R and Tensorflow!

Why Use R with Tensorflow?

Repo Contents

If you're completely new to using R with Tensorflow, the hello-keras.Rmd notebook will give you an introduction.

Reproducibility.Rmd shows how to share and reuse models.

Visualization.Rmd is a playground for leveraging R's data visualization methods for introspecting experiment results.

Examples - You can find 2 examples in the R folder:

• mnist_mlp: Trains a simple deep NN on the MNIST dataset.

• cats_and_dogs_small: Convnet for dogs-versus-cats classification.

How to Use

Download large datasets to the downloads/ folder and git will ignore them.

Models are written to the models/ folder.

Additional Resources

• Installing RStudio's Tensorflow packages - https://keras.rstudio.com/
• R Interface to CloudML (training in the cloud) - https://tensorflow.rstudio.com/tools/cloudml/articles/getting_started.html
• Run RStudio in the Cloud with IBM Watson Studio - https://ibm.biz/BdYRNi
• IBM's Model Asset Exchange - https://developer.ibm.com/code/exchanges/models/
• Example Shiny App using Keras and LIME! https://jjallaire.shinyapps.io/keras-customer-churn/