#ONNX
4 related articles
Offline Voice Control with ONNX: Custom Wake Words and Multi-Command Recognition
Learn how to build offline voice control with custom wake words and ONNX: a two-stage architecture separating wake and command recognition, a multi-keyword model for fixed commands, and R0/R1 two-round training to govern false triggers, plus ONNX and TFLite deployment checklists.
Fixing GatherBlockQuantized: Running 4-bit Gemma in Browsers Without WebGPU
Transformers.js 4.x throws a GatherBlockQuantized error when loading 4-bit Gemma ONNX models without WebGPU, because the default asyncify build lacks that operator's CPU kernel. Here's the fix: manually set wasmPaths to point at the regular WebAssembly build.
Building a Small Multilingual Offline Voice-Control Pipeline with ONNX
Learn how to build a small multilingual offline voice-control pipeline with ONNX, covering KWS vs ASR, runtime pipeline design, multi-platform deployment (Python/Web/Android/ESP32-S3), and pre-launch testing.
TutorialsONNX Model Deployment in Practice: A Complete Guide from PyTorch Export to Cross-Framework Inference
Complete guide to ONNX neural network exchange format: PyTorch and TensorFlow model export, ONNX Runtime inference, MNIST classifier deployment, and HuggingFace model downloads with runnable code examples.