Merge pull request #55 from jeffxtang/update4pt19

xta0 · web-flow · commit f2b9aa196821 · 2021-06-16T13:52:59.000-07:00
Update of Podfile's, README's, and models in the remaining demo apps for PyTorch 1.9
diff --git a/HelloWorld/HelloWorld/HelloWorld/Assets.xcassets/AppIcon.appiconset/Contents.json b/HelloWorld/HelloWorld/HelloWorld/Assets.xcassets/AppIcon.appiconset/Contents.json
@@ -2,97 +2,103 @@
   "images" : [
     {
       "idiom" : "iphone",
-      "size" : "20x20",
-      "scale" : "2x"
+      "scale" : "2x",
+      "size" : "20x20"
     },
     {
       "idiom" : "iphone",
-      "size" : "20x20",
-      "scale" : "3x"
+      "scale" : "3x",
+      "size" : "20x20"
     },
     {
+      "filename" : "Icon-Small@2x-1.png",
       "idiom" : "iphone",
-      "size" : "29x29",
-      "scale" : "2x"
+      "scale" : "2x",
+      "size" : "29x29"
     },
     {
+      "filename" : "Icon-Small@3x.png",
       "idiom" : "iphone",
-      "size" : "29x29",
-      "scale" : "3x"
+      "scale" : "3x",
+      "size" : "29x29"
     },
     {
+      "filename" : "Icon-Small-40@2x.png",
       "idiom" : "iphone",
-      "size" : "40x40",
-      "scale" : "2x"
+      "scale" : "2x",
+      "size" : "40x40"
     },
     {
+      "filename" : "Icon-Small-60@2x.png",
       "idiom" : "iphone",
-      "size" : "40x40",
-      "scale" : "3x"
+      "scale" : "3x",
+      "size" : "40x40"
     },
     {
+      "filename" : "Icon-60@2x.png",
       "idiom" : "iphone",
-      "size" : "60x60",
-      "scale" : "2x"
+      "scale" : "2x",
+      "size" : "60x60"
     },
     {
+      "filename" : "Icon-60@3x.png",
       "idiom" : "iphone",
-      "size" : "60x60",
-      "scale" : "3x"
+      "scale" : "3x",
+      "size" : "60x60"
     },
     {
       "idiom" : "ipad",
-      "size" : "20x20",
-      "scale" : "1x"
+      "scale" : "1x",
+      "size" : "20x20"
     },
     {
       "idiom" : "ipad",
-      "size" : "20x20",
-      "scale" : "2x"
+      "scale" : "2x",
+      "size" : "20x20"
     },
     {
       "idiom" : "ipad",
-      "size" : "29x29",
-      "scale" : "1x"
+      "scale" : "1x",
+      "size" : "29x29"
     },
     {
       "idiom" : "ipad",
-      "size" : "29x29",
-      "scale" : "2x"
+      "scale" : "2x",
+      "size" : "29x29"
     },
     {
       "idiom" : "ipad",
-      "size" : "40x40",
-      "scale" : "1x"
+      "scale" : "1x",
+      "size" : "40x40"
     },
     {
       "idiom" : "ipad",
-      "size" : "40x40",
-      "scale" : "2x"
+      "scale" : "2x",
+      "size" : "40x40"
     },
     {
       "idiom" : "ipad",
-      "size" : "76x76",
-      "scale" : "1x"
+      "scale" : "1x",
+      "size" : "76x76"
     },
     {
       "idiom" : "ipad",
-      "size" : "76x76",
-      "scale" : "2x"
+      "scale" : "2x",
+      "size" : "76x76"
     },
     {
       "idiom" : "ipad",
-      "size" : "83.5x83.5",
-      "scale" : "2x"
+      "scale" : "2x",
+      "size" : "83.5x83.5"
     },
     {
       "idiom" : "ios-marketing",
-      "size" : "1024x1024",
-      "scale" : "1x"
+      "scale" : "1x",
+      "size" : "1024x1024"
     }
   ],
   "info" : {
-    "version" : 1,
-    "author" : "xcode"
+    "author" : "xcode",
+    "version" : 1
   }
-}
+}
diff --git a/HelloWorld/HelloWorld/HelloWorld/Assets.xcassets/AppIcon.appiconset/Icon-60@2x.png b/HelloWorld/HelloWorld/HelloWorld/Assets.xcassets/AppIcon.appiconset/Icon-60@2x.png
diff --git a/HelloWorld/HelloWorld/HelloWorld/Assets.xcassets/AppIcon.appiconset/Icon-60@3x.png b/HelloWorld/HelloWorld/HelloWorld/Assets.xcassets/AppIcon.appiconset/Icon-60@3x.png
diff --git a/HelloWorld/HelloWorld/HelloWorld/Assets.xcassets/AppIcon.appiconset/Icon-Small-40@2x.png b/HelloWorld/HelloWorld/HelloWorld/Assets.xcassets/AppIcon.appiconset/Icon-Small-40@2x.png
diff --git a/HelloWorld/HelloWorld/HelloWorld/Assets.xcassets/AppIcon.appiconset/Icon-Small-60@2x.png b/HelloWorld/HelloWorld/HelloWorld/Assets.xcassets/AppIcon.appiconset/Icon-Small-60@2x.png
diff --git a/HelloWorld/HelloWorld/HelloWorld/Assets.xcassets/AppIcon.appiconset/Icon-Small@2x-1.png b/HelloWorld/HelloWorld/HelloWorld/Assets.xcassets/AppIcon.appiconset/Icon-Small@2x-1.png
diff --git a/HelloWorld/HelloWorld/HelloWorld/Assets.xcassets/AppIcon.appiconset/Icon-Small@3x.png b/HelloWorld/HelloWorld/HelloWorld/Assets.xcassets/AppIcon.appiconset/Icon-Small@3x.png
diff --git a/HelloWorld/HelloWorld/HelloWorld/Assets.xcassets/Contents.json b/HelloWorld/HelloWorld/HelloWorld/Assets.xcassets/Contents.json
@@ -1,6 +1,6 @@
 {
   "info" : {
-    "version" : 1,
-    "author" : "xcode"
+    "author" : "xcode",
+    "version" : 1
   }
-}
+}
diff --git a/HelloWorld/HelloWorld/HelloWorld/model/model.pt b/HelloWorld/HelloWorld/HelloWorld/model/model.pt
diff --git a/HelloWorld/HelloWorld/Podfile b/HelloWorld/HelloWorld/Podfile
@@ -1,5 +1,5 @@
 
 platform :ios, '12.0'
 target 'HelloWorld' do
-    pod 'LibTorch', '~> 1.6.0'
+    pod 'LibTorch', '~> 1.9.0'
 end
diff --git a/HelloWorld/README.md b/HelloWorld/README.md
@@ -14,7 +14,7 @@ The model we are going to use is [MobileNet v2](https://pytorch.org/hub/pytorch_
 > We highly recommend following the [Pytorch Github page](https://github.com/pytorch/pytorch) to set up the Python development environment on your local machine.
 
 ```shell
-pip install torchvision
+pip install torch torchvision
 ```
 
 Once we have TorchVision installed successfully, navigate to the HelloWorld folder and run `trace_model.py` to generate our model. The script contains the code of tracing and saving a [torchscript model](https://pytorch.org/tutorials/beginner/Intro_to_TorchScript_tutorial.html) that can be run on mobile devices.
@@ -23,7 +23,7 @@ Once we have TorchVision installed successfully, navigate to the HelloWorld fold
 python trace_model.py
 ```
 
-If everything works well, `model.pt` should be generated in the `HelloWorld` folder. Now copy the model file to our application folder `HelloWorld/model`.
+If everything works well, `model.pt` should be generated and saved in the `HelloWorld/HelloWorld/model`.
 
 ### Install LibTorch via Cocoapods
 
diff --git a/HelloWorld/trace_model.py b/HelloWorld/trace_model.py
@@ -7,4 +7,4 @@
 example = torch.rand(1, 3, 224, 224)
 traced_script_module = torch.jit.trace(model, example)
 torchscript_model_optimized = optimize_for_mobile(traced_script_module)
-torchscript_model_optimized.save("model.pt")
+torchscript_model_optimized.save("HelloWorld/HelloWorld/model/model.pt")
diff --git a/ObjectDetection/Podfile b/ObjectDetection/Podfile
@@ -6,5 +6,5 @@ target 'ObjectDetection' do
   use_frameworks!
 
   # Pods for ObjectDetection
-  pod 'LibTorch', '~>1.7.0'
+  pod 'LibTorch', '~>1.9.0'
 end
diff --git a/ObjectDetection/README.md b/ObjectDetection/README.md
@@ -6,9 +6,9 @@
 
 ## Prerequisites
 
-* PyTorch 1.7 or later (Optional)
+* PyTorch 1.9.0 or later (Optional)
 * Python 3.8 (Optional)
-* iOS Pytorch pod library 1.7
+* iOS Cocoapods library LibTorch 1.9.0
 * Xcode 12 or later
 
 ## Quick Start
@@ -19,7 +19,7 @@ To Test Run the Object Detection iOS App, follow the steps below:
 
 If you don't have the PyTorch environment set up to run the script, you can download the model file [here](https://drive.google.com/file/d/1QOxNfpy_j_1KbuhN8INw2AgAC82nEw0F/view?usp=sharing) to the `ios-demo-app/ObjectDetection/ObjectDetection` folder, then skip the rest of this step and go to step 2 directly.
 
-Be aware that the downloadable model file was created with PyTorch 1.7.0, matching the iOS LibTorch library 1.7.0 specified in the `Podfile`. If you use a different version of PyTorch to create your model by following the instructions below, make sure you specify the same iOS LibTorch version in the `Podfile` to avoid possible errors caused by the version mismatch. Furthermore, if you want to use the latest prototype features in the PyTorch master branch to create the model, follow the steps at [Building PyTorch iOS Libraries from Source](https://pytorch.org/mobile/ios/#build-pytorch-ios-libraries-from-source) on how to use the model in iOS.
+Be aware that the downloadable model file was created with PyTorch 1.9.0, matching the iOS LibTorch library 1.9.0 specified in the `Podfile`. If you use a different version of PyTorch to create your model by following the instructions below, make sure you specify the same iOS LibTorch version in the `Podfile` to avoid possible errors caused by the version mismatch. Furthermore, if you want to use the latest prototype features in the PyTorch master branch to create the model, follow the steps at [Building PyTorch iOS Libraries from Source](https://pytorch.org/mobile/ios/#build-pytorch-ios-libraries-from-source) on how to use the model in iOS.
 
 The Python script `export.py` in the `models` folder of the [YOLOv5 repo](https://github.com/ultralytics/yolov5) is used to generate a TorchScript-formatted YOLOv5 model named `yolov5s.torchscript.pt` for mobile apps.
 
diff --git a/PyTorchDemo/Podfile b/PyTorchDemo/Podfile
@@ -1,4 +1,4 @@
 platform :ios, '14.3'
 target 'PyTorchDemo' do
-  pod 'LibTorch', '~>1.7.0'
+  pod 'LibTorch', '~>1.9.0'
 end
diff --git a/QuestionAnswering/Podfile b/QuestionAnswering/Podfile
@@ -6,5 +6,5 @@ target 'QuestionAnswering' do
   use_frameworks!
 
   # Pods for QuestionAnswering
-  pod 'LibTorch', '~>1.7.0'
+  pod 'LibTorch', '~>1.9.0'
 end
diff --git a/QuestionAnswering/README.md b/QuestionAnswering/README.md
@@ -8,9 +8,9 @@ In this demo app, we'll show how to quantize and convert the Huggingface's Disti
 
 ## Prerequisites
 
-* PyTorch 1.7 or later (Optional)
+* PyTorch 1.9 or later (Optional)
 * Python 3.8 (Optional)
-* iOS Pytorch pod library 1.7
+* iOS Cocoapods library LibTorch 1.9.0
 * Xcode 12 or later
 
 ## Quick Start
@@ -21,13 +21,13 @@ To Test Run the iOS QA demo app, run the following commands on a Terminal:
 
 If you don't have PyTorch installed or want to have a quick try of the demo app, you can download the scripted QA model compressed in a zip file [here](https://drive.google.com/file/d/1RWZa_5oSQg5AfInkn344DN3FJ5WbbZbq/view?usp=sharing), then unzip, drag and drop it to the project, and continue to Step 2.
 
-Be aware that the downloadable model file was created with PyTorch 1.7.0, matching the iOS LibTorch library 1.7.0 specified in the `Podfile`. If you use a different version of PyTorch to create your model by following the instructions below, make sure you specify the same iOS LibTorch version in the `Podfile` to avoid possible errors caused by the version mismatch. Furthermore, if you want to use the latest prototype features in the PyTorch master branch to create the model, follow the steps at [Building PyTorch iOS Libraries from Source](https://pytorch.org/mobile/ios/#build-pytorch-ios-libraries-from-source) on how to use the model in iOS.
+Be aware that the downloadable model file was created with PyTorch 1.9.0, matching the iOS LibTorch library 1.9.0 specified in the `Podfile`. If you use a different version of PyTorch to create your model by following the instructions below, make sure you specify the same iOS LibTorch version in the `Podfile` to avoid possible errors caused by the version mismatch. Furthermore, if you want to use the latest prototype features in the PyTorch master branch to create the model, follow the steps at [Building PyTorch iOS Libraries from Source](https://pytorch.org/mobile/ios/#build-pytorch-ios-libraries-from-source) on how to use the model in iOS.
 
-With PyTorch 1.7 installed, first install the Huggingface `transformers` by running `pip install transformers` (the versions that have been tested are 4.0.0 and 4.1.1), then run `python convert_distilbert_qa.py`.
+With PyTorch 1.9.0 installed, first install the Huggingface `transformers` by running `pip install transformers`, then run `python convert_distilbert_qa.py`.
 
 Note that a pre-defined question and text, resulting in the size of the input tokens (of question and text) being 360, is used in the `convert_distilbert_qa.py`, and 360 is the maximum token size for the user text and question in the app. If the token size of the inputs of the text and question is less than 360, padding will be needed to make the model work correctly.
 
-After the script completes, drag and drop the model file qa360_quantized.pt to the iOS app project. [Dynamic quantization](https://pytorch.org/tutorials/intermediate/dynamic_quantization_bert_tutorial.html) is used to quantize the model to reduce its size to half, without causing inference difference in question answering - you can verify this changing the last 4 lines of code in `convert_distilbert_qa.py` from:
+After the script completes, drag and drop the model file `qa360_quantized.pt` to the iOS app project. [Dynamic quantization](https://pytorch.org/tutorials/intermediate/dynamic_quantization_bert_tutorial.html) is used to quantize the model to reduce its size to half, without causing inference difference in question answering - you can verify this changing the last 4 lines of code in `convert_distilbert_qa.py` from:
 
 ```
 model_dynamic_quantized = torch.quantization.quantize_dynamic(model, qconfig_spec={torch.nn.Linear}, dtype=torch.qint8)
diff --git a/Seq2SeqNMT/Podfile b/Seq2SeqNMT/Podfile
@@ -1,4 +1,4 @@
 platform :ios, '12.0'
 target 'Seq2SeqNMT' do
-  pod 'LibTorch', '~>1.7.0'
+  pod 'LibTorch', '~>1.9.0'
 end
diff --git a/Seq2SeqNMT/README.md b/Seq2SeqNMT/README.md
@@ -12,9 +12,9 @@ This iOS demo app shows:
 
 ## Prerequisites
 
-* PyTorch 1.7 or later (Optional)
+* PyTorch 1.9.0 or later (Optional)
 * Python 3.8 (Optional)
-* iOS Pytorch pod library 1.7
+* iOS Cocoapods library LibTorch 1.9.0
 * Xcode 12 or later
 
 ## Quick Start
@@ -25,11 +25,11 @@ To Test Run the Object Detection iOS App, follow the steps below:
 
 If you don't have the PyTorch environment set up to run the script, you can download the PyTorch trained and optimized NMT encoder and decoder models compressed in a zip [here](https://drive.google.com/file/d/1Ju9ceHi5e87UW1P09-XIvPVdMjOs5kiE/view?usp=sharing), unzip it, copy to the iOS app project folder, and continue to Step 2.
 
-Be aware that the downloadable model files were created with PyTorch 1.7.0, matching the iOS LibTorch library 1.7.0 specified in the `Podfile`. If you use a different version of PyTorch to create your model by following the instructions below, make sure you specify the same iOS LibTorch version in the `Podfile` to avoid possible errors caused by the version mismatch. Furthermore, if you want to use the latest prototype features in the PyTorch master branch to create the model, follow the steps at [Building PyTorch iOS Libraries from Source](https://pytorch.org/mobile/ios/#build-pytorch-ios-libraries-from-source) on how to use the model in iOS.
+Be aware that the downloadable model files were created with PyTorch 1.9.0, matching the iOS LibTorch library 1.9.0 specified in the `Podfile`. If you use a different version of PyTorch to create your model by following the instructions below, make sure you specify the same iOS LibTorch version in the `Podfile` to avoid possible errors caused by the version mismatch. Furthermore, if you want to use the latest prototype features in the PyTorch master branch to create the model, follow the steps at [Building PyTorch iOS Libraries from Source](https://pytorch.org/mobile/ios/#build-pytorch-ios-libraries-from-source) on how to use the model in iOS.
 
 If you have a good GPU and want to train your model from scratch, uncomment the line `trainIters(encoder, decoder, 450100, print_every=5000)` in `seq2seq_nmt.py` before running `python seq2seq2_nmt.py` to go through the whole process of training, saving, loading, optimizing and saving the final mobile-ready models.
 
-To just convert a pre-trained model `seq2seq_mt_150000.pt` to the TorchScript model used on mobile, download [seq2seq_mt_150000.pt](https://drive.google.com/file/d/1f91PvlkxS8JS0xGpMRZ3fmr0Ev80Guxk/view?usp=sharing) first to the same directory as `seq2seq2_nmt.py`, then run `python seq2seq2_nmt.py`. After `optimized_encoder_150k.pth` and `optimized_decoder_150k.pth` are generated, copy them to the iOS app. Note that dynamic quantization is applied to the decoder in `seq2seq2_nmt.py` for its `nn.Linear` parameters to reduce the decoder model size from 29MB to 18MB.
+To just convert a pre-trained model `seq2seq_mt_150000.pt` to the TorchScript model used on mobile, download [seq2seq_mt_150000.pt](https://drive.google.com/file/d/1f91PvlkxS8JS0xGpMRZ3fmr0Ev80Guxk/view?usp=sharing) first to the same directory as `seq2seq2_nmt.py`, then run `python seq2seq_nmt.py`. After `optimized_encoder_150k.pth` and `optimized_decoder_150k.pth` are generated, copy them to the iOS app. Note that dynamic quantization is applied to the decoder in `seq2seq2_nmt.py` for its `nn.Linear` parameters to reduce the decoder model size from 29MB to 18MB.
 
 ### 2. Use LibTorch
 
diff --git a/TorchVideo/Podfile b/TorchVideo/Podfile
@@ -6,5 +6,5 @@ target 'TorchVideo' do
   use_frameworks!
 
   # Pods for TorchVideo
-  pod 'LibTorch', '~>1.8.0'
+  pod 'LibTorch', '~>1.9.0'
 end
diff --git a/TorchVideo/README.md b/TorchVideo/README.md
@@ -6,9 +6,9 @@ The newly released open-sourced [PyTorchVideo](https://github.com/facebookresear
 
 ## Prerequisites
 
-* PyTorch 1.8.0/1.8.1, torchvision 0.9.1, PyTorchVideo (Optional)
+* PyTorch 1.9.0, torchvision 0.10.0, PyTorchVideo 0.1.1 (Optional)
 * Python 3.8 or above (Optional)
-* iOS PyTorch pod library 1.8.0
+* iOS Cocoapods library LibTorch 1.9.0
 * Xcode 12 or later
 
 
@@ -18,19 +18,19 @@ The newly released open-sourced [PyTorchVideo](https://github.com/facebookresear
 
 If you don't have the PyTorch environment set up to run the script, you can download the model file `video_classification.pt` [here](https://drive.google.com/file/d/1qweDu7QZv7xJA7Sx_UIxjvcS7y1rQ2kE/view) to the `ios-demo-app/TorchVideo/TorchVideo` folder, then skip the rest of this step and go to step 2 directly.
 
-Be aware that the downloadable model file was created with PyTorch 1.8.1, matching the iOS LibTorch library 1.8.0 specified in the `Podfile`. If you use a different version of PyTorch to create your model by following the instructions below, make sure you specify the same iOS LibTorch version in the `Podfile` to avoid possible errors caused by the version mismatch. Furthermore, if you want to use the latest prototype features in the PyTorch master branch to create the model, follow the steps at [Building PyTorch iOS Libraries from Source](https://pytorch.org/mobile/ios/#build-pytorch-ios-libraries-from-source) on how to use the model in iOS.
+Be aware that the downloadable model file was created with PyTorch 1.9.0, matching the iOS LibTorch library 1.9.0 specified in the `Podfile`. If you use a different version of PyTorch to create your model by following the instructions below, make sure you specify the same iOS LibTorch version in the `Podfile` to avoid possible errors caused by the version mismatch. Furthermore, if you want to use the latest prototype features in the PyTorch master branch to create the model, follow the steps at [Building PyTorch iOS Libraries from Source](https://pytorch.org/mobile/ios/#build-pytorch-ios-libraries-from-source) on how to use the model in iOS.
 
 To create the model yourself, simply run the following commands:
 ```
-conda create -n pt181 python=3.8.5
-conda activate pt181
+conda create -n pt19 python=3.8.5
+conda activate pt19
 pip install torch torchvision
+pip install pytorchvideo
 
 # pip list|grep torch
-# torch             1.8.1
-# torchvision       0.9.1
-
-pip install pytorchvideo
+# torch             1.9.0
+# torchvision       0.10.0
+# pytorchvideo      0.1.1
 
 cd ios-demo-app/TorchVideo
 python build_model.py

Original file line number	Diff line number	Diff line change
`@@ -1,6 +1,6 @@`
`1`	`1`	`{`
`2`	`2`	`"info" : {`
`3`		`- "version" : 1,`
`4`		`- "author" : "xcode"`
	`3`	`+ "author" : "xcode",`
	`4`	`+ "version" : 1`
`5`	`5`	`}`
`6`		`-}`
	`6`	`+}`