+: Accurate quantization of large language models by equivalent and optimal shifting and scaling
Figure 4 for Outlier Suppression+: Accurate quantization of large language models by equivalent and optimal shifting and scaling
Viaarxiv icon

Outlier Suppression: Pushing the Limit of Low-bit Transformer Language Models

Sep 27, 2022
Xiuying Wei, Yunchen Zhang, Xiangguo Zhang, Ruihao Gong, Shanghang Zhang, Qi Zhang, Fengwei Yu, Xianglong Liu

Figure 1 for Outlier Suppression: Pushing the Limit of Low-bit Transformer Language Models
Figure 2 for Outlier Suppression: Pushing the Limit of Low-bit Transformer Language Models
Figure 3 for Outlier Suppression: Pushing the Limit of Low-bit Transformer Language Models
Figure 4 for Outlier Suppression: Pushing the Limit of Low-bit Transformer Language Models
Viaarxiv icon

QDrop: Randomly Dropping Quantization for Extremely Low-bit Post-Training Quantization

Mar 11, 2022
Xiuying Wei, Ruihao Gong, Yuhang Li, Xianglong Liu, Fengwei Yu

Figure 1 for QDrop: Randomly Dropping Quantization for Extremely Low-bit Post-Training Quantization
Figure 2 for QDrop: Randomly Dropping Quantization for Extremely Low-bit Post-Training Quantization
Figure 3 for QDrop: Randomly Dropping Quantization for Extremely Low-bit Post-Training Quantization
Figure 4 for QDrop: Randomly Dropping Quantization for Extremely Low-bit Post-Training Quantization
Viaarxiv icon