Skip to content

Latest commit

 

History

History
22 lines (18 loc) · 997 Bytes

README.md

File metadata and controls

22 lines (18 loc) · 997 Bytes

Doorbell.io

Audio Classifier with Tensorflow.js and ReactJS This is just a specific use case, you can train your own model and classify any sound you like!

How To Use

  • Clone this repository
  • Run npm install
  • Start using npm run

How to Train Your Own Model

  • Go to https://teachablemachine.withgoogle.com/train
  • Create a new Audio Project
  • Record a couple of samples for Background Noise (each of 20 seconds). Extract the samples for each recording.
  • Move on to the next class (default name will be "Class 2"). Similarly record samples for the sound you want to classifiy.
  • You can add more classes similarly and rename them. For better accuracy, try to record more samples for each class.
  • Train the model, for at least 100 epochs.
  • Export the model and upload the model to the cloud. Get the shareable link.
  • Replace the current model link in AudioClassifier.js with your shareable link.
  • (Optional) Change the variable tts_key for TTS.
  • That's all.