ShawnHymel/tflite-speech-recognition
Demo for training a convolutional neural network to classify words and deploy the model to a Raspberry Pi using TensorFlow Lite.
This project helps embedded systems developers or hobbyists create a custom wake word detection system for small, low-power devices like a Raspberry Pi. It takes recorded speech audio as input and outputs a trained machine learning model that can recognize a specific spoken word, enabling voice-controlled applications. The primary users are engineers or makers building voice interfaces for their hardware projects.
105 stars. No commits in the last 6 months.
Use this if you need to train a simple speech recognition model to detect a specific "wake word" and deploy it onto a Raspberry Pi or similar embedded device.
Not ideal if you need to recognize a large vocabulary of words, require high accuracy in noisy environments, or are working with complex conversational AI applications.
Stars
105
Forks
54
Language
Jupyter Notebook
License
—
Category
Last pushed
Mar 02, 2020
Commits (30d)
0
Get this data via API
curl "https://pt-edge.onrender.com/api/v1/quality/voice-ai/ShawnHymel/tflite-speech-recognition"
Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.
Higher-rated alternatives
Picovoice/porcupine
On-device wake word detection powered by deep learning
MycroftAI/mycroft-precise
A lightweight, simple-to-use, RNN wake word listener
arcosoph/nanowakeword
A lightweight, open-source, and intelligent wake word detection engine. Train custom,...
mozilla/DeepSpeech
DeepSpeech is an open source embedded (offline, on-device) speech-to-text engine which can run...
OAID/cortex-m-kws
Cortex M KWS example with Tengine Lite.