Hands-On Transfer Learning with Python

Summary

We looked at a completely new problem and case study in this chapter, involving audio identification and classification. Concepts surrounding audio data and signals were covered, including effective techniques to visualize and understand this data type.

We also looked at effective feature engineering techniques and how we could use transfer learning to extract effective features from image representations of audio data. This shows us the promise of transfer learning and how you can leverage knowledge from one domain (images) to another domain (audio) and build an extremely robust and effective classifier. Finally, we built a complete end-to-end pipeline for identifying and classifying new samples of audio data. Do check out further ...