boostr 0.1.0

ML framework built on numr - attention, quantization, model architectures
Documentation 
1
2
3
4
5
6
7
8
9
10
11

pub mod flash;
pub mod fused_qkv;
pub mod mla;
pub mod paged_attention;
pub mod varlen_attention;

pub use flash::{AttentionOps, FlashAttentionOps};
pub use fused_qkv::FusedQkvOps;
pub use mla::MlaOps;
pub use paged_attention::PagedAttentionOps;
pub use varlen_attention::VarLenAttentionOps;