Why TensorFlow eager execution matters - Sigrid Keydana

In current deep learning with Keras and TensorFlow, when you've mastered the basics and are ready to dive into more involved applications (such as generative networks, sequence-to-sequence or attention mechanisms), you may find that surprisingly, the learning curve doesn't get much flatter. This is largely due to restrictions imposed by TensorFlow's traditional static graph paradigm. With TensorFlow Eager Execution, available since summer and announced to be the default mode in the upcoming major release, model architectures become more flexible, readable, composable, and last not least, debuggable. In this session, we'll see how with Eager, we can code sophisticated architectures like Generative Adversarial Networks (GANs) and Variational Autoencoders (VAEs) in a straightforward way.

About the Author

Sigrid Keydana

Sigrid has experience a psychologist, software developer and data scientist. She is passionate about exploring the frontiers of deep learning and helping users employ the power of deep learning from R. When she’s not tuning neural networks she loves doing outdoor sports- among her favorites are cycling, running and cross country skiing.

Follow on Twitter Follow on Linkedin Visit Website More Content by Sigrid Keydana
Previous Video
Working with names and expressions in your tidy eval code - Lionel Henry
Working with names and expressions in your tidy eval code - Lionel Henry

In practice there are two main flavors of tidy eval functions: functions that select columns, such as `dply...

Next Video
Visualizing uncertainty with hypothetical outcomes plots - Claus Wilke
Visualizing uncertainty with hypothetical outcomes plots - Claus Wilke

Uncertainty is a key component of statistical inference. However, uncertainty is not easy to convey effecti...