Overview
In this section you will find tutorials that can be used to get started with TensorFlow for R or, for more advanced users, to discover best practices for loading data, building complex models and solving common problems.
The best place to get started with TensorFlow is using Keras - a Deep Learning API created by François Chollet and ported to R by JJ Allaire. Keras makes it easy to get started, and it allows you to progressively build more complex workflows as you need to use advanced models and techniques.
For beginners
We recommend the following tutorials for your first contact with TensorFlow. Feel free to navigate through the ‘beginners’ section in the sidebar.
- Quickstart: the minimal getting started guide to Keras.
- Basic ML with Keras: use Keras to solve basic Machine Learning tasks.
-
Load data: learn to efficiently load data to TensorFlow using
tfdatasets.
For experts
- Advanced Quickstart: learn the subclassing API and how to create custom loops.
- Customization: build custom layers and training loops in TensorFlow.
- Distributed Training: distribute your model training across multiple GPU’s or machines.
We also provide tutorials focused on different types of data:
- Images: Build more advanced models for classification and segmentation of images.
- Structured Data: Build models for structured data.