Publications

(* indicates equal contributions)

2024

  1. MiniCache: KV Cache Compression in Depth Dimension for Large Language Models
    Akide Liu ,  Jing Liu ,  Zizheng Pan ,  Yefei He ,  Gholamreza Haffari ,  and  Bohan Zhuang
    In Conference on Neural Information Processing Systems (NeurIPS) , 2024
  2. ZipCache: Accurate and Efficient KV Cache Quantization with Salient Token Identification
    Yefei He ,  Luoming Zhang ,  Weijia Wu ,  Jing Liu ,  Hong Zhou ,  and  Bohan Zhuang
    In Conference on Neural Information Processing Systems (NeurIPS) , 2024
  3. Stitched ViTs are Flexible Vision Backbones
    Zizheng Pan ,  Jing Liu ,  Haoyu He ,  Jianfei Cai ,  and  Bohan Zhuang
    In European Conference on Computer Vision (ECCV) , 2024
  4. Efficient Stitchable Task Adaptation
    Haoyu He ,  Zizheng Pan ,  Jing Liu ,  Jianfei Cai ,  and  Bohan Zhuang
    In Conference on Computer Vision and Pattern Recognition (CVPR) , 2024
  5. CVPR Highlight
    Tfmq-dm: Temporal feature maintenance quantization for diffusion models
    Yushi Huang* ,  Ruihao Gong* ,  Jing Liu ,  Tianlong Chen ,  and  Xianglong Liu
    In Conference on Computer Vision and Pattern Recognition (CVPR) , 2024
    Spotlight (top 11% of the accepted papers)
  6. QLLM: Accurate and Efficient Low-Bitwidth Quantization for Large Language Models
    Jing Liu ,  Ruihao Gong ,  Xiuying Wei ,  Zhiwei Dong ,  Jianfei Cai ,  and  Bohan Zhuang
    In International Conference on Learning Representations (ICLR) , 2024
  7. ICLR Spotlight
    EfficientDM: Efficient Quantization-Aware Fine-Tuning of Low-Bit Diffusion Models
    Yefei He ,  Jing Liu ,  Weijia Wu ,  Hong Zhou ,  and  Bohan Zhuang
    In International Conference on Learning Representations (ICLR) , 2024
    Spotlight (top 5% of the accepted papers)

2023

  1. Pruning self-attentions into convolutional layers in single path
    Haoyu He ,  Jing Liu ,  Zizheng Pan ,  Jianfei Cai ,  Jing Zhang ,  Dacheng Tao ,  and  Bohan Zhuang
    IEEE Transactions on Pattern Analysis and Machine Intelligence (TPAMI), 2023
  2. PTQD: Accurate Post-Training Quantization for Diffusion Models
    Yefei He ,  Luping Liu ,  Jing Liu ,  Weijia Wu ,  Hong Zhou ,  and  Bohan Zhuang
    In Conference on Neural Information Processing Systems (NeurIPS) , 2023
  3. BiViT: Extremely Compressed Binary Vision Transformers
    Yefei He ,  Zhenyu Lou ,  Luoming Zhang ,  Jing Liu ,  Weijia Wu ,  Hong Zhou ,  and  Bohan Zhuang
    In International Conference on Computer Vision (ICCV) , 2023
  4. Single-path bit sharing for automatic loss-aware model compression
    Jing Liu ,  Bohan Zhuang ,  Peng Chen ,  Chunhua Shen ,  Jianfei Cai ,  and  Mingkui Tan
    IEEE Transactions on Pattern Analysis and Machine Intelligence, 2023
  5. A Survey on Efficient Training of Transformers
    Bohan Zhuang ,  Jing Liu ,  Zizheng Pan ,  Haoyu He ,  Yuetian Weng ,  and  Chunhua Shen
    In International Joint Conference on Artificial Intelligence (IJCAI) , 2023
    Survey Track
  6. Dynamic Focus-Aware Positional Queries for Semantic Segmentation
    Haoyu He ,  Jianfei Cai ,  Zizheng Pan ,  Jing Liu ,  Jing Zhang ,  Dacheng Tao ,  and  Bohan Zhuang
    In Conference on Computer Vision and Pattern Recognition (CVPR) , 2023

2022

  1. NeurIPS Spotlight
    EcoFormer: Energy-Saving Attention with Linear Complexity
    Jing Liu* ,  Zizheng Pan* ,  Haoyu He ,  Jianfei Cai ,  and  Bohan Zhuang
    In Conference on Neural Information Processing Systems (NeurIPS) , 2022
    Spotlight (top 5% of the accepted papers)
  2. Less is more: Pay less attention in vision transformers
    Zizheng Pan ,  Bohan Zhuang ,  Haoyu He ,  Jing Liu ,  and  Jianfei Cai
    In AAAI Conference on Artificial Intelligence (AAAI) , 2022

2021

  1. Scalable Vision Transformers With Hierarchical Pooling
    Zizheng Pan ,  Bohan Zhuang ,  Jing Liu ,  Haoyu He ,  and  Jianfei Cai
    In International Conference on Computer Vision (ICCV) , 2021
  2. arXiv
    Sharpness-aware quantization for deep neural networks
    Jing Liu ,  Jianfei Cai ,  and  Bohan Zhuang
    arXiv preprint arXiv:2111.12273, 2021
  3. arXiv
    Mesa: A memory-saving training framework for transformers
    Zizheng Pan ,  Peng Chen ,  Haoyu He ,  Jing Liu ,  Jianfei Cai ,  and  Bohan Zhuang
    arXiv preprint arXiv:2111.11124, 2021
  4. Discrimination-aware network pruning for deep model compression
    Jing Liu* ,  Bohan Zhuang* ,  Zhuangwei Zhuang* ,  Yong Guo ,  Junzhou Huang ,  Jinhui Zhu ,  and  Mingkui Tan*
    IEEE Transactions on Pattern Analysis and Machine Intelligence (TPAMI), 2021
  5. Effective training of convolutional neural networks with low-bitwidth weights and activations
    Bohan Zhuang* ,  Mingkui Tan* ,  Jing Liu* ,  Lingqiao Liu ,  Ian Reid ,  and  Chunhua Shen
    IEEE Transactions on Pattern Analysis and Machine Intelligence (TPAMI), 2021
  6. CVPR Oral
    AQD: Towards Accurate Quantized Object Detection
    Peng Chen* ,  Jing Liu* ,  Bohan Zhuang ,  Mingkui Tan ,  and  Chunhua Shen
    In Conference on Computer Vision and Pattern Recognition (CVPR) , 2021
    Oral Presentation (top 4% of the accepted papers)

2020

  1. Deep transferring quantization
    Zheng Xie* ,  Zhiquan Wen* ,  Jing Liu* ,  Zhiqiang Liu ,  Xixian Wu ,  and  Mingkui Tan
    In European Conference on Computer Vision (ECCV) , 2020
  2. Generative low-bitwidth data free quantization
    Shoukai Xu* ,  Haokun Li* ,  Bohan Zhuang* ,  Jing Liu ,  Jiezhang Cao ,  Chuangrun Liang ,  and  Mingkui Tan
    In European Conference on Computer Vision (ECCV) , 2020

2018

  1. Discrimination-aware Channel Pruning for Deep Neural Networks
    Zhuangwei Zhuang* ,  Mingkui Tan* ,  Bohan Zhuang* ,  Jing Liu* ,  Yong Guo ,  Qingyao Wu ,  Junzhou Huang ,  and  Jinhui Zhu
    In Conference on Neural Information Processing Systems (NeurIPS) , 2018