Speech Command Recognition

Speak a keyword or upload a 1-second clip.

Commands: yes, no, up, down, left, right, on, off, stop, go

Model

CNN: 95.29%, instant // AST: 97.23%, slower on CPU


CNN: 1.21M params, 95.29% acc, instant // AST: 86.2M params, 97.23% acc, first load ~5s

GitHub