Zhaoxiang Liu

Beyond Real: Imaginary Extension of Rotary Position Embeddings for Long-Context LLMs

MeanCache: From Instantaneous to Average Velocity for Accelerating Flow Matching Inference

Sparser Block-Sparse Attention via Token Permutation

Mixture of Heterogeneous Grouped Experts for Language Modeling

PathXfer: Few-Shot Visual Fidelity Transfer for Compressed Multi-to-Few Step Sampling

LeMiCa: Lexicographic Minimax Path Caching for Efficient Diffusion-Based Video Generation