How do I build a custom voice recognition model for multiple people?

TensorFlow Special Interest Groups TF.js
, ,
29

That tutorial shows a basic start. Doing longer sentences requires more complex models. Even if you train with longer sentences, the accuracy will probably start to drop a lot.

Following the @Bhack post above, XLSR-Wav2Vec2 model might help you.
I hope it’s published to TFHub at some point

30

tf.js is not sure when it will be available
wav2vec2 model

1 Like
32

Take a look also at facebook/wav2vec2-large-robust · Hugging Face

34
2 Likes
36

I can’t found the Euphonia and Parrotron Easy training tool or installation tool ?
Where are these tools ?

1 Like