OSSForks/pytorch
1
0
Fork 0
mirror of https://github.com/zebrajr/pytorch.git synced 2025-12-07 00:21:07 +01:00
Code Issues Actions 38 Packages Projects Releases Wiki Activity
62,303 Commits 8,304 Branches 150 Tags 1.5 GiB
00ee38c661
Commit Graph

16 Commits

Author SHA1 Message Date
BowenBao
00ee38c661 [ONNX] Export module as function (#105618) 
Introduce `Modularize` pass that analyzes the flat `fx.GraphModule` and creates nested
layers of sub `fx.GraphModule`s along with the `call_module` fx nodes that invokes them.
The analysis is done on the meta data "nn_module_stack", which captures the `nn.Module`
each flat `fx.Node` belongs to.

`FxOnnxInterpreter` is updated to support `call_module`. The related sub module linked
by `node.target` is exported as an ONNX model local function. The `call_module` node itself
is exported as an ONNX node, associated with the ONNX model local function by op_type.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/105618
Approved by: https://github.com/justinchuby
 2023-07-25 03:28:31 +00:00
BowenBao
980589b04d [ONNX] Suppress ORT warnings in unit tests (#105624) 
As title, these warnings are too noisy and made CI test logs hard to read.

Pull Request resolved: https://github.com/pytorch/pytorch/pull/105624
Approved by: https://github.com/justinchuby
 2023-07-20 07:21:21 +00:00
Thiago Crepaldi
82091d666c [ONNX] Refactor Input/Output Adapter (#100490) 
This PR refactors how InputAdapter and OutputAdapter is used throughout the exporter.

During refactoring, API issues with passes (torch.onnx._internal.fx._pass.Transform) were identified and should be tackled on another API. In short, some passes can modify the input/output of the model and the input/output adapter must be in sync with such change, otherwise, the adapters will not reflect the actual model input/output. The first instance of this issue was with `ReplaceGetAttrWithPlaceholder` pass that adds new inputs to the model. In order to work this around, a new input adapt step to append new inputs (generated by the pass) was introduced. That resulted in the number of inputs of the ONNX model to mismatch the numer of inputs of the pytorch model, though.

Follow up on https://github.com/pytorch/pytorch/pull/98421
Pull Request resolved: https://github.com/pytorch/pytorch/pull/100490
Approved by: https://github.com/BowenBao
 2023-05-06 16:01:49 +00:00
AllenTiTaiWang
843ead134c [ONNX] Add supported ops into test_fx_op_consistency - 1st batch (#100265) 
Pull Request resolved: https://github.com/pytorch/pytorch/pull/100265
Approved by: https://github.com/justinchuby
 2023-05-03 14:42:25 +00:00
AllenTiTaiWang
d23dbfff60 [ONNX] Add RemoveConstantInputStep to adapt torch inputs to ONNX inputs (#100252) 
Pull Request resolved: https://github.com/pytorch/pytorch/pull/100252
Approved by: https://github.com/BowenBao
 2023-05-03 01:50:47 +00:00
BowenBao
0bcb9dac4f [ONNX] Non-global diagnostic context (#100219) 
Summary:

* `dynamo_export`, and everything within now access diagnostic context through a maintained local
  variable, instead of global.
* Refactored `diagnose_call` decorator to require local diagnostic context, instead of accessing global.
* Modified `test_fx_to_onnx_*.py` tests to only log '*.sarif' logs when `verbose=True`.
* Temporarily removed diagnostics for `OnnxFunction`, as they don't have access to diagnostic context
  anymore. These diagnostics will be the responsibility of `onnxscript`, and they will return once
  diagnostics system is integrated there.

Pull Request resolved: https://github.com/pytorch/pytorch/pull/100219
Approved by: https://github.com/justinchuby
 2023-05-01 19:58:53 +00:00
AllenTiTaiWang
cba07ffe0c [ONNX] Add xfail into subtests of op consistency and retire fixme (#100173) 
Pull Request resolved: https://github.com/pytorch/pytorch/pull/100173
Approved by: https://github.com/justinchuby
 2023-04-28 02:25:59 +00:00
AllenTiTaiWang
1dba53cbab [ONNX] Refactor test_op_consistenct.py and test_fx_op_consistency.py (#100172) 
## Summary
<!--
copilot:summary
-->
### <samp>🤖 Generated by Copilot at 9255aa3</samp>

This pull request refactors the ONNX operator testing code to use a common module `./test/onnx/onnx_test_common.py` that defines constants, types, classes, and functions for testing ONNX operators. This improves the code quality, readability, and maintainability.

## Walkthrough
<!--
copilot:walkthrough
-->
### <samp>🤖 Generated by Copilot at 9255aa3</samp>

*  Refactor the common code for testing ONNX operators from different files into `./test/onnx/onnx_test_common.py` ([link](https://github.com/pytorch/pytorch/pull/100172/files?diff=unified&w=0#diff-1b38383dc1a0228a835d83bb7c4ba2d0c1bcd41297be5c6336572c525846166eL10-R24), [link](https://github.com/pytorch/pytorch/pull/100172/files?diff=unified&w=0#diff-1b38383dc1a0228a835d83bb7c4ba2d0c1bcd41297be5c6336572c525846166eR33), [link](https://github.com/pytorch/pytorch/pull/100172/files?diff=unified&w=0#diff-1b38383dc1a0228a835d83bb7c4ba2d0c1bcd41297be5c6336572c525846166eR367-R623))
* Remove the unused and duplicated imports, constants, types, and classes for testing ONNX operators from `./test/onnx/test_fx_op_consistency.py` and `./test/onnx/test_op_consistency.py` ([link](https://github.com/pytorch/pytorch/pull/100172/files?diff=unified&w=0#diff-db2f78a51511bb172cbfde1b2f68272b8b33049abe2571cded27bcd0f3ae5fa4L28-R29), [link](https://github.com/pytorch/pytorch/pull/100172/files?diff=unified&w=0#diff-db2f78a51511bb172cbfde1b2f68272b8b33049abe2571cded27bcd0f3ae5fa4L43-R42), [link](https://github.com/pytorch/pytorch/pull/100172/files?diff=unified&w=0#diff-e968c9cb6fc6631cab526cb3a9fe66358c4c6e757e2a223a224b976471bcb753L28-R29), [link](https://github.com/pytorch/pytorch/pull/100172/files?diff=unified&w=0#diff-e968c9cb6fc6631cab526cb3a9fe66358c4c6e757e2a223a224b976471bcb753L43-R44))
* Import the `unittest`, `opinfo_core`, and `onnx_test_common` modules and the `fixme`, `skip`, and `xfail` functions in `./test/onnx/test_fx_op_consistency.py` and `./test/onnx/test_op_consistency.py` ( [link](https://github.com/pytorch/pytorch/pull/100172/files?diff=unified&w=0#diff-db2f78a51511bb172cbfde1b2f68272b8b33049abe2571cded27bcd0f3ae5fa4R36), [link](https://github.com/pytorch/pytorch/pull/100172/files?diff=unified&w=0#diff-e968c9cb6fc6631cab526cb3a9fe66358c4c6e757e2a223a224b976471bcb753L37-R37))
* Update the references to the constants, types, functions, and classes for testing ONNX operators in `./test/onnx/test_fx_op_consistency.py` and `./test/onnx/test_op_consistency.py` to use the definitions from `./test/onnx/onnx_test_common.py` ([link](https://github.com/pytorch/pytorch/pull/100172/files?diff=unified&w=0#diff-db2f78a51511bb172cbfde1b2f68272b8b33049abe2571cded27bcd0f3ae5fa4L324-R80), [link](https://github.com/pytorch/pytorch/pull/100172/files?diff=unified&w=0#diff-db2f78a51511bb172cbfde1b2f68272b8b33049abe2571cded27bcd0f3ae5fa4L389-R135), [link](https://github.com/pytorch/pytorch/pull/100172/files?diff=unified&w=0#diff-db2f78a51511bb172cbfde1b2f68272b8b33049abe2571cded27bcd0f3ae5fa4L405-R151), [link](https://github.com/pytorch/pytorch/pull/100172/files?diff=unified&w=0#diff-db2f78a51511bb172cbfde1b2f68272b8b33049abe2571cded27bcd0f3ae5fa4L455-R204), [link](https://github.com/pytorch/pytorch/pull/100172/files?diff=unified&w=0#diff-e968c9cb6fc6631cab526cb3a9fe66358c4c6e757e2a223a224b976471bcb753L333-R107), [link](https://github.com/pytorch/pytorch/pull/100172/files?diff=unified&w=0#diff-e968c9cb6fc6631cab526cb3a9fe66358c4c6e757e2a223a224b976471bcb753L434-R183), [link](https://github.com/pytorch/pytorch/pull/100172/files?diff=unified&w=0#diff-e968c9cb6fc6631cab526cb3a9fe66358c4c6e757e2a223a224b976471bcb753L448-R197), [link](https://github.com/pytorch/pytorch/pull/100172/files?diff=unified&w=0#diff-e968c9cb6fc6631cab526cb3a9fe66358c4c6e757e2a223a224b976471bcb753L494-R246))
Pull Request resolved: https://github.com/pytorch/pytorch/pull/100172
Approved by: https://github.com/justinchuby
 2023-04-27 21:32:04 +00:00
AllenTiTaiWang
bd7191111f [ONNX] Add additional_test_kwargs into test_fx_to_onnx_with_onnxruntime.py (#99434) 
1. Expand additional_test_inputs to include kwargs
2. Revisit and update tests status by adding ops
3. Disabling dtype -1 assignment avoids potential bugs
4. Expand input/output to accept buit-in type, but they are not dynamically captured by dynamo.export right now, and they would be added as constant input to op.targets.
5. Move run_test_with_fx_to_onnx_exporter_and_onnx_runtime to onnx_test_common.py

<!--
copilot:all
-->
### <samp>🤖 Generated by Copilot at 3c03579</samp>

### Summary
🛠️🧪🚀

<!--
1.  🛠️ for updating the `filter_incompatible_and_dtype_convert_kwargs` function
2.  🧪 for updating the test function and test cases
3.  🚀 for adding support for new operators and scalar types
-->
This pull request improves the ONNX export support for scalar types and some ATen operators in PyTorch. It updates the test framework, the input and output adapters, the function dispatcher and the ONNX script generator to handle these cases. It also fixes or removes some failing or outdated tests.

> _We defy the limits of the ONNX script_
> _We export the models with scalar and copy_
> _We filter and convert the kwargs of dtype_
> _We run the tests with FX and docstring_

### Walkthrough
*  Update the `_InputArgsType` type annotation and the `_run_test_with_fx_to_onnx_exporter_and_onnx_runtime` function signature and docstring to handle int, float and bool inputs for some ONNX operators ([link](https://github.com/pytorch/pytorch/pull/99434/files?diff=unified&w=0#diff-c8fa56eefd7f98fb4f9739d57df57f02ede77e28528133736010a6d06651ebcbL44-R46), [link](https://github.com/pytorch/pytorch/pull/99434/files?diff=unified&w=0#diff-c8fa56eefd7f98fb4f9739d57df57f02ede77e28528133736010a6d06651ebcbL144-R157), [link](https://github.com/pytorch/pytorch/pull/99434/files?diff=unified&w=0#diff-c8fa56eefd7f98fb4f9739d57df57f02ede77e28528133736010a6d06651ebcbL155-R164), [link](https://github.com/pytorch/pytorch/pull/99434/files?diff=unified&w=0#diff-c8fa56eefd7f98fb4f9739d57df57f02ede77e28528133736010a6d06651ebcbL162-R172), [link](https://github.com/pytorch/pytorch/pull/99434/files?diff=unified&w=0#diff-c8fa56eefd7f98fb4f9739d57df57f02ede77e28528133736010a6d06651ebcbL201-R224), [link](https://github.com/pytorch/pytorch/pull/99434/files?diff=unified&w=0#diff-0795f54fd1f38cfbf2c4a863a4efc9f40f2ea020a2b1612605c361b8d8d35862L197-R199), [link](https://github.com/pytorch/pytorch/pull/99434/files?diff=unified&w=0#diff-0795f54fd1f38cfbf2c4a863a4efc9f40f2ea020a2b1612605c361b8d8d35862L291-R293))
* Update the `filter_incompatible_and_dtype_convert_kwargs` function to omit the `dtype` argument if it is None ([link](https://github.com/pytorch/pytorch/pull/99434/files?diff=unified&w=0#diff-cabc3e58713d6fe7ab764ade4f2692f6753402322a7b542397cad16fcc72cf4bL203-R205))
* Update the test cases in `test_fx_to_onnx_with_onnxruntime.py` to use the `input_kwargs` parameter as a mapping, to fix the format of the `additional_test_inputs` parameter, and to add or remove `xfail`, `skip_dynamic_fx_test` and `skip_min_ort_version` decorators as needed ([link](https://github.com/pytorch/pytorch/pull/99434/files?diff=unified&w=0#diff-c8fa56eefd7f98fb4f9739d57df57f02ede77e28528133736010a6d06651ebcbL320-R336), [link](https://github.com/pytorch/pytorch/pull/99434/files?diff=unified&w=0#diff-c8fa56eefd7f98fb4f9739d57df57f02ede77e28528133736010a6d06651ebcbL330-R353), [link](https://github.com/pytorch/pytorch/pull/99434/files?diff=unified&w=0#diff-c8fa56eefd7f98fb4f9739d57df57f02ede77e28528133736010a6d06651ebcbL357-R380), [link](https://github.com/pytorch/pytorch/pull/99434/files?diff=unified&w=0#diff-c8fa56eefd7f98fb4f9739d57df57f02ede77e28528133736010a6d06651ebcbL452-L470), [link](https://github.com/pytorch/pytorch/pull/99434/files?diff=unified&w=0#diff-c8fa56eefd7f98fb4f9739d57df57f02ede77e28528133736010a6d06651ebcbL488-R486), [link](https://github.com/pytorch/pytorch/pull/99434/files?diff=unified&w=0#diff-c8fa56eefd7f98fb4f9739d57df57f02ede77e28528133736010a6d06651ebcbL509-R510), [link](https://github.com/pytorch/pytorch/pull/99434/files?diff=unified&w=0#diff-c8fa56eefd7f98fb4f9739d57df57f02ede77e28528133736010a6d06651ebcbR543), [link](https://github.com/pytorch/pytorch/pull/99434/files?diff=unified&w=0#diff-c8fa56eefd7f98fb4f9739d57df57f02ede77e28528133736010a6d06651ebcbL559-R565), [link](https://github.com/pytorch/pytorch/pull/99434/files?diff=unified&w=0#diff-c8fa56eefd7f98fb4f9739d57df57f02ede77e28528133736010a6d06651ebcbL578-R580), [link](https://github.com/pytorch/pytorch/pull/99434/files?diff=unified&w=0#diff-c8fa56eefd7f98fb4f9739d57df57f02ede77e28528133736010a6d06651ebcbL597-R599), [link](https://github.com/pytorch/pytorch/pull/99434/files?diff=unified&w=0#diff-c8fa56eefd7f98fb4f9739d57df57f02ede77e28528133736010a6d06651ebcbL611-R620), [link](https://github.com/pytorch/pytorch/pull/99434/files?diff=unified&w=0#diff-c8fa56eefd7f98fb4f9739d57df57f02ede77e28528133736010a6d06651ebcbL636-R636), [link](https://github.com/pytorch/pytorch/pull/99434/files?diff=unified&w=0#diff-c8fa56eefd7f98fb4f9739d57df57f02ede77e28528133736010a6d06651ebcbL656-R659), [link](https://github.com/pytorch/pytorch/pull/99434/files?diff=unified&w=0#diff-c8fa56eefd7f98fb4f9739d57df57f02ede77e28528133736010a6d06651ebcbL672-R675), [link](https://github.com/pytorch/pytorch/pull/99434/files?diff=unified&w=0#diff-c8fa56eefd7f98fb4f9739d57df57f02ede77e28528133736010a6d06651ebcbL691-R698), [link](https://github.com/pytorch/pytorch/pull/99434/files?diff=unified&w=0#diff-c8fa56eefd7f98fb4f9739d57df57f02ede77e28528133736010a6d06651ebcbL709-R714), [link](https://github.com/pytorch/pytorch/pull/99434/files?diff=unified&w=0#diff-c8fa56eefd7f98fb4f9739d57df57f02ede77e28528133736010a6d06651ebcbL732-R730), [link](https://github.com/pytorch/pytorch/pull/99434/files?diff=unified&w=0#diff-c8fa56eefd7f98fb4f9739d57df57f02ede77e28528133736010a6d06651ebcbL752-R750), [link](https://github.com/pytorch/pytorch/pull/99434/files?diff=unified&w=0#diff-c8fa56eefd7f98fb4f9739d57df57f02ede77e28528133736010a6d06651ebcbL773-R771), [link](https://github.com/pytorch/pytorch/pull/99434/files?diff=unified&w=0#diff-c8fa56eefd7f98fb4f9739d57df57f02ede77e28528133736010a6d06651ebcbR797-R803), [link](https://github.com/pytorch/pytorch/pull/99434/files?diff=unified&w=0#diff-c8fa56eefd7f98fb4f9739d57df57f02ede77e28528133736010a6d06651ebcbL807-R816))

Pull Request resolved: https://github.com/pytorch/pytorch/pull/99434
Approved by: https://github.com/justinchuby
 2023-04-22 04:03:50 +00:00
Aaron Gokaslan
3d82d8d0ed [BE] Enable more flake8-comprehensions checks (#94601) 
I applied some flake8 fixes and enabled checking for them in the linter. I also enabled some checks for my previous comprehensions PR.

This is a follow up to #94323 where I enable the flake8 checkers for the fixes I made and fix a few more of them.

Pull Request resolved: https://github.com/pytorch/pytorch/pull/94601
Approved by: https://github.com/ezyang
 2023-02-10 23:40:29 +00:00
BowenBao
79f9672249 [ONNX] Use VerificationOptions to wrap option arguments (#89807) 
Pull Request resolved: https://github.com/pytorch/pytorch/pull/89807
Approved by: https://github.com/justinchuby, https://github.com/titaiwangms
 2022-12-09 23:49:51 +00:00
BowenBao
500fd65531 [ONNX] Create common ExportTestCase base class (#88145) 
Refactor out a common base class `ExportTestCase`, for common things in `setUp`.

Pull Request resolved: https://github.com/pytorch/pytorch/pull/88145
Approved by: https://github.com/justinchuby, https://github.com/abock, https://github.com/AllenTiTaiWang
 2022-11-10 21:51:59 +00:00
Justin Chu
2fa8142cf9 [ONNX] Rename constants for clarity (#84645) 
Rename constants to make them more clear. Fix styles to upper case.

Removed `onnx_stable_opsets` because it can be computed from `ONNX_MIN_OPSET` and `ONNX_MAX_OPSET`.

Fixes #84643

Pull Request resolved: https://github.com/pytorch/pytorch/pull/84645
Approved by: https://github.com/BowenBao
 2022-09-09 01:22:14 +00:00
titaiwang
5bceaadb70 [ONNX] Add script/trace different flatten and move optional type tests to runtime (#83184) 
fix #78119

Why:
As in onnx tests verification code, we used to only consider tracing output, which ignores None type, this PR enables runtime test to keep None type in torch in script mode.

1. Move Optional Type tests from no runtime to runtime, as it's supported by ONNXRUNTIME.
2. Add ignoreNone flag for output comparison of internal tests
Pull Request resolved: https://github.com/pytorch/pytorch/pull/83184
Approved by: https://github.com/justinchuby, https://github.com/BowenBao
 2022-08-30 18:23:24 +00:00
Li-Huai (Allan) Lin
d9a7e93aaf [ONNX] Add dtype check in onnx verification (#79263) 
Currently we don't have a dtype check in verifying the consistency between PyTorch and ONNX outputs. As a result, some of dtype inconsistencies were found and reported: #77842 #77845

This is a POC.

Failed workflows:
- [linux-xenial-py3.7-clang7-onnx / test (default, 2, 2, linux.2xlarge)]
  - inconsistent shape
    - TestONNXRuntime_opset10.test_all (#79371)
    - TestONNXRuntime_opset10.test_any (#79371)
    - TestONNXRuntime_opset10.test_argmin_argmax (#79503)
    - TestONNXRuntime_opset10.test_hardshrink (#79695)
    - TestONNXRuntime_opset10.test_linalg_norm (#79506)
    - TestONNXRuntime_opset10.test_linalg_vector_norm (#79506)
    - TestONNXRuntime_opset10.test_prelu_scalar (#79846)
    - TestONNXRuntime_opset10.test_softshrink (#79695)
    - TestONNXRuntime_opset10.test_sum_empty_tensor (skipped)
    - TestONNXRuntime_opset10.test_tolist (skipped)
  - inconsistent dtype
    - test_arithmetic_prim_bool (skipped)
    - test_arithmeticOps_with_low_precision (skipped)
    - test_arithmetic_prim_float (skipped)
    - test_logical_and (#79339)
    - test_logical_or (#79339)
    - test_logical_xor (#79339)
    - test_pow (skipped)
    - test_primitive_input_floating (skipped)
    - test_quantize_per_tensor (#79690)
    - test_quantized_adaptive_avg_pool2d (#79690)
    - test_quantized_arithmetic (#79690)
    - test_quantized_arithmetic_qfunctional (#79690)
    - test_quantized_conv2d (#79690)
    - test_quantized_conv2d_relu (#79690)
    - test_quantized_flatten (#79690)
    - test_quantized_hardsigmoid (#79690)
    - test_quantized_hardswish (#79690)
    - test_quantized_linear (#79690)
    - test_quantized_sigmoid (#79690)
    - test_item (skipped)
    - test_full_like_value (skipped)
    - TestONNXRuntime_opset7.test_div_rounding_mode (skipped)
    - TestONNXRuntime_opset8.test_div_rounding_mode (skipped)
    - TestONNXRuntime_opset9.test_div_rounding_mode (skipped)
    - TestONNXRuntime_opset9_IRv4.test_div_rounding_mode (skipped)
    - test_outer (skipped)
    - test_symbolic_shape_inference_arange_2 (skipped)
Pull Request resolved: https://github.com/pytorch/pytorch/pull/79263
Approved by: https://github.com/justinchuby, https://github.com/BowenBao
 2022-08-10 07:14:12 +00:00
Justin Chu
773d80747c [ONNX] Clean up unit tests, rename files and improve import style (#81141) 
- Rename `test_pytorch_common` -> `pytorch_test_common`, `test_onnx_common` -> `onnx_test_common`, removing the test_ prefix to show that the files are not test cases
- Remove import * in `test_pytorch_common` and adjust to import from `testing._internal.common_utils` (where functions are actually defined) instead
- Import modules only in `test_pytorch_onnx_onnxruntime` (too many to handle in a single PR in other tests) (The skips are exceptions)
Pull Request resolved: https://github.com/pytorch/pytorch/pull/81141
Approved by: https://github.com/BowenBao
 2022-07-12 00:00:49 +00:00