Haofeng Huang

SageAttention2: Efficient Attention with Thorough Outlier Smoothing and Per-thread INT4 Quantization

Faster Video Diffusion with Trainable Sparse Attention

SageAttention3: Microscaling FP4 Attention for Inference and An Exploration of 8-Bit Training

SpargeAttention: Accurate and Training-free Sparse Attention Accelerating Any Model Inference

Mixture of Attention Spans: Optimizing LLM Inference Efficiency with Heterogeneous Sliding-Window Lengths

XAttention: Block Sparse Attention with Antidiagonal Scoring

ViDiT-Q: Efficient and Accurate Quantization of Diffusion Transformers for Image and Video Generation

MoA: Mixture of Sparse Attention for Automatic Large Language Model Compression