turboquant-rs 0.4.1

TurboQuant KV-Cache Quantization — 3-bit compression with zero accuracy loss (Zandieh et al., ICLR 2026)
Documentation 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21

//! CUDA kernel wrappers for TurboQuant operations.
//!
//! All modules are gated behind `#[cfg(feature = "cuda")]`.

#[cfg(feature = "cuda")]
pub mod attention;
#[cfg(feature = "cuda")]
pub(crate) mod ffi;
#[cfg(feature = "cuda")]
pub(crate) mod quantize;

#[cfg(feature = "cuda")]
mod error_check;

/// Post-launch CUDA error check. `#[doc(hidden)] pub` so the
/// `tests/cuda_error_check_tests.rs` integration test can call it from
/// the test crate; same SemVer-visible convention as
/// [`crate::test_utils`].
#[cfg(feature = "cuda")]
#[doc(hidden)]
pub use error_check::check_cuda_kernel_launch;