Gongfan Fang

Ph.D. Candidate | xML Lab | National University of Singapore.
me.jpeg

Hi there! I'm Gongfan Fang, currently a last-year Ph.D. Candidate at the xML Lab @ National University of Singapore, under the supervision of Prof. Xinchao Wang (Presidential Young Professor) . Before joining xML Lab, I earned my Bachelor’s degree in 2019 and completed my Master’s degree in 2022 at the Visual Intelligence and Pattern Analysis (VIPA) Lab @ Zhejiang University, adviced by Prof. Mingli Song.

My research focuses on LLMs, Diffusion and Efficient Generative Models. My previous work includes Hybrid Reasoning LLM, Efficient LLM, and Fast Diffusion Transformers. I'm also actively contributing to several projects such as Torch-Pruning, a top framework for compressing foundation models.

I'm currently on the job market and open to both academic and industrial opportunities for 2026. Please feel free to reach out via email (gongfan at u.nus.edu) if there’s a potential fit.



News

Feb, 2025 🍺 One first-author paper TinyFusion and three co-authored papers were accepted by CVPR’25.
Dec, 2024 🎵 I’m deeply honored to be awarded the 2024 ByteDance Scholarship (10~15 recipients per year).
Sep, 2024 🚀 Two first-author papers MaskLLM (Spotlight) and Remix-DiT were accepted by NeurIPS’24.

Selected Publications

  1. fang2025thinkless.png
    ArXiv’25
    Thinkless: LLM Learns When to Think
    Gongfan Fang, Xinyin Ma, and Xinchao Wang
    arXiv preprint arXiv:2505.13379, 2025
    National University of Singapore
    Hybrid Reasoning LLMs via Decoupled GRPO | Cuts 50%-90% of Unnecessary Thinking | Stop Overthinking 1+1=?
    arXiv Bib Code
  2. fang2024tinyfusion.png
    CVPR’25
    TinyFusion: Diffusion Transformers Learned Shallow
    Gongfan Fang, Kunjun Li, Xinyin Ma, and Xinchao Wang
    Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2025
    National University of Singapore
    CVPR’25 Highlight (3%) | Learning Fast DiTs at 7% Training Costs | 2x Faster Inference
    arXiv Bib Code
  3. fang2024maskllm.png
    NeurIPS’24
    MaskLLM: Learnable Semi-structured Sparsity for Large Language Models
    Gongfan Fang, Hongxu Yin, Saurav Muralidharan, Greg Heinrich, Jeff Pool, Jan Kautz, Pavlo Molchanov, and Xinchao Wang
    Advances in Neural Information Processing Systems, 2024
    NVIDIA Research, National University of Singapore
    NeurIPS’24 Spotlight (2%) | Pre-training of Sparse LLM | The First Scalable Algorithm for N:M Sparsity in LLMs | 1.4x Faster with 30%+ Memory Saving
    arXiv Webpage Bib Code
  4. fang2023depgraph.png
    CVPR’23
    DepGraph: Towards Any Structural Pruning
    Gongfan Fang, Xinyin Ma, Mingli Song, Michael Bi Mi, and Xinchao Wang
    Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2023
    National University of Singapore, Zhejiang University, Huawei
    400+ Citations | #Model-Compression Top-5 | Pruning of Foundation Models
    arXiv Bib PDF Code
  5. ma2023deepcache.png
    CVPR’24
    DeepCache: Accelerating Diffusion Models for Free
    Xinyin Ma, Gongfan Fang, and Xinchao Wang
    Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2024
    National University of Singapore
    Training-free and almost lossless | 2-7x Speedup on Diffusion Models
    arXiv Bib Code
  6. ma2023llm_pruner.png
    NeurIPS’23
    LLM-Pruner: On the Structural Pruning of Large Language Models
    Xinyin Ma, Gongfan Fang, and Xinchao Wang
    Advances in Neural Information Processing Systems, 2023
    National University of Singapore
    The First Structured Pruning Method for LLMs | Low-cost Pruning and Training
    arXiv Bib Code

Education

2022.07 - 2026.06 - Ph.D. in Electrical and Computer Engineering, National University of Singapore.

2019.09 - 2022.04 - M.Eng. in Computer Science, College of Computer Science and Technology, Zhejiang University.

2015.09 - 2019.06 - B.S. in Computer Science, College of Computer Science and Technology, Zhejiang University.