reve-rs

Pure-Rust inference for the REVE (Representation for EEG with Versatile Embeddings) foundation model, built on Burn 0.20.

REVE is pretrained on 60,000+ hours of EEG data from 92 datasets spanning 25,000 subjects. Its key innovation is a 4D Fourier positional encoding scheme that enables generalization across arbitrary electrode configurations without retraining.

Architecture

EEG [B, C, T]
    │
    ├─ Overlapping Patch Extraction (unfold)
    │  → [B, C, n_patches, patch_size]
    │
    ├─ Linear Patch Embedding
    │  → [B, C*n_patches, embed_dim]
    │
    ├─ 4D Positional Encoding (Fourier + MLP)
    │  (x, y, z, t) → [B, C*n_patches, embed_dim]
    │
    ├─ Transformer Encoder (RMSNorm, GEGLU, Multi-Head Attention)
    │  → [B, C*n_patches, embed_dim]
    │
    └─ Classification Head (Flatten+LN+Linear or Attention Pooling)
       → [B, n_outputs]

Quick Start

use reve_rs::{ReveEncoder, data};
use std::path::Path;

// Load model
let (model, ms) = ReveEncoder::<B>::load(
    Path::new("config.json"),
    Path::new("model.safetensors"),
    device,
)?;

// Build input batch
let batch = data::build_batch::<B>(
    signal,     // [C, T] flat f32
    positions,  // [C, 3] flat f32
    n_channels,
    n_samples,
    &device,
);

// Run inference
let output = model.run_batch(&batch)?;
println!("Output: {:?}", output.shape);

Build

# CPU (default — NdArray + Rayon)
cargo build --release

# CPU with Apple Accelerate BLAS (macOS)
cargo build --release --features blas-accelerate

# GPU (Metal on macOS)
cargo build --release --no-default-features --features metal

# GPU (Vulkan on Linux/Windows)
cargo build --release --no-default-features --features vulkan

CLI Inference

# Download weights (requires HuggingFace access)
cargo run --release --features hf-download --bin download_weights -- --repo brain-bzh/reve-base

# Run inference
cargo run --release --bin infer -- --weights data/model.safetensors --config data/config.json -v

Pretrained Weights

Weights are on HuggingFace:

Note: You must agree to the data usage terms on HuggingFace before downloading.

Features

References

• El Ouahidi et al. (2025). REVE: A Foundation Model for EEG — Adapting to Any Setup with Large-Scale Pretraining on 25,000 Subjects. NeurIPS 2025.
• braindecode Python implementation

License

Apache-2.0