Tianhui Song (宋天惠)
Multimedia Computing Group
Department of Computer Science and Technology
Nanjing University
Email: tianhuisong [at] smail.nju.edu.cn

About Me

I am an MS at Department of Computer Science and Technology, Nanjing University, under the supervision of Prof. Limin Wang. I received the B.S. degree from Nanjing University in 2023.
My current research interests lies in Multi-Modal Understanding and Generation.

News

Publications [ Google Scholar ] [ Github ]

Scaling Unified Multimodal Models to Long-Horizon Interleaved Generation
Z. Peng, Y. Cui, W. Lin, Z. Li, Q. Gu, T. Song, C. Jia, G. Zhang, X. Qiu, Q. Wang, M. Yang, Z. Zhong, Y. Tian, H. Liu, J. He
in arXiv preprint arXiv:2610.17080.
[ Paper ]
HYDRA-X: Native Unified Multimodal Models with Holistic Visual Tokenizers
G. Zhang, X. Qiu, Y. Cui, T. Song, C. Li, J. Li, T. Huang, X. Zhang, Y. Li, J. Wu, M. Yang, Z. Zhong, L. Bo, L. Wang
in Conference on Neural Information Processing Systems (NeurIPS), 2026.
[ Paper ]
Towards Pixel-Level VLM Perception via Simple Points Prediction
T. Song, H. Lu, H. Yang, L. Sui, H. Wu, Z. Zhou, Z. Huang, Y. Bao, Y. Charles, X. Zhou, L. Wang
in arXiv preprint arXiv:2602.02276.
[ Paper ] [ Project Page ]
DMM: Building a Versatile Image Generation Model via Distillation-Based Model Merging
T. Song, W. Feng, S. Wang, X. Li, T. Ge, B. Zheng, L. Wang
in arXiv preprint arXiv:2504.07491.
[ Paper ] [ Code ]
Differentiable Solver Search for Fast Diffusion Sampling
S. Wang, Z. Li, Q. Zhang, T. Song, X. Li, T. Ge, B. Zheng, L. Wang
in International Conference on Machine Learning (ICML), 2025.
[ Paper ] [ Code ]
FlowDCN: Exploring DCN-like Architectures for Fast Image Generation with Arbitrary Resolution
S. Wang, Z. Li, T. Song, X. Li, T. Ge, B. Zheng, L. Wang
in Conference on Neural Information Processing Systems (NeurIPS), 2024.
[ Paper ] [ Code ]
SPLAM: Accelerating Image Generation with Sub-Path Linear Approximation Model
T. Song, C. Xu, W. Feng, X. Li, T. Ge, B. Zheng, L. Wang
in European Conference on Computer Vision (ECCV), 2024. (Oral)
[ Paper ] [ Code ]
MixFormerV2: Efficient Fully Transformer Tracking
T. Song, Y. Cui, G. Wu, L. Wang
in Conference on Neural Information Processing Systems (NeurIPS), 2023.
[ Paper ] [ Code ]

Last Updated on 21st Oct., 2026

Published with GitHub Pages