TensorFlow in Swift

This repository is the Swift Chapter for 简单粗暴TensorFlow | A Concise Handbook of TensorFlow

MNIST Example in Book:

Book Chapter: Tensorflow Handbook - Swift

• Github Repository: https://github.com/huan/tensorflow-handbook-swift
• Github Pages: https://huan.github.io/tensorflow-handbook-swift/

USAGE

For those who had Swift installed

If you had installed Swift locally, then you can run swift directly:

$ swift run
Downloading train-images-idx3-ubyte ...
Downloading train-labels-idx1-ubyte ...
Reading data.
Constructing data tensors.
Test Accuracy: 0.9116667

For those who had no Swift installed (Docker)

If you are using Windows, or Ubuntu 19 (Swift for Tensorflow only supports Mac & Ubuntu 18.04 at Aug 2019), then the only way which you can play with this repository is to use Docker.

# use `nvidia-docker` if you want to use GPU
docker run -ti --rm \
  --privileged \
  --userns=host \
  -v "$(pwd)":/notebooks \
  zixia/swift \
  swift run

See:

1. Docker parameters - hashicorp/nomad#1904 (comment)
2. Docker Repo of zixia/swift - https://github.com/huan/docker-swift-tensorflow

TODO

RESOURCES

• Colab notebooks for Swift for TensorFlow
• Repository for reviews and evaluations of converted Swift Notebooks as part of GSoC '19 program
• fast.ai Embracing Swift for Deep Learning
• High Performance Numeric Programming with Swift: Explorations and Reflections
• Swift for TensorFlow Design Overview

Videos

• Swift for TensorFlow (TensorFlow @ O’Reilly AI Conference, San Francisco '18)
• Swift for TensorFlow: The Next-Generation Machine Learning Framework (TF Dev Summit '19)

Author

Huan (李卓桓), Google Developers Expert in Machine Learning (ML GDE), zixia@zixia.net

Copyright & License

• Code & Docs © 2018 - now Huan (李卓桓) <zixia@zixia.net>
• Code released under the Apache-2.0 License
• Docs released under Creative Commons