PinnedLoading

1. thu-ml/SageAttentionthu-ml/SageAttentionPublic

   Quantized Attention achieves speedup of 2-5x and 3-11x compared to FlashAttention and xformers, without lossing end-to-end metrics across language, image, and video models.

   Cuda 2k 152

2. thu-ml/SpargeAttnthu-ml/SpargeAttnPublic

   SpargeAttention: A training-free sparse attention that can accelerate any model inference.

   Cuda 649 48

3. SPH_ProjectSPH_ProjectPublic

   SPH Realization of Fluid Simulation. Featuring Large Scale Simulation, Rigid-Fluid Coupling and High Viscosity Fluid.

   Python 178 13

4. mit-han-lab/llm-awqmit-han-lab/llm-awqPublic

   [MLSys 2024 Best Paper Award] AWQ: Activation-aware Weight Quantization for LLM Compression and Acceleration

   Python 3.1k 264

5. thu-nics/MoAthu-nics/MoAPublic

   [CoLM'25] The official implementation of the paper <MoA: Mixture of Sparse Attention for Automatic Large Language Model Compression>

   Python 139 7

6. mit-han-lab/omniservemit-han-lab/omniservePublic

   [MLSys'25] QServe: W4A8KV4 Quantization and System Co-design for Efficient LLM Serving; [MLSys'25] LServe: Efficient Long-sequence LLM Serving with Unified Sparse Attention

   C++ 717 48