Is there a TensorFlow version of the fast.ai library? Or is it possible to implement these various functions in TensorFlow in a convenient manner? I work with TensorFlow at my job and want to implement many of the techniques taught. Any advice how I can do that?
There is a repository which has some of the techniques implemented. It is for Keras and not a direct implementation for Tensorflow. Nevertheless, it could serve as a starting point.
https://github.com/nathanhubens/Learning-Rate
The implementation heavily depends on what you are planning to use in Tensorflow. Starting to implement it with Eager is kind of obvious as the debugging is a lot easier. I wouldn’t try to directly implement it for the higher level APIs, such as Estimators. At least in the previous versions, it was not too smooth for me to customize them.
Please don’t cross post your questions to multiple categories.