Update the readme file. (#713)

Summary:
Pull Request resolved: #713

Simpler build instructions due to the new helper scripts.

Reviewed By: kimishpatel

Differential Revision: D50071422

fbshipit-source-id: 550aacdb1a20b7513461ba31368ebaaf2111cdd2

Author: shoumikhin

examples/demo-apps/apple_ios/README.md

@@ -1,176 +1,96 @@
-# iOS Demo App: ExecuTorch Setup
+# ExecuTorch: iOS Demo App Setup
 
-This guide explains how to setup ExecuTorch for iOS using a demo app. The app
-employs a MobileNet v3 model (exported to ExecuTorch) to process live camera
-images.
+This README guides you through the setup of ExecuTorch on iOS using a demo app.
+This app utilizes the MobileNet v3 model to process live camera images.
 
-## Pre-setup
+## Prerequisites
 
-1. Install Xcode and Command Line Tools:
+1. **Install Xcode 15 and Command Line Tools**
 
    ```bash
    xcode-select --install
    ```
 
-2. Install `buck2` binary (using MacOS Apple Silicon build as example):
+2. **Python 3.10+ and `pip`** (Pre-installed from MacOS 13.5+)
 
-   ```bash
-   curl -L -O https://github.com/facebook/buck2/releases/download/2023-07-18/buck2-aarch64-apple-darwin.zst
-   pip3 install zstd
-   zstd -cdq buck2-aarch64-apple-darwin.zst > /tmp/buck2 && chmod +x /tmp/buck2
-   ```
-
-3. Install [Cmake](cmake.org/download) and link it in a system directory or
-   `$PATH`:
+   [Download](https://www.python.org/downloads/macos/) and install Python 3.10
+   or 3.11, if needed, and verify the versions:
 
    ```bash
-   ln -s /Applications/CMake.app/Contents/bin/cmake /usr/bin/cmake
+   which python3 pip
+   python3 --version
+   pip --version
    ```
 
-4. Clone ExecuTorch repository and update submodules:
+3. **Follow the [Getting Started](../../../docs/source/getting-started-setup.md)
+   Tutorial**
 
-   ```bash
-   git clone https://github.com/pytorch/executorch.git
-   cd executorch
-   git submodule sync
-   git submodule update --init
-   ```
+4. **Backend Dependency Installation**
 
-5. Verify Python 3.10+ (standard since MacOS 13.5) and `pip3` installation:
+   Install additional dependencies for **CoreML**:
 
    ```bash
-   which python3 pip
-   python3 --version
+   ./backends/apple/coreml/scripts/install_requirements.sh
    ```
 
-6. Install PyTorch dependencies:
+   And **Metal Performance Shaders**:
 
    ```bash
-   ./install_requirements.sh
+   ./backends/apple/mps/install_requirements.sh
    ```
 
-## Flatbuffers Compiler Setup
-
-Run the following in the `flatbuffers` directory:
-
-```bash
-cd third-party/flatbuffers
-rm -rf cmake-out && mkdir cmake-out && cd cmake-out
-cmake .. && cmake --build . --target flatc
-cd ../../..
-```
-
-## ExecuTorch Configuration
-
-Configure the libraries for iOS:
-
-```bash
-rm -rf cmake-out && mkdir cmake-out && cd cmake-out
-cmake .. -G Xcode \
-    -DCMAKE_TOOLCHAIN_FILE=../third-party/pytorch/cmake/iOS.cmake \
-    -DBUCK2=/tmp/buck2 \
-    -DPYTHON_EXECUTABLE=$(which python3) \
-    -DFLATC_EXECUTABLE=$(realpath ../third-party/flatbuffers/cmake-out/flatc) \
-    -DEXECUTORCH_BUILD_EXTENSION_DATA_LOADER=ON \
-    -DCMAKE_ARCHIVE_OUTPUT_DIRECTORY=$(pwd)
-```
-
-Append `-DIOS_PLATFORM=SIMULATOR` for Simulator configuration to build libraries
-for `x86` architecture instead of `arm64`, which is the default.
-
-Append `-DEXECUTORCH_BUILD_COREML_DELGATE=ON` to build CoreML backend libraries.
-
-Append `-DEXECUTORCH_BUILD_XNNPACK=ON` to build XNNPACK backend libraries.
-
-## Building and Copying Libraries
-
-1. Build the libraries:
-
-```bash
-cmake --build . --config Release
-```
-
-2. Copy the libraries to the appropriate location to link against them:
+## Model Export & Bundling
 
-Navigate to the build artifacts directory:
+1. **Export MobileNet v3 model with CoreML, MPS and XNNPACK delegates**
 
-```bash
-cd Release
-```
-
-Copy the core libraries:
-
-```bash
-mkdir -p ../../examples/demo-apps/apple_ios/ExecuTorchDemo/ExecuTorchDemo/Frameworks/executorch/
-cp libexecutorch.a \
-   libextension_data_loader.a \
-   ../../examples/demo-apps/apple_ios/ExecuTorchDemo/ExecuTorchDemo/Frameworks/executorch/
-```
-
-For Portable CPU operators, copy additional libraries:
-
-```bash
-mkdir -p ../../examples/demo-apps/apple_ios/ExecuTorchDemo/ExecuTorchDemo/Frameworks/portable/
-cp libportable_kernels.a \
-   libportable_ops_lib.a \
-   ../../examples/demo-apps/apple_ios/ExecuTorchDemo/ExecuTorchDemo/Frameworks/portable/
-```
-
-For CoreML delegate backend, copy additional libraries:
-
-```bash
-mkdir -p ../../examples/demo-apps/apple_ios/ExecuTorchDemo/ExecuTorchDemo/Frameworks/coreml/
-cp libcoremldelegate.a \
-   ../../examples/demo-apps/apple_ios/ExecuTorchDemo/ExecuTorchDemo/Frameworks/coreml/
-```
-
-For XNNPACK delegate backend, copy additional libraries:
+   ```bash
+   python3 -m examples.portable.scripts.export --model_name="mv3"
+   python3 -m examples.xnnpack.aot_compiler --delegate --model_name="mv3"
+   python3 -m examples.apple.coreml.scripts.export_and_delegate --model_name="mv3"
+   python3 -m examples.apple.mps.scripts.mps_example --model_name="mv3"
 
-```bash
-mkdir -p ../../examples/demo-apps/apple_ios/ExecuTorchDemo/ExecuTorchDemo/Frameworks/xnnpack/
-cp libclog.a \
-   libcpuinfo.a \
-   libpthreadpool.a \
-   libxnnpack_backend.a \
-   libXNNPACK.a \
-   ../../examples/demo-apps/apple_ios/ExecuTorchDemo/ExecuTorchDemo/Frameworks/xnnpack/
-```
+   mkdir -p examples/demo-apps/apple_ios/ExecuTorchDemo/ExecuTorchDemo/Resources/Models/MobileNet/
+   mv mv3*.pte examples/demo-apps/apple_ios/ExecuTorchDemo/ExecuTorchDemo/Resources/Models/MobileNet/
+   ```
 
-Then return to the `executorch` directory:
+2. **Download MobileNet model labels**
 
-```bash
-cd ../..
-```
+   ```bash
+   curl https://raw.githubusercontent.com/pytorch/hub/master/imagenet_classes.txt \
+     -o examples/demo-apps/apple_ios/ExecuTorchDemo/ExecuTorchDemo/Resources/Models/MobileNet/imagenet_classes.txt
+   ```
 
-## Model Download and Bundling
+## ExecuTorch & Backend Building
 
-1. Download MobileNet model labels and bundle them with the app:
+1. **Build frameworks**
 
-```bash
-mkdir -p examples/demo-apps/apple_ios/ExecuTorchDemo/ExecuTorchDemo/Resources/Models/MobileNet/
-curl https://raw.githubusercontent.com/pytorch/hub/master/imagenet_classes.txt -o examples/demo-apps/apple_ios/ExecuTorchDemo/ExecuTorchDemo/Resources/Models/MobileNet/imagenet_classes.txt
-```
+   ```bash
+   ./build/build_apple_frameworks.sh --Release --coreml --mps --xnnpack
+   ```
 
-2. Export a MobileNet v3 model backed with XNNPACK delegate and bundle it with
-   the app:
+2. **Move frameworks for app linking**
 
-```bash
-export FLATC_EXECUTABLE=$(realpath third-party/flatbuffers/cmake-out/flatc)
-python3 -m examples.portable.scripts.export --model_name="mv3"
-python3 -m examples.xnnpack.aot_compiler --model_name="mv3" --delegate
-python3 -m examples.export.coreml_export_and_delegate -m "mv3"
-cp mv3.pte mv3_coreml.pte mv3_xnnpack_fp32.pte examples/demo-apps/apple_ios/ExecuTorchDemo/ExecuTorchDemo/Resources/Models/MobileNet/
-```
+   ```bash
+   mv cmake-out examples/demo-apps/apple_ios/ExecuTorchDemo/ExecuTorchDemo/Frameworks
+   ```
 
 ## Final Steps
 
-1. Open the project with Xcode:
+1. **Open project in Xcode**
 
-```bash
-open executorch/examples/demo-apps/apple_ios/ExecuTorchDemo/ExecuTorchDemo.xcodeproj
-```
+   ```bash
+   open examples/demo-apps/apple_ios/ExecuTorchDemo/ExecuTorchDemo.xcodeproj
+   ```
+
+2. **Run tests in Xcode** (Cmd + U) or command line:
 
-2. Set the Header Search Paths for `MobileNetClassifier` target to the directory
-   containing the `executorch` folder.
+   ```bash
+   xcrun simctl create executorch "iPhone 15"
+   xcodebuild clean test \
+        -project examples/demo-apps/apple_ios/ExecuTorchDemo/ExecuTorchDemo.xcodeproj \
+        -scheme App \
+        -destination name=executorch
+   xcrun simctl delete executorch
+   ```
 
-3. Run the app (Cmd + R) and tests (Cmd + U).
+3. **Setup Code Signing and run app** (Cmd + R).