Pinned Loading

1. bytedance/FlexPrefill bytedance/FlexPrefill Public

   Code for paper: [ICLR2025 Oral] FlexPrefill: A Context-Aware Sparse Attention Mechanism for Efficient Long-Sequence Inference

   Python 61 2

2. native-sparse-attention-triton native-sparse-attention-triton Public

   Efficient triton implementation of Native Sparse Attention.

   Python 116 5