Explore the latest features and best practices for using TensorFlow Estimators in this 26-minute conference talk from EuroPython 2019. Dive into the simplified yet flexible API updates, including canned estimators for quick predictions and advanced tensor operations for custom neural networks. Learn how to efficiently handle large datasets using Datasets, leverage pre-trained modules from TensorFlow Hub to boost performance, and design neural networks following industry best practices. Gain insights into different levels of Estimator abstraction and discover how to optimize your machine learning pipeline. Suitable for those with basic knowledge of neural networks and TensorFlow conventions, this talk aims to demystify recent changes and empower you to create more effective machine learning models.
Overview
Syllabus
Mai Giménez - Tensorflow Estimators
Taught by
EuroPython Conference