Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/70960
This patch uses some bytecode introspection logic to see if a boolean is being used as an assert condition and if so, it records the assert in the fx graph and allows the trace to continue.
Test Plan: Imported from OSS
Reviewed By: jamesr66a
Differential Revision: D33570397
Pulled By: zdevito
fbshipit-source-id: 99d26cede8fe42c96d4032d9353c1ede7eb3d969
(cherry picked from commit 30d002da25)
Summary:
Previously resulted in `AttributeError: module 'operator' has no attribute 'and'`
and/or are python keywords, so they are renamed to `operator.and_` and `operator.or_`
Pull Request resolved: https://github.com/pytorch/pytorch/pull/65196
Reviewed By: Chillee
Differential Revision: D31020336
Pulled By: jansel
fbshipit-source-id: 51d888151fe78c0c1197ecaf161976b219c59694
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/64842
Change the `__torch_function__` to follow best guidelines of using classmethods.
I am not sure how to handle the case where multiple tracer objects are given as input but given that before we were getting an arbitrary tracer from there based on the "self" that was arbitrarily chosen by the torch_function caller, the new implementation is not worst?
Let me know what you think!
Test Plan: Imported from OSS
Reviewed By: heitorschueroff
Differential Revision: D30991423
Pulled By: albanD
fbshipit-source-id: d28940df230b543952b278a0eb2d61cf7ae123ce
Summary:
Fixes https://github.com/pytorch/pytorch/issues/61733
Allow FX tracer to trace control flow (if/while) statements when parameter shapes are in the condition.
If the user specifies the new "param_shapes_constant" option when constructing a tracer, the model's parameter shape attribute will be evaluated and the resulting constant will be emitted into the IR during tracing.
Also added a new test
`
python test/fx/test_fx_param_shape_control_flow.py
`
The test also performs a somewhat whitebox style testing to check the generated Python code from the IR.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/61820
Reviewed By: bdhirsh
Differential Revision: D29969299
Pulled By: jerryzhenleicai
fbshipit-source-id: 99aae824bdfec880be69258de7ead5c8cd59eddc
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/61780
These changes would allow objects to control how they are handled when they are an argument to a torch.fx call_module node from within their source. Previously, we have been using a custom Tracer with an overridden create_arg() method and branching based on class name to handle args that are unusual (data classes, etc).
Reviewed By: suo, houseroad
Differential Revision: D27976120
fbshipit-source-id: 0c5249c5f8398368ca0fbec0ad8a07ccf99b7da4
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/45193
This change makes it possible to subclass the tracer to add additional
behavior when you know something about the shape of the Proxy objects,
by overriding the defaults for how the tracer tries to make something iterable,
looks for keys for **kwargs, or tries to convert to a boolean.
An example test shows how this can be used to tag inputs with shapes.
It can also be used combined with create_node to do type propagation during
tracing to fullfil requests like iter.
Test Plan: Imported from OSS
Reviewed By: jamesr66a
Differential Revision: D24258993
Pulled By: zdevito
fbshipit-source-id: 6ece686bec292e51707bbc7860a1003d0c1321e8
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/44566
The Delegate objects were confusing. They were suppose to be a way to
configure how tracing works, but in some cases they appeared necessary
for consturcting graphs, which was not true. This makes the organization
clearer by removing Delgate and moving its functionality into a Tracer class,
similar to how pickle has a Pickler class.
Test Plan: Imported from OSS
Reviewed By: jamesr66a
Differential Revision: D23683177
Pulled By: zdevito
fbshipit-source-id: 7605a34e65dfac9a487c0bada39a23ca1327ab00
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/43083
This adds type annotations to all classes, arguments, and returns
for fx. This should make it easier to understand the code, and
encourage users of the library to also write typed code.
Test Plan: Imported from OSS
Reviewed By: ezyang
Differential Revision: D23145853
Pulled By: zdevito
fbshipit-source-id: 648d91df3f9620578c1c51408003cd5152e34514
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/43082
Fixes all present errors in mypy. Does not try to add annotations everywhere.
Test Plan: Imported from OSS
Reviewed By: jamesr66a
Differential Revision: D23145854
Pulled By: zdevito
fbshipit-source-id: 18e483ed605e89ed8125971e84da1a83128765b7
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/42991
Have Node both be a record of the operator in the graph, and the
way we _build_ the graph made it difficult to keep the IR datastructure
separate from the proxying logic in the build.
Among other issues this means that typos when using nodes would add
things to the graph:
```
for node in graph.nodes:
node.grph # does not error, returns an node.Attribute object!
```
This separates the builder into a Proxy object. Graph/Node no longer
need to understand `delegate` objects since they are now just pure IR.
This separates the `symbolic_trace` (proxy.py/symbolic_trace.py) from
the IR (node.py, graph.py).
This also allows us to add `create_arg` to the delegate object,
allowing the customization of how aggregate arguments are handled
when converting to a graph.
Test Plan: Imported from OSS
Reviewed By: jamesr66a
Differential Revision: D23099786
Pulled By: zdevito
fbshipit-source-id: 6f207a8c237e5eb2f326b63b0d702c3ebcb254e4
