About me

I am Zhijun Tu, a researcher at Huawei Fundation Model Department. My research interests lie in efficient systems for large language model and multimodal generation, and model compression.

I received my B.S. and M.S. degrees from Xi’an Jiaotong University in 2019 and 2022, respectively, where I was advised by Prof. Pengju Ren. In 2021, I was a research intern at Huawei Noah’s Ark Lab, advised by Dr. Xinghao Chen and led by Dr. Yunhe Wang.

My primary focus is to advance modern foundation models and generative AI for real-world mobile ecosystems, ultimately democratizing powerful intelligence on everyday devices. To this end, our team is seeking self-motivated Research Interns to collaborate on pioneering projects in efficient multimodal systems. If you are driven to advance cutting-edge multimodal systems and next-generation generative AI, feel free to drop me an email with your CV.

🔥 News

• [2026.05] One paper accepted by ICML 2026 (BinaryLLM).
• [2026.03] One paper accepted by IEEE Transactions on Image Processing (TAD-SR).
• [2025.11] One paper accepted by AAAI 2026 (MoR).
• [2025.05] One paper accepted by ICML 2025 (Diff-MoE).
• [2025.02] One paper accepted by CVPR 2025 (RaSS).
• [2025.01] Two papers accepted by ICLR 2025 (CBQ, AugKD).
• [2024.11] One paper accepted by AAAI 2025 (DiT-SR).
• [2024.09] One paper accepted by NeurIPS 2024 (U-DiTs).
• [2024.07] One paper accepted by ECCV 2024 (PQ-SAM).
• [2024.02] Released our new survey paper: “A Survey on Transformer Compression”.
• [2023.09] One paper accepted by NeurIPS 2023 Track on Datasets and Benchmarks.
• [2023.04] Won the Winner Award in NTIRE Challenge on Image Denoising@CVPR2023.
• [2023.02] One paper accepted by CVPR 2023 (PTQ4SR).
• [2022.07] One paper accepted by ECCV 2022 (AdaBin).

Selected Publications

Rethinking 1-bit Optimization Leveraging Pre-trained Large Language Models

Zhijun Tu, Jian Li, Yuanyuan Xi, Siqi Liu, Chuanjian Liu, Hanting Chen, Jie Hu, Yunhe Wang

ICML 2026 [Paper]

Mixture of Ranks with Degradation-Aware Routing for One-Step Real-World Image Super-Resolution

Xiao He, Zhijun Tu^†, Kun Cheng, Mingrui Zhu, Jie Hu, Nannan Wang, Xinbo Gao

^†Project Leader, AAAI 2026 [Paper]

One Step Diffusion-based Super-Resolution with Time-Aware Distillation

Xiao He, Haoao Tang, Zhijun Tu^†, Junchao Zhang, Kun Cheng, Hanting Chen, Yong Guo, Mingrui Zhu, Jie Hu, Nannan Wang, Xinbo Gao

^†Project Leader, IEEE TIP [Paper] [Code]

Diff-MoE: Diffusion Transformer with Time-Aware and Space-Adaptive Experts

Kun Cheng*, Xiao He*, Lei Yu, Zhijun Tu^†, Mingrui Zhu, Nannan Wang, Xinbo Gao, Jie Hu

^†Project Leader, ICML 2025 [Paper] [Code]

RaSS: Improving Denoising Diffusion Samplers with Reinforced Active Sampling Scheduler

Xin Ding, Lei Yu, Xin Li, Zhijun Tu^†, Hanting Chen, Jie Hu, Zhibo Chen

^†Project Leader, CVPR 2025 [Paper]

CBQ: Cross-block Quantization for Large Language Models

Xin Ding, Xiaoyu Liu, Zhijun Tu^†, Yun Zhang, Wei Li, Jie Hu, Hanting Chen, Yehui Tang, Zhiwei Xiong, Baoqun Yin, Yunhe Wang

^†Project Leader, ICLR 2025 Spotlight [Paper]

Effective Diffusion Transformer Architecture for Image Super-Resolution

Kun Cheng*, Lei Yu*, Zhijun Tu^†, Xiao He, Liyu Chen, Yong Guo, Mingrui Zhu, Nannan Wang, Xinbo Gao, Jie Hu

^†Project Leader, AAAI 2025 [Paper] [Code]

U-DiTs: Downsample Tokens in U-Shaped Diffusion Transformers

Yuchuan Tian*, Zhijun Tu*, Hanting Chen, Jie Hu,Chao Xu, Yunhe Wang

NeurIPS 2024 [Paper] [Code]

A Survey on Transformer Compression

Yehui Tang, Yunhe Wang, Jianyuan Guo, Zhijun Tu, Kai Han, Hailin Hu, and Dacheng Tao

Arxiv 2024 [Paper]

IPT-V2: Efficient Image Processing Transformer using Hierarchical Attentions

Zhijun Tu*, Kunpeng Du*, Hanting Chen, Hailing Wang, Wei Li, Jie Hu, Yunhe Wang

Winner Award@NTIRE, CVPR 2023 [Paper] [Challenge Report] [Award]

Toward Accurate Post-Training Quantization for Image Super Resolution

Zhijun Tu, Jie Hu, Hanting Chen, Yunhe Wang

CVPR 2023 [Paper] [Code]

AdaBin: Improving Binary Neural Networks with Adaptive Binary Sets

Zhijun Tu, Xinghao Chen, Pengju Ren, Yunhe Wang

ECCV 2022 [Paper] [Code]