Pinned Loading

1. thu-ml/SageAttention thu-ml/SageAttention Public

   [ICLR2025, ICML2025, NeurIPS2025 Spotlight] Quantized Attention achieves speedup of 2-5x compared to FlashAttention, without lossing end-to-end metrics across language, image, and video models.

   Cuda 2.6k 243

2. thu-ml/SpargeAttn thu-ml/SpargeAttn Public

   [ICML2025] SpargeAttention: A training-free sparse attention that accelerates any model inference.

   Cuda 747 62

3. CardinalityEstimationTestbed CardinalityEstimationTestbed Public

   CardinalityEstimationTestbed

   Python 48 14

4. Sparse_SageAttention_API Sparse_SageAttention_API Public

   Python 54 6

5. attention-survey/Efficient_Attention_Survey attention-survey/Efficient_Attention_Survey Public

   A Survey of Efficient Attention Methods: Hardware-efficient, Sparse, Compact, and Linear Attention

   204 4