Summary:
Previously `torch.jit.trace` relies on AutoGrad hooks to infer name of tensors in computation, including those of function/method arguments. This often doesn't work out because:
- These names often do not exist
- Tracer uses argument name of first tensor operation on each tensor as inferred argument names. These tensor operations have programmatically-generated names like `argument_1`
This PR extracts argument names directly from Python functions and pass them down to tracer, which then assigns them to correct graph inputs. This way, we always have the correct argument names captured in IR.
This is useful for both debugging and supporting using `InterfaceType` to represent traced modules.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/51775
Reviewed By: izdeby
Differential Revision: D26273105
Pulled By: gmagogsfm
fbshipit-source-id: 934a385041137dc3731bb6fa8657b11532fed9e5
Summary:
The attributes in `dir(mod)` may not be valid, this will throw error when calling `getattr`.
Use `hasattr` to test if it is valid.
Here is an example:
```python
class A:
def __init__(self, x):
if x:
self._attr = 1
property
def val(self):
return getattr(self, '_attr')
a = A(False)
print('val' in dir(a))
print(hasattr(a, 'val'))
b = A(True)
print('val' in dir(b))
print(hasattr(b, 'val'))
```
And the outputs:
```
True
False
True
True
```
Pull Request resolved: https://github.com/pytorch/pytorch/pull/50680
Reviewed By: malfet
Differential Revision: D26103975
Pulled By: eellison
fbshipit-source-id: 67a799afe7d726153c91654d483937c5e198ba94
Summary:
These unused variables were identified by [pyflakes](https://pypi.org/project/pyflakes/). They can be safely removed to simplify the code.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/50181
Reviewed By: gchanan
Differential Revision: D25844270
fbshipit-source-id: 0e648ffe8c6db6daf56788a13ba89806923cbb76
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/48741
**Summary**
This commit fixes a bug in the handling of `ast.Subscript` inside
`get_annotation_str`. `annotation.value` (which contains the AST node
representing the container name) should also be processed using
`get_annotation_str`.
**Test Plan**
This commit adds a unit test to `TestClassType` based on the test case
from the issue that reported this bug.
**Fixes**
This commit fixes#47570.
Test Plan: Imported from OSS
Reviewed By: ppwwyyxx
Differential Revision: D25286013
Pulled By: SplitInfinity
fbshipit-source-id: 61a9e5dc16d9f87b80578f78d537f91332093e52
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/47734
**Summary**
This commit allows `torch.device` to be resolved properly when used in
class types that are recursively scripted. This is accomplished by augmenting
the resolution callback used during recursively class scripting to include
the type annotations used on class method declarations.
Classes that are not explicitly annotated with `torch.jit.script` are
implicitly scripted during the compilation of a function or class method
that uses them. One key difference between this method of class type
compilation and explicit scripting is that the former uses a resolution callback
that can only resolve variables that class methods close over (see
`_jit_internal.createResolutionCallbackForClassMethods`). This does
not include type annotations and default arguments. This means that
builtin types like `torch.Tensor` and `torch.device` cannot be resolved
using the resolution callback. This issue does not arise when explicitly
scripting classes because the resolution callback for that code path is
constructed from scope of the class definition
(see `_jit_internal.createResolutionCallbackFromFrame`). `torch.Tensor`
and `torch.device` are almost always present in that scope, usually from
`import`ing `torch`.
**Test Plan**
This commit adds a new unit test to `TestClassType`,
`test_recursive_script_builtin_type_resolution`.
**Fixes**
This commit closes#47405.
Test Plan: Imported from OSS
Reviewed By: eellison
Differential Revision: D24995374
Pulled By: SplitInfinity
fbshipit-source-id: db68212634cacf81cfaeda8095a1fe5105fa73b7
Summary:
If there is no annotation given, we want to show users that the type is inferred
Pull Request resolved: https://github.com/pytorch/pytorch/pull/46969
Test Plan:
Added a new test case that throws an error with the expected error message
Fixes https://github.com/pytorch/pytorch/issues/46326
Reviewed By: ZolotukhinM
Differential Revision: D24614450
Pulled By: gmagogsfm
fbshipit-source-id: dec555a53bfaa9cdefd3b21b5142f5e522847504
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/45261
**Summary**
This commit enables `unused` syntax for ignoring
properties. Inoring properties is more intuitive with this feature enabled.
`ignore` is not supported because class type properties cannot be
executed in Python (because they exist only as TorchScript types) like
an `ignored` function and module properties that cannot be scripted
are not added to the `ScriptModule` wrapper so that they
may execute in Python.
**Test Plan**
This commit updates the existing unit tests for class type and module
properties to test properties ignored using `unused`.
Test Plan: Imported from OSS
Reviewed By: navahgar, Krovatkin, mannatsingh
Differential Revision: D23971881
Pulled By: SplitInfinity
fbshipit-source-id: 8d3cc1bbede7753d6b6f416619e4660c56311d33
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/44586
**Summary**
This commit disallows plain `Optional` type annotations without
any contained types both in type comments and in-line as
Python3-style type annotations.
**Test Plan**
This commit adds a unit test for these two situations.
Test Plan: Imported from OSS
Reviewed By: gmagogsfm
Differential Revision: D23721517
Pulled By: SplitInfinity
fbshipit-source-id: ead411e94aa0ccce227af74eb0341e2a5331370a
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/44585
**Summary**
This commit disallows plain `Tuple` type annotations without any
contained types both in type comments and in-line as Python3-style
type annotations.
**Test Plan**
This commit adds a unit test for these two situations.
Test Plan: Imported from OSS
Reviewed By: gmagogsfm
Differential Revision: D23721515
Pulled By: SplitInfinity
fbshipit-source-id: e11c77a4fac0b81cd535c37a31b9f4129c276592
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/44584
**Summary**
This commit extends the work done in #38130 and disallows plain
Python3-style `List` type annotations.
**Test Plan**
This commit extends `TestList.test_no_element_type_annotation` to the
Python3-style type annotation.
Test Plan: Imported from OSS
Reviewed By: gmagogsfm
Differential Revision: D23721514
Pulled By: SplitInfinity
fbshipit-source-id: 48957868286f44ab6d5bf5e1bf97f0a4ebf955df
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/44334
**Summary**
This commit detects and prohibits the case in which `typing.Dict` is
used as an annotation without type arguments (i.e. `typing.Dict[K, V]`).
At present, `typing.Dict` is always assumed to have two arguments, and
when it is used without them, `typing.Dict.__args__` is nonempty and
contains some `typing.TypeVar` instances, which have no JIT type equivalent.
Consequently, trying to convert `typing.Dict` to a JIT type results in
a `c10::DictType` with `nullptr` for its key and value types, which can cause
a segmentation fault.
This is fixed by returning a `DictType` from
`jit.annotations.try_ann_to_type` only if the key and value types are converted
successfully to a JIT type and returning `None` otherwise.
**Test Plan**
This commit adds a unit test to `TestDict` that tests the plain `Dict`
annotations throw an error.
**Fixes**
This commit closes#43530.
Test Plan: Imported from OSS
Reviewed By: gmagogsfm
Differential Revision: D23610766
Pulled By: SplitInfinity
fbshipit-source-id: 036b10eff6e3206e0da3131cfb4997d8189c4fec
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/41371
**Summary**
This commit enables the use of `torch.no_grad()` in a with item of a
with statement within JIT. Note that the use of this context manager as
a decorator is not supported.
**Test Plan**
This commit adds a test case to the existing with statements tests for
`torch.no_grad()`.
**Fixes**
This commit fixes#40259.
Test Plan: Imported from OSS
Reviewed By: gmagogsfm
Differential Revision: D22649519
Pulled By: SplitInfinity
fbshipit-source-id: 7fa675d04835377666dfd0ca4e6bc393dc541ab9
Summary:
fixes https://github.com/pytorch/pytorch/issues/39566
`typing.Final` is a thing since python 3.8, and on python 3.8, `typing_extensions.Final` is an alias of `typing.Final`, therefore, `ann.__module__ == 'typing_extensions'` will become False when using 3.8 and `typing_extensions` is installed.
~~I don't know why the test is skipped, seems like due to historical reason when python 2.7 was still a thing?~~ Edit: I know now, the `Final` for `<3.7` don't have `__origin__`
Pull Request resolved: https://github.com/pytorch/pytorch/pull/39568
Reviewed By: smessmer
Differential Revision: D23043388
Pulled By: malfet
fbshipit-source-id: cc87a9e4e38090d784e9cea630e1c543897a1697
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/42611
**Summary**
This commit modifies the Python frontend to ignore static functions on
Torchscript classes when compiling them. They are currently included
along with methods, which causes the first argument of the
staticfunction to be unconditionally inferred to be of the type of the
class it belongs to (regardless of how it is annotated or whether it is
annotated at all). This can lead to compilation errors depending on
how that argument is used in the body of the function.
Static functions are instead imported and scripted as if they were
standalone functions.
**Test Plan**
This commit augments the unit test for static methods in `test_class_types.py`
to test that static functions can call each other and the class
constructor.
**Fixes**
This commit fixes#39308.
Test Plan: Imported from OSS
Reviewed By: ZolotukhinM
Differential Revision: D22958163
Pulled By: SplitInfinity
fbshipit-source-id: 45c3c372792299e6e5288e1dbb727291e977a2af
Summary:
Raise and assert used to have a hard-coded error message "Exception". User provided error message was ignored. This PR adds support to represent user's error message in TorchScript.
This breaks backward compatibility because now we actually need to script the user's error message, which can potentially contain unscriptable expressions. Such programs can break when scripting, but saved models can still continue to work.
Increased an op count in test_mobile_optimizer.py because now we need aten::format to form the actual exception message.
This is built upon an WIP PR: https://github.com/pytorch/pytorch/pull/34112 by driazati
Pull Request resolved: https://github.com/pytorch/pytorch/pull/41907
Reviewed By: ngimel
Differential Revision: D22778301
Pulled By: gmagogsfm
fbshipit-source-id: 2b94f0db4ae9fe70c4cd03f4048e519ea96323ad
Summary:
* Add EnumType and AnyEnumType as first-class jit type
* Add Enum-typed IValue
* Enhanced aten::eq to support Enum
Supported:
Enum-typed function targuments
using Enum type and comparing them
TODO:
Add PyThon sugared value for Enum
Support getting name/value attrs of enums
Support Enum-typed return values
Support enum values of different types in same Enum class
Support serialization and deserialization
Pull Request resolved: https://github.com/pytorch/pytorch/pull/41390
Reviewed By: eellison
Differential Revision: D22524388
Pulled By: gmagogsfm
fbshipit-source-id: 1627154a64e752d8457cd53270f3d14aea4b1150
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/40270
Original commit changeset: 1227e243ab94
D22082806 (1e03d603c6) broke the model generation of pyper models. We trace the namedtuple as input. To unblock the development of PyPer project, let's revert the diff first.
Sorry about the inconvenience, SplitInfinity
ghstack-source-id: 106217609
Test Plan: buck run dper3/dper3_models/experimental/pytorch/feed:feed_generation_script -- --model_files_dir=/tmp/
Reviewed By: alyssawangqq
Differential Revision: D22132960
fbshipit-source-id: ce9278c8462602a341e231ea890e46f74e743ddf
Summary:
**Summary**
This commit modifies type inference for `nn.Module` instance attributes
such that the type of a `NamedTuple` attribute is inferred correctly and
such that the field names of this `NamedTuple` instance can be used in
scripted methods. At present, the type of this attribute is inferred to be
`Tuple[T, U, ..., V]`, so the field must be referred to by index and
cannot be referred to by name.
**Test Plan**
This commit adds a unit test to test that a field of a `NamedTuple`
attribute can be referred to by name in a scripted method.
**Fixes**
This commit fixes https://github.com/pytorch/pytorch/issues/37668.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/39116
Differential Revision: D22082806
Pulled By: SplitInfinity
fbshipit-source-id: 1227e243ab941376cd5e382fb093751e88dc8846
Summary:
- add call out to python resolver in parseArgsFromDecl, parserReturnFromDecl
- add support in python resolver for nested subexpressions
- wrap python resolver call in exception handling to fall back to c++ path
- add tests for newly resolvable types
- closes https://github.com/pytorch/pytorch/issues/38728
Fixes bug where SourceRange objects did not include the final closing ']' for a subscript expression. E.g. range for 'List[int]' previously included only 'List[int'.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/39269
Differential Revision: D21956402
Pulled By: wconstab
fbshipit-source-id: 5d783260322eb1e04e20bc931a8e9d9179765f13
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/27510
We could delete polyfill typing bc requirements.txt require user to
install typing as a dependency whether in py2 or py3, so those typing
actually not getting used either ways.
Test Plan: Imported from OSS
Differential Revision: D20673393
fbshipit-source-id: ea5276824c6e275c1f991f8c12329040b0058d2b
Summary:
This adds some machinery so that we use Python to resolve types to a value and the corresponding resolution logic in `annotations.py` instead of using the string.
This PR also `slowTests` a random test since it was taking > 1 min whereas all the other tests take < 10 seconds.
Fixes#31864Fixes#31950
](https://our.intern.facebook.com/intern/diff/20144407/)
Pull Request resolved: https://github.com/pytorch/pytorch/pull/29623
Pulled By: driazati
Differential Revision: D20144407
fbshipit-source-id: ef3699f6b86039d8b4646ffc42c21bd1132d1681
Summary:
Uses new overload mechanism for rnns, making it so that python & torchscript go through the same path and using an API that is in line with the one specified
in https://docs.python.org/3/library/typing.html#typing.overload
This brings the TorchScriptable rnns closer to the base implementation; unifying them should be done in a follow up PR but there are still a few limitations that make it difficult to do so.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/29614
Differential Revision: D18486982
Pulled By: eellison
fbshipit-source-id: aaaea66a4a7f12d2e46199ca254f9e8f7475500e
Summary:
Uses new overload mechanism for rnns, making it so that python & torchscript go through the same path and using an API that is in line with the one specified
in https://docs.python.org/3/library/typing.html#typing.overload
This brings the TorchScriptable rnns closer to the base implementation; unifying them should be done in a follow up PR but there are still a few limitations that make it difficult to do so.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/29614
Differential Revision: D18458751
Pulled By: eellison
fbshipit-source-id: 07c71838f21cb5425e8d6dbd4a512f774c8c2970
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/28988
Make ModuleList, Sequential, ModuleDict go through the same pathway as other modules, cleaning up a bunch of code and allowing them to define custom forwards and other methods.
EDIT: Previously, we would ignore an nn.Sequential attribute if it was not in `__constants__` ("did you forget to add it to Constants"). This PR scripts it even if it is not in `__constants__`. Is that what we want?
Test Plan: Imported from OSS
Differential Revision: D18402821
Pulled By: eellison
fbshipit-source-id: dd4f28fb0df0d1ba4ad1b3bc34ba141959a433f7
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/29610
`DeprecationWarning` is intended for developers (and so is ignored in
certain circumstances). `FutureWarning` is the user-facing deprecation
warning. This fixes fbcode failures.
Test Plan: Imported from OSS
Differential Revision: D18446393
Pulled By: suo
fbshipit-source-id: ded11a007f0a62132a9839b733157a97cf9006e9
Summary:
Type objects in python have an attribute `__abstractmethods__` that throws when it is accessed, so we were failing with an AttributeError whenever a type was used in TorchScript.
This pr prevents that error from happening. We can't just throw when a type is used because it could be used to access a static method: https://github.com/pytorch/pytorch/pull/27163
Pull Request resolved: https://github.com/pytorch/pytorch/pull/28053
Differential Revision: D18332347
Pulled By: eellison
fbshipit-source-id: 9c7f2220f92674ad4d903621d9762cecc566ab0d
