MESIE 16-Channel Acoustic Beamforming Transformer v1 (ALPHA WebAssembly Edition)
Zero-Latency Client-Side Inference for Browser (React/Next.js) & Mobile (React Native/iOS/Android) Published by ItsNotAI LABS (Dallas, Texas)
The MESIE-MultiAudio-v1-alpha repository contains the lightweight ONNX & WebAssembly (WASM) client-side package for running inference directly inside web browsers and mobile apps with 0ms server latency and zero cloud API infrastructure costs.
⚡ WebAssembly (WASM) SDK Features
- Direct In-Browser Inference: Compiles PyTorch weights into ONNX Runtime WASM (
model.onnx). - Zero Server Latency: Executes directly on client CPU threads using SIMD & multi-threading.
- Full TypeScript SDK Included: Includes
@sovereign-engine/wasm-sdksource and pre-built bundles. - 100% Offline Capable: Zero network dependencies after initial WASM module download.
🚀 Quick Start (React / Next.js)
// Loaded in Browser via WebAssembly (WASM) & ONNX Runtime
🔬 Model Parameter Integrity
- Trainable Parameters:
797,584 - Positional Encoding Buffer:
640,000 - ONNX Model Size:
5.5 MB - WASM Inference Latency:
1.384 ms
📄 License
Apache 2.0 License © ItsNotAI LABS