Swift for TensorFlow in 3 Hours

Quick Start

You won't need to go beyond this quickstart section if everything here works.

To copy the tutorial notebooks into your Google Drive:

1. Open the shared Google Drive folder, containing the notebooks:

   • https://lab.to/s4tf_tfworld_notebooks

2. Add the shared TFWorld2019 folder to your Google Drive:

   • Click the dropdown next to the folders name, and then clicking "Add to My Drive", as shown in the image below.

3. Once the TFWorld2019 folder is in your Google Drive:

   • Open whichever notebook(s) you want to work with by double clicking them, and then clicking "Google Colaboratory", as shown in the image below.

   
   
   

   • If you see the notebook as raw text, click the "Open with" dropdown at the top, and click "Google Colaboratory", or if you don't see Colab as an option, click "Connect more apps":

   
   
   

   • Then add "Google Colaboratory":

   
   
   

   • Then open it with Colab:

   
   
   

4. Once your notebook is open, run it in Playground mode to allow cells to compile and run, as shown in the image below.

   
   
   

To create a blank Swift notebook in Google Colab

Click on this link: lab.to/s4tf_blank

You'll need to save and name the notebook once you've created it!

Other useful resources

Why Swift for TensorFlow?

• https://github.com/tensorflow/swift/blob/master/docs/WhySwiftForTensorFlow.md

Swift tour and book:

• Tour: https://docs.swift.org/swift-book/GuidedTour/GuidedTour.html
• Book: https://docs.swift.org/swift-book/

Tutorials from the Swift for TensorFlow team:

• Basic Swift: https://www.tensorflow.org/swift/tutorials/a_swift_tour
• Protocol-oriented Programming and Generics: https://www.tensorflow.org/swift/tutorials/protocol_oriented_generics
• Python interopability: https://www.tensorflow.org/swift/tutorials/python_interoperability
• Model training: https://www.tensorflow.org/swift/tutorials/model_training_walkthrough
• Custom differentiation: https://www.tensorflow.org/swift/tutorials/custom_differentiation
• Raw TensorFlow operators: https://www.tensorflow.org/swift/tutorials/raw_tensorflow_operators

The Swift Models repository, from the TensorFlow team:

• https://github.com/tensorflow/swift-models

The Swift for TensorFlow Discussion Group:

• https://groups.google.com/a/tensorflow.org/forum/#!forum/swift

The Swift APIs for TensorFlow:

• https://github.com/tensorflow/swift-apis

The Differentiable Programming Proposal:

• https://docs.google.com/document/d/1bPepWLfRQa6CtXqKA8CDQ87uZHixNav-TFjLSisuKag/edit
• https://forums.swift.org/t/differentiable-programming-mega-proposal/28547
• https://github.com/dan-zheng/swift/blob/differentiable-programming/docs/DifferentiableProgramming.md

SwiftAI's (FastAI) high-level Swift ML API (very early):

• https://github.com/fastai/swiftai

Why FastAI is embracing Swift for Deep Learning:

• https://www.fast.ai/2019/03/06/fastai-swift/

FastAI's lessons on Swift for Deep Learning:

• Basics of Swift for Deep Learning: https://course.fast.ai/videos/?lesson=13
  • Slides from this lesson: https://docs.google.com/presentation/d/1dc6o2o-uYGnJeCeyvgsgyk05dBMneArxdICW5vF75oU/edit#slide=id.p
• C interop; Protocols; Putting it all together: https://course.fast.ai/videos/?lesson=14
  • A participant's notes on this lesson: https://medium.com/@lankinen/fast-ai-lesson-14-notes-part-2-v3-be4667394295

The FastAI forum:

• https://forums.fast.ai/c/swiftai

Skip the Foreign Function Interface (FFI), a talk from the LLVM/Apple team:

• Slides: http://llvm.org/devmtg/2014-10/Slides/Skip%20the%20FFI.pdf
• Video: http://llvm.org/devmtg/2014-10/Videos/Skip%20The%20FFI!%20Embedding%20Clang%20for%20C-720.mov

Value SEMANTICS (not value types!) Talk:

• https://academy.realm.io/posts/swift-gallagher-value-semantics/

Facebook AI's paper on Tensor Comprehensions:

• https://arxiv.org/pdf/1802.04730.pdf