search

anjandeepsahni / automatic_speech_recognition

Speech to text transcription using RNN (Listen, Attend and Spell).
MIT License
11 stars 6 forks source link

How can we do this own custom dataset? #1

Open razasaddiqi opened 4 years ago

razasaddiqi commented 4 years ago
  1. can you please tell what should be the structure of dataset to feed into this model and what feature you extracted from audio from all out there such as mel spectrogram, filterbank, mfcc etc.
  2. how will i map each recording to it's relevant trancript characters.
Ishu-dev commented 8 months ago

Please add some reply as to what should be the structure of data