Crate zuna_rs 

zuna-rs — ZUNA EEG Foundation Model inference in Rust

Pure-Rust inference for the ZUNA EEG foundation model. Two inference engines are available behind Cargo features:

feature	module	runtime
burn	crate root (ZunaEncoder, …)	Burn 0.20
rlx	rlx	RLX compiler/runtime

FIF preprocessing is shared (preprocess_fif_cpu) via exg.

Backends

Burn (with --features burn,ndarray): ndarray, blas-accelerate, wgpu, burn-mlx / mlx.

RLX (with --features rlx): cpu, metal, mlx, gpu, cuda, rocm, tpu, and BLAS variants.

Compare both engines (add --features burn,ndarray for the Burn side):

cargo run --example backend_compare --release \
    --no-default-features \
    --features burn,rlx,ndarray,rlx-cpu,rlx-metal,metal,wgpu,mlx,rlx-mlx

Re-exports

pub use rlx::EpochEmbedding;

pub use rlx::EpochOutput;

pub use rlx::InferenceResult;

pub use rlx::ZunaDecoder;

pub use rlx::ZunaEncoder;

pub use rlx::ZunaInference;

pub use config::DataConfig;

pub use config::InferConfig;

pub use config::ModelConfig;

pub use data::FifInfo;

pub use data::PreprocessedEpoch;

pub use data::PreprocessedFif;

pub use data::invert_reshape;

pub use data::preprocess_fif_cpu;

pub use channel_positions::MontageLayout;

pub use channel_positions::channel_xyz;

pub use channel_positions::montage_channels;

pub use channel_positions::nearest_channel;

pub use channel_positions::normalise;

pub use csv_loader::CsvInfo;

pub use csv_loader::CsvLoadOptions;

pub use csv_loader::PaddingStrategy;

pub use csv_loader::load_from_csv;

pub use csv_loader::load_from_named_tensor;

pub use csv_loader::load_from_raw_tensor;

pub use csv_export::fif_to_csv;

Modules

channel_positions

EEG channel position lookup from embedded standard montage files.

config

csv_export

Export FIF recordings to CSV files.

csv_loader

CSV and raw-tensor loading for ZUNA inference.

data

Data preparation for ZUNA inference.

rlx

RLX-backed ZUNA inference (rlx::Graph + rlx::Session).

Functions

init_threads

Configure the global Rayon thread pool (Burn NdArray + RLX CPU).