Merge pull request pytorch#736 from cccclai/site

Update mobile homepage for 1.9

Author: brianjo

_mobile/android.md

@@ -34,9 +34,10 @@ model.eval()
 example = torch.rand(1, 3, 224, 224)
 traced_script_module = torch.jit.trace(model, example)
 traced_script_module_optimized = optimize_for_mobile(traced_script_module)
-traced_script_module_optimized.save("app/src/main/assets/model.pt")
+traced_script_module_optimized._save_for_lite_interpreter("app/src/main/assets/model.ptl")
+
 ```
-If everything works well, we should have our model - `model.pt` generated in the assets folder of android application.
+If everything works well, we should have our model - `model.ptl` generated in the assets folder of android application.
 That will be packaged inside android application as `asset` and can be used on the device.
 
 More details about TorchScript you can find in [tutorials on pytorch.org](https://pytorch.org/docs/stable/jit.html)
@@ -64,8 +65,8 @@ repositories {
 }
 
 dependencies {
-    implementation 'org.pytorch:pytorch_android:1.8.0'
-    implementation 'org.pytorch:pytorch_android_torchvision:1.8.0'
+    implementation 'org.pytorch:pytorch_android_lite:1.9.0'
+    implementation 'org.pytorch:pytorch_android_torchvision:1.9.0'
 }
 ```
 Where `org.pytorch:pytorch_android` is the main dependency with PyTorch Android API, including libtorch native library for all 4 android abis (armeabi-v7a, arm64-v8a, x86, x86_64).
@@ -81,11 +82,11 @@ As a first step we read `image.jpg` to `android.graphics.Bitmap` using the stand
 Bitmap bitmap = BitmapFactory.decodeStream(getAssets().open("image.jpg"));
 ```
 
-#### 5. Loading TorchScript Module
+#### 5. Loading Mobile Module
 ```
-Module module = Module.load(assetFilePath(this, "model.pt"));
+Module module = Module.load(assetFilePath(this, "model.ptl"));
 ```
-`org.pytorch.Module` represents `torch::jit::script::Module` that can be loaded with `load` method specifying file path to the serialized to file model.
+`org.pytorch.Module` represents `torch::jit::mobile::Module` that can be loaded with `load` method specifying file path to the serialized to file model.
 
 #### 6. Preparing Input
 ```
@@ -389,6 +390,21 @@ SELECTED_OP_LIST=MobileNetV2.yaml scripts/build_pytorch_android.sh arm64-v8a
 
 After successful build you can integrate the result aar files to your android gradle project, following the steps from previous section of this tutorial (Building PyTorch Android from Source).
 
+## Use PyTorch JIT interpreter
+
+PyTorch JIT interpreter is the default interpreter before 1.9 (a version of our PyTorch interpreter that is not as size-efficient). It will still be supported in 1.9, and can be used via `build.gradle`:
+```
+repositories {
+    jcenter()
+}
+
+dependencies {
+    implementation 'org.pytorch:pytorch_android:1.9.0'
+    implementation 'org.pytorch:pytorch_android_torchvision:1.9.0'
+}
+```
+
+
 ## Android Tutorials
 
 Watch the following [video](https://youtu.be/5Lxuu16_28o) as PyTorch Partner Engineer Brad Heintz walks through steps for setting up the PyTorch Runtime for Android projects:

_mobile/home.md

@@ -24,15 +24,14 @@ PyTorch Mobile is in beta stage right now, and is already in wide scale producti
 * Support for tracing and scripting via TorchScript IR
 * Support for XNNPACK floating point kernel libraries for Arm CPUs
 * Integration of QNNPACK for 8-bit quantized kernels. Includes support for per-channel quantization, dynamic quantization and more
-* Build level optimization and selective compilation depending on the operators needed for user applications, i.e., the final binary size of the app is determined by the actual operators the app needs
+* Provides an [efficient mobile interpreter in Android and iOS](https://pytorch.org/tutorials/prototype/lite_interpreter.html). Also supports build level optimization and selective compilation depending on the operators needed for user applications (i.e., the final binary size of the app is determined by the actual operators the app needs).
 * Streamline model optimization via optimize_for_mobile
 * Support for hardware backends like GPU, DSP, and NPU will be available soon in Beta
 
 
 ## Prototypes
 We have launched the following features in prototype, available in the PyTorch nightly releases, and would love to get your feedback on the [PyTorch forums](https://discuss.pytorch.org/c/mobile/18):
 
-* Runtime binary size reduction via our [Lite Interpreter](https://pytorch.org/tutorials/prototype/lite_interpreter.html)
 * GPU support on [iOS via Metal](https://pytorch.org/tutorials/prototype/ios_gpu_workflow.html)
 * GPU support on [Android via Vulkan](https://pytorch.org/tutorials/prototype/vulkan_workflow.html)
 * DSP and NPU support on Android via [Google NNAPI](https://pytorch.org/tutorials/prototype/nnapi_mobilenetv2.html)

_mobile/ios.md

@@ -94,10 +94,10 @@ private lazy var module: TorchModule = {
     }
 }()
 ```
-Note that the `TorchModule` Class is an Objective-C wrapper of `torch::jit::script::Module`.
+Note that the `TorchModule` Class is an Objective-C wrapper of `torch::jit::mobile::Module`.
 
 ```cpp
-torch::jit::script::Module module = torch::jit::load(filePath.UTF8String);
+torch::jit::mobile::Module module = torch::jit::_load_for_mobile(filePath.UTF8String);
 ```
 Since Swift can not talk to C++ directly, we have to either use an Objective-C class as a bridge, or create a C wrapper for the C++ library. For demo purpose, we're going to wrap everything in this Objective-C class.
 
@@ -251,7 +251,8 @@ To use the custom built libraries the project, replace `#import <LibTorch/LibTor
 #include "caffe2/core/timer.h"
 #include "caffe2/utils/string_utils.h"
 #include "torch/csrc/autograd/grad_mode.h"
-#include "torch/csrc/jit/serialization/import.h"
+#include "torch/csrc/jit/mobile/import.h"
+#include "torch/csrc/jit/mobile/module.h"
 #include "torch/script.h"
 ```
 
@@ -289,6 +290,12 @@ SELECTED_OP_LIST=MobileNetV2.yaml BUILD_PYTORCH_MOBILE=1 IOS_ARCH=arm64 ./script
 torch::jit::GraphOptimizerEnabledGuard guard(false);
 ```
 
+## Use PyTorch JIT interpreter
+PyTorch JIT interpreter is the default interpreter before 1.9 (a version of our PyTorch interpreter that is not as size-efficient). It will still be supported in 1.9, and can be used in CocoaPods:
+```
+pod 'LibTorch', '~>1.9.0'
+```
+
 ## iOS Tutorials
 
 Watch the following [video](https://youtu.be/amTepUIR93k) as PyTorch Partner Engineer Brad Heintz walks through steps for setting up the PyTorch Runtime for iOS projects: