Add ONNX registry tutorial

advanced_source/super_resolution_with_onnxruntime.py

@@ -2,11 +2,11 @@
 (optional) Exporting a Model from PyTorch to ONNX and Running it using ONNX Runtime
 ===================================================================================
 
-.. Note::
+.. note::
     As of PyTorch 2.1, there are two versions of ONNX Exporter.
 
-    * ``torch.onnx.dynamo_export`is the newest (still in beta) exporter based on the TorchDynamo technology released with PyTorch 2.0
-    * ``torch.onnx.export`` is based on TorchScript backend and has been available since PyTorch 1.2.0
+    * ``torch.onnx.dynamo_export`is the newest (still in beta) exporter based on the TorchDynamo technology released with PyTorch 2.0.
+    * ``torch.onnx.export`` is based on TorchScript backend and has been available since PyTorch 1.2.0.
 
 In this tutorial, we describe how to convert a model defined
 in PyTorch into the ONNX format using the TorchScript ``torch.onnx.export` ONNX exporter.

beginner_source/onnx/README.txt

@@ -8,3 +8,7 @@ ONNX
 2. export_simple_model_to_onnx_tutorial.py
     Export a PyTorch model to ONNX
     https://pytorch.org/tutorials/beginner/onnx/export_simple_model_to_onnx_tutorial.html
+
+3. onnx_registry_tutorial.py
+    Introduction to ONNX Registry
+    https://pytorch.org/tutorials/beginner/onnx/onnx_registry_tutorial.html

beginner_source/onnx/export_simple_model_to_onnx_tutorial.py

@@ -1,7 +1,8 @@
 # -*- coding: utf-8 -*-
 """
 `Introduction to ONNX <intro_onnx.html>`_ ||
-**Export a PyTorch model to ONNX**
+**Export a PyTorch model to ONNX** ||
+`Introduction to ONNX Registry <onnx_registry_tutorial.html>`_
 
 Export a PyTorch model to ONNX
 ==============================
@@ -104,7 +105,7 @@ def forward(self, x):
 export_output.save("my_image_classifier.onnx")
 
 ######################################################################
-# The ONNX file can be loaded back into memory and checked if it is well formed with the following code:
+# You can load the ONNX file back into memory and check if it is well formed with the following code:
 
 import onnx
 onnx_model = onnx.load("my_image_classifier.onnx")
@@ -115,7 +116,7 @@ def forward(self, x):
 # ----------------------------------------------
 #
 # Now that we have our model saved in a file, we can visualize it with `Netron <https://github.com/lutzroeder/netron>`_.
-# Netron can either be installed on macos, Linux or Windows computers, or run directly from the browser.
+# Netron can either be installed on macOS, Linux or Windows computers, or run directly from the browser.
 # Let's try the web version by opening the following link: https://netron.app/.
 #
 # .. image:: ../../_static/img/onnx/netron_web_ui.png

beginner_source/onnx/intro_onnx.py

@@ -1,6 +1,7 @@
 """
 **Introduction to ONNX** ||
-`Export a PyTorch model to ONNX <export_simple_model_to_onnx_tutorial.html>`_
+`Export a PyTorch model to ONNX <export_simple_model_to_onnx_tutorial.html>`_ ||
+`Introduction to ONNX Registry <onnx_registry_tutorial.html>`_
 
 Introduction to ONNX
 ====================
@@ -32,10 +33,12 @@
 Dependencies
 ------------
 
+PyTorch 2.1.0 or newer is required.
+
 The ONNX exporter depends on extra Python packages:
 
-  - `ONNX <https://onnx.ai>`_
-  - `ONNX Script <https://onnxscript.ai>`_
+  - `ONNX <https://onnx.ai>`_ standard implementation
+  - `ONNX Script <https://onnxscript.ai>`_  is leveraged to enable developers to naturally author ONNX operators, functions and models using a subset of Python in an expressive, and yet simple fashion.
 
 They can be installed through `pip <https://pypi.org/project/pip/>`_: