[JIT] Update docs for recently added features #45232

SplitInfinity · 2020-09-23T21:26:32Z

Stack from ghstack:

[JIT] Add recursive scripting for class type module attributes #45234 [JIT] Add recursive scripting for class type module attributes
[JIT] Add API for ignoring arbitrary module attributes #45262 [JIT] Add API for ignoring arbitrary module attributes
[JIT] Enable @unused syntax for ignoring properties #45261 [JIT] Enable @unused syntax for ignoring properties
[JIT] Add test for ignored class type property #45233 [JIT] Add test for ignored class type property
[JIT] Update docs for recently added features #45232 [JIT] Update docs for recently added features

Summary
This commit updates the TorchScript language reference to include
documentation on recently-added TorchScript enums. It also removed
torch.no_grad from the list of known unsupported torch modules and
classes because it is now supported.

Test Plan
Continuous integration.

Differential Revision: D23971884

**Summary** This commit updates the TorchScript language reference to include documentation on recently-added TorchScript enums. It also removed `torch.no_grad` from the list of known unsupported `torch` modules and classes because it is now supported. **Test Plan** Continuous integration. [ghstack-poisoned]

SplitInfinity · 2020-09-23T21:33:46Z

Where's a good place to mention property support? They're not part of the core Python language so jit_python_reference.rst is not the right place to put documentation about them. A new section in jit_language_reference.rst, maybe?

SplitInfinity · 2020-09-23T21:34:25Z

docs/source/jit_unsupported.rst

  * :class:`torch.nn.RNN`
  * :class:`torch.nn.AdaptiveLogSoftmaxWithLoss`
  * :class:`torch.autograd.Function`
-  * :class:`torch.autograd.no_grad`


Is there a good place to mention that record_function is supported now?

docs/source/jit_language_reference.rst

dr-ci · 2020-09-23T22:02:48Z

💊 CI failures summary and remediations

As of commit 5b009a6 (more details on the Dr. CI page):

💚 💚 Looks good so far! There are no failures yet. 💚 💚

This comment was automatically generated by Dr. CI (expand for details).

Follow this link to opt-out of these comments for your Pull Requests.

Please report bugs/suggestions on the GitHub issue tracker or post in the (internal) Dr. CI Users group.

See how this bot performed.

This comment has been revised 22 times.

**Summary** This commit updates the TorchScript language reference to include documentation on recently-added TorchScript enums. It also removed `torch.no_grad` from the list of known unsupported `torch` modules and classes because it is now supported. **Test Plan** Continuous integration. [ghstack-poisoned]

eellison

Looks good, a couple comments

eellison · 2020-09-24T00:19:27Z

docs/source/jit_language_reference.rst

+
+.. warning::
+
+    TorchScript enum support is experimental..


IMO we don't need to mark it as experimental - there may be bugs but that's true of everything else too 🤷

eellison · 2020-09-24T00:19:40Z

docs/source/jit_language_reference.rst



+.. _TorchScript Enums:
+.. _TorchScript Enums:


Is the triple repeat here intentional?

I copied over the class types section, I don't know how to write .rst lol

eellison · 2020-09-24T18:53:27Z

docs/source/jit_language_reference.rst

+
+    TorchScript enum support is experimental..
+
+Python enums can be used in TorchScript without any extra annotation or code:


Personally I would cut down on this section a bit - it's pretty long for what is an easy & simple feature to use (enums). see how short the named tuple section is

What do you think I should cut out? Most of it is code samples.

eellison

Meant to accept - these are just suggestions though, it looks good obviously feel free to keep whatever

eellison · 2020-09-24T18:59:18Z

docs/source/jit_language_reference.rst

+
+::
+
+    from enum import Enum
+
+    class IntEnum(Enum):
+        FOO = 1
+        BAR = 2
+
+    class FloatEnum(Enum):
+        FOO = 1.2
+        BAR = 2.3
+
+    class StringEnum(Enum):
+        FOO = "foo as in foo bar"
+        BAR = "bar as in foo bar"
+


Would cut out this part

eellison · 2020-09-24T18:59:29Z

docs/source/jit_language_reference.rst

+
+    from enum import Enum
+
+    class Color(Enum):
+        RED = 1
+        GREEN = 2
+
+    @torch.jit.script
+    def return_enum(cond: bool):
+        if cond:
+            return Color.RED
+        else:
+            return Color.GREEN
+
+    self.assertEqual(return_enum(True), Color.RED)
+    self.assertEqual(return_enum(False), Color.GREEN)
+


And this part

eellison · 2020-09-24T19:00:35Z

docs/source/jit_language_reference.rst

+        if cond:
+            return Color.RED
+        else:
+            return Color.GREEN


And then maybe include this example where you access the field name Color.RED or make it part of the other function

**Summary** This commit updates the TorchScript language reference to include documentation on recently-added TorchScript enums. It also removed `torch.no_grad` from the list of known unsupported `torch` modules and classes because it is now supported. **Test Plan** Continuous integration. [ghstack-poisoned]

codecov · 2020-09-27T10:45:15Z

Codecov Report

Merging #45232 into gh/splitinfinity/49/base will decrease coverage by 0.00%.
The diff coverage is n/a.

@@                     Coverage Diff                      @@
##           gh/splitinfinity/49/base   #45232      +/-   ##
============================================================
- Coverage                     68.05%   68.05%   -0.01%     
============================================================
  Files                           396      396              
  Lines                         51235    51235              
============================================================
- Hits                          34868    34867       -1     
- Misses                        16367    16368       +1

Impacted Files	Coverage Δ
torch/testing/_internal/expecttest.py	`77.55% <0.00%> (-1.03%)`	⬇️

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 36c3fbc...5b009a6. Read the comment docs.

**Summary** This commit updates the TorchScript language reference to include documentation on recently-added TorchScript enums. It also removed `torch.no_grad` from the list of known unsupported `torch` modules and classes because it is now supported. **Test Plan** Continuous integration. [ghstack-poisoned]

facebook-github-bot · 2020-09-29T02:20:31Z

@SplitInfinity merged this pull request in 4af4b71.

This was referenced Sep 23, 2020

[JIT] Add test for ignored class type property #45233

Closed

[JIT] Add recursive scripting for class type module attributes #45234

Closed

SplitInfinity requested review from eellison and gmagogsfm September 23, 2020 21:31

SplitInfinity commented Sep 23, 2020

View reviewed changes

gmagogsfm reviewed Sep 23, 2020

View reviewed changes

docs/source/jit_language_reference.rst Outdated Show resolved Hide resolved

docs/source/jit_language_reference.rst Outdated Show resolved Hide resolved

This was referenced Sep 24, 2020

[JIT] Enable @unused syntax for ignoring properties #45261

Closed

[JIT] Add API for ignoring arbitrary module attributes #45262

Closed

eellison reviewed Sep 24, 2020

View reviewed changes

eellison approved these changes Sep 24, 2020

View reviewed changes

Meghan Lele added 3 commits September 24, 2020 12:34

Meghan Lele added 2 commits September 28, 2020 12:42

facebook-github-bot closed this in 4af4b71 Sep 29, 2020

facebook-github-bot added the merged label Sep 29, 2020

facebook-github-bot deleted the gh/splitinfinity/49/head branch October 2, 2020 14:17

mruberry added the Merged label Oct 28, 2020


		TorchScript enum support is experimental..

		Python enums can be used in TorchScript without any extra annotation or code:

[JIT] Update docs for recently added features #45232

[JIT] Update docs for recently added features #45232

Uh oh!

Conversation

SplitInfinity commented Sep 23, 2020 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

SplitInfinity commented Sep 23, 2020

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

dr-ci bot commented Sep 23, 2020 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

💊 CI failures summary and remediations

Uh oh!

eellison left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

eellison Sep 24, 2020 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

eellison left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

codecov bot commented Sep 27, 2020 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Codecov Report

Uh oh!

facebook-github-bot commented Sep 29, 2020

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

6 participants

SplitInfinity commented Sep 23, 2020 •

edited

Loading

dr-ci bot commented Sep 23, 2020 •

edited

Loading

eellison Sep 24, 2020 •

edited

Loading

codecov bot commented Sep 27, 2020 •

edited

Loading