Tensorflow 101: Building end-to-end machine learning pipelines using Estimators

By Dora Jambor and Peng Yu

Supporting material used in our workshop on Tensorflow Tstimators. The workshop focused on the ease and simplicity of building end-to-end machine learning pipelines for production environments using Tensorflow's high-level estimators API.

In the notebook worksheet, we walk through this process by building an image classifier on the Fashion MNIST Dataset. This dataset contains 60K training and 10K test images of clothing articles, each classified into 10 labels. Each image has been greyscaled and resized to 28x28 pixels. We use the Estimators API to train, evaluate, predict and export an image classifier model.

Slides on the workshop will be soon available, covering more details on Tensorflow basics and Estimators.

Handcrafted in New York City, design & code by
Dora Jambor