Pinned Loading

1. Image-Local-Attention Image-Local-Attention Public

   A better PyTorch implementation of image local attention which reduces the GPU memory by an order of magnitude.

   Python 141 19

2. GBDTMO GBDTMO Public

   Gradient boosted decision trees for multiple outputs. Better generalization ability, faster training and inference.

   C++ 46 10

3. FlashWindowAttention FlashWindowAttention Public

   Speedup the attention computation of Swin Transformer

   Python 21 3

4. MEDUSA-Plus MEDUSA-Plus Public

   MEDUSA+: Acceleration Multiple Heads Decoding for LLM via Dynamic Tree Attention

   Jupyter Notebook