Tensorflow training eats up cpu without progressing

context: Tensorflow training not progressing · Issue #58333 · tensorflow/tensorflow · GitHub

Basically I would go down the second notebook from nicknochnack’s repo with no issues until it came to training the model, from which there would be no fatal errors and no progress. Changing apis was mentioned in the discussion, but it is not clear how I go about doing that. Do I have to follow the given “instructions” in the log, and go into each file in my project to manually replace text?