Uploaded April 2021 | Updated September 2026, 2 weeks ago

---

Variational AutoEncoders are wonderful Deep Learning beasts to generate data. They have mostly been used to produce images. But they have the capacity to generate all kinds of data, speech included.

In my new tutorial, you can learn how to train a Variational AutoEncoder on the Free Sound Digits Dataset, with Python, TensorFlow and Keras.

Code:
github.com/musikalkemist/generating-sound-with-neural-networks/tree/main/13%20Training%20VAE%20with%20audio%20data

===============================

Interested in hiring me as a consultant/freelancer?
valeriovelardo.com/​​​​​​​

Join The Sound Of AI Slack community:
https://valeriovelardo.com/the-sound-...​

Follow Valerio on Facebook:
facebook.com/TheSoundOfAI​​

Connect with Valerio on Linkedin:
https://www.linkedin.com/in/valeriove...​

Follow Valerio on Twitter:
twitter.com/musikalkemist​​​​​​

===============================

Content:
0:00 Intro
1:28 Loading Free Sound Digits Dataset
6:12 Reshaping the data
9:43 Final edits
12:30 Training the VAE
13:02 Coming next