Name		Name	Last commit message	Last commit date
parent directory ..
mnist_cnn_custom_estimator		mnist_cnn_custom_estimator
01_README_mnist_simple.md		01_README_mnist_simple.md
README.md		README.md
dataset.py		dataset.py
mnist_eager_keras-debug.ipynb		mnist_eager_keras-debug.ipynb
mnist_eager_keras.ipynb		mnist_eager_keras.ipynb
mnist_estimator.ipynb		mnist_estimator.ipynb
mnist_estimator.py		mnist_estimator.py
mnist_simple.ipynb		mnist_simple.ipynb
mnist_simple.py		mnist_simple.py

README.md

The high-level TensorFlow APIs, via MNIST & 'fashion MNIST'

This directory contains a series of labs/examples that introduce the the high-level TensorFlow APIs in tf.estimator and tf.data via the 'MNIST' (and 'Fashion MNIST') data sets.

mnist_estimator.ipynb: Introducing the high-level TensorFlow APIs in tf.estimator and tf.data to easily build a LinearClassifier, as well a DNNClassifier with hidden layers. Introducing TensorBoard.

Run this example as a colab notebook.
mnist_eager_keras-debug.ipynb and mnist_eager_keras.ipynb: This pair of notebooks show an example of using Keras with TensorFlow eager mode, on the 'Fashion MNIST' dataset. These notebooks require TensorFlow >=1.7.

Run these examples as colab notebooks: mnist_eager_keras-debug.ipynb or mnist_eager_keras.ipynb. (We'll start with the "debug" version in the workshop).
Building Custom CNN Estimators: Where 'canned' Estimators aren't available for your use case, you can build a custom one, to get all the advantages of using an Estimator, including support for distributed training. You can use Keras layers to do this. Examples show how to do this for variants of CNNs, with both Keras and TensorFlow layers.
Click through to the README for links to run these examples as colab notebooks.

01_README_mnist_simple: A simple version of a linear classifier for MNIST that uses the low-level TensorFlow APIs.

Run this example as a colab notebook.