5 projects
flash-attn-4
Flash Attention CUTE (CUDA Template Engine) implementation
attn-gym
Helpful tools and examples for working with flex-attention
dist-moe
CuTe DSL distributed mixture-of-experts kernels for PyTorch
transformer-nuggets
A place to store reusable transformer components found around the interwebs
flash-attn-cute
Flash Attention CUTE - package coming soon