This is an example application for TensorFlow Lite on iOS. It uses Classification-by-Retrieval to train classification models out of a small sample of user-provided images.

These instructions walk you through building and running the demo on an iOS device.

• Device with iOS 14.0 or above

• Xcode 12.5 or above

• Valid Apple Developer ID

• Xcode command-line tools (run xcode-select --install)

If this is a new install, you will need to run the Xcode application once to agree to the license before continuing.

Note: The demo app requires a camera and must be executed on a real iOS device. You can build it and run with the iPhone Simulator, but the app will raise a Camera not found exception.

1. Clone this GitHub repository to your workstation.

   $ git clone https://github.com/tensorflow/examples.git
   

2. Change directory to the Classification-by-Retrieval example

   $ cd examples/lite/examples/classification_by_retrieval
   

3. Build for a simulator with the following command:

   $ bazel build -c opt --config=ios_x86_64 ios:ImageClassifierBuilder
   

4. To build for a device:

   1. You'll first need to obtain a mobile provisioning profile from Apple (documentation).

   2. Make a symlink to your profile like so:

      $ ln -s <path/to/your/profile.mobileprovision> ProvisioningProfile.mobileprovision
      

   3. Uncomment all occurrences of ProvisioningProfile.mobileprovision in ios/BUILD.

   4. Finally, build for a device with the following command:

      $ bazel build -c opt --config=ios_arm64 ios:ImageClassifierBuilder
      

You'll have to grant permissions for the app to use the device's camera and Photo Library.

First you'll create a model out of albums of your Photo Library, then you'll be able to see how it performs live: point the camera at things you trained your model with and enjoy seeing how it classifies them!