Earlybird: predictive speech-to-text in 29M parameters | Sahil Mahendrakar

Earlybird: predictive speech-to-text in 29M parameters | Sahil Mahendrakar
www.sahilmahendrakar.com/writings/earlybird
Earlybird: predictive speech-to-text in 29M parameters
A tiny streaming speech-to-text model that predicts words before the speaker finishes saying them with a confidence score.