Explore the latest features of TensorFlow 2.0 in this 43-minute EuroPython Conference talk. Delve into the framework's evolution, focusing on Keras as the new standard API, Eager Mode & Graph execution, production pipeline improvements, and the extended ecosystem. Learn how TensorFlow 2.0 addresses previous usability challenges while maintaining its strengths in performance and scalability. Gain insights into the modernized, Pythonic API that rivals competitors like PyTorch, and discover how these changes benefit both researchers and practitioners in the field of deep learning and data science. Examine practical examples of model specification, data handling, training processes, and exporting models using the new TensorFlow 2.0 paradigm.
Overview
Syllabus
Introduction
TensorFlow 1110
What is TensorFlow 20
Function not session
Jeff Graph
Karis
Karis Layers
Karis Loss
Model API
Sequential Model
Functional API
Channel API
Schoener
Fetch Data
Data Source
Python in Trouble
Compile
Training
Model arguments
Gradient Tape
Custom Training
Exporting Models
Conclusion
Question
Taught by
EuroPython Conference
