Explore Keras Preprocessing Layers in this 37-minute video featuring Google Software Engineer Matthew Watson. Learn how these layers streamline model development workflows as Watson builds an end-to-end model. Discover the problem-solving capabilities of preprocessing layers, understand their various types, and see how to transform string inputs into numeric data. Follow along as a simple model is constructed, new features are added, and performance is optimized using tf.data. Gain insights into multi-worker training and key takeaways for efficient model development. Access additional resources, including Watson's GitHub, preprocessing layers guide, and tutorials on text and image loading to further enhance your TensorFlow skills.
Overview
Syllabus
- Introduction
- Identifying the problem
- What are Keras preprocessing layers
- Preprocessing layers that are offered
- Transforming inputs from strings to a numeric input
- Building a simple model
- Adding a new feature
- Better performance with tf.data
- Multi worker training
- Takeaways
Taught by
TensorFlow
