Deterministic Tensorflow Part 1: Model Training

Reproducibility is critical to any scientific endeavour, and machine learning is no exception. Releasing code that generates results from papers is an important step in addressing this, but difficu...