> ## Documentation Index
> Fetch the complete documentation index at: https://newtorch.aboneda.com/llms.txt
> Use this file to discover all available pages before exploring further.

# FlashAttention

> Memory-efficient attention implementation

# FlashAttention

IO-aware attention, fused kernels, massive memory savings.