Do not import transformer_engine on import
#3056
Merged
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
muellerzr
reviewed
Aug 28, 2024
tests/test_imports.py
Outdated
Comment on lines
87
to
100
| @require_transformer_engine | ||
| class LazyImportTester(TempDirTestCase): | ||
| """ | ||
| Test suite which checks if specific packages are lazy-loaded. | ||
|
|
||
| Eager-import will trigger circular import in some case, | ||
| e.g. in huggingface/accelerate#3056. | ||
| """ | ||
|
|
||
| # @require_transformer_engine | ||
| def test_te_import(self): | ||
| output = run_import_time("import accelerate, accelerate.utils.transformer_engine") | ||
|
|
||
| self.assertFalse(' transformer_engine' in output, '`transformer_engine` should not be imported on import') |
There was a problem hiding this comment.
We have no CI for transformerengine, so not sure this test is actually valuable imo
There was a problem hiding this comment.
Thanks for your review! I added this test sort of as a self-contained doc. But it's indeed not much valuable if not in CI. I've just add a one line comment in the other file, and I can undo the test if you think that's better.
|
The docs for this PR live here. All of your documentation changes will be reflected on that endpoint. The docs are available until 30 days after the last update. |
muellerzr
approved these changes
Aug 28, 2024
Co-authored-by: Zach Mueller <muellerzr@gmail.com>
|
@oraluben for the failing quality test, do |
|
Thanks for helping landing this! |
pytorchmergebot
pushed a commit
to pytorch/pytorch
that referenced
this pull request
Aug 31, 2024
Currently, if installed, `onnxruntime` will be imported when importing `torch._inductor` (which will be imported by some other library, e.g. transformer-engine): ``` /mnt/c.py(53)<module>() -> from torch._inductor.utils import maybe_profile /usr/local/lib/python3.10/site-packages/torch/_inductor/utils.py(49)<module>() -> import torch._export /usr/local/lib/python3.10/site-packages/torch/_export/__init__.py(25)<module>() -> import torch._dynamo /usr/local/lib/python3.10/site-packages/torch/_dynamo/__init__.py(2)<module>() -> from . import convert_frame, eval_frame, resume_execution /usr/local/lib/python3.10/site-packages/torch/_dynamo/convert_frame.py(48)<module>() -> from . import config, exc, trace_rules /usr/local/lib/python3.10/site-packages/torch/_dynamo/trace_rules.py(52)<module>() -> from .variables import ( /usr/local/lib/python3.10/site-packages/torch/_dynamo/variables/__init__.py(38)<module>() -> from .higher_order_ops import ( /usr/local/lib/python3.10/site-packages/torch/_dynamo/variables/higher_order_ops.py(14)<module>() -> import torch.onnx.operators /usr/local/lib/python3.10/site-packages/torch/onnx/__init__.py(62)<module>() -> from ._internal.onnxruntime import ( /usr/local/lib/python3.10/site-packages/torch/onnx/_internal/onnxruntime.py(37)<module>() -> import onnxruntime # type: ignore[import] ``` This issue breaks generated triton kernel because it imported torch, and unexpected runtime libraries as well. I've also added a test for this specific case under `test/onnx`, perhaps we should add more somewhere else? Related issue: huggingface/accelerate#3056 Pull Request resolved: #134662 Approved by: https://github.com/justinchuby
tolleybot
pushed a commit
to tolleybot/pytorch
that referenced
this pull request
Sep 14, 2024
Currently, if installed, `onnxruntime` will be imported when importing `torch._inductor` (which will be imported by some other library, e.g. transformer-engine): ``` /mnt/c.py(53)<module>() -> from torch._inductor.utils import maybe_profile /usr/local/lib/python3.10/site-packages/torch/_inductor/utils.py(49)<module>() -> import torch._export /usr/local/lib/python3.10/site-packages/torch/_export/__init__.py(25)<module>() -> import torch._dynamo /usr/local/lib/python3.10/site-packages/torch/_dynamo/__init__.py(2)<module>() -> from . import convert_frame, eval_frame, resume_execution /usr/local/lib/python3.10/site-packages/torch/_dynamo/convert_frame.py(48)<module>() -> from . import config, exc, trace_rules /usr/local/lib/python3.10/site-packages/torch/_dynamo/trace_rules.py(52)<module>() -> from .variables import ( /usr/local/lib/python3.10/site-packages/torch/_dynamo/variables/__init__.py(38)<module>() -> from .higher_order_ops import ( /usr/local/lib/python3.10/site-packages/torch/_dynamo/variables/higher_order_ops.py(14)<module>() -> import torch.onnx.operators /usr/local/lib/python3.10/site-packages/torch/onnx/__init__.py(62)<module>() -> from ._internal.onnxruntime import ( /usr/local/lib/python3.10/site-packages/torch/onnx/_internal/onnxruntime.py(37)<module>() -> import onnxruntime # type: ignore[import] ``` This issue breaks generated triton kernel because it imported torch, and unexpected runtime libraries as well. I've also added a test for this specific case under `test/onnx`, perhaps we should add more somewhere else? Related issue: huggingface/accelerate#3056 Pull Request resolved: pytorch#134662 Approved by: https://github.com/justinchuby
Chao1Han
pushed a commit
to Chao1Han/pytorch
that referenced
this pull request
Sep 20, 2024
Currently, if installed, `onnxruntime` will be imported when importing `torch._inductor` (which will be imported by some other library, e.g. transformer-engine): ``` /mnt/c.py(53)<module>() -> from torch._inductor.utils import maybe_profile /usr/local/lib/python3.10/site-packages/torch/_inductor/utils.py(49)<module>() -> import torch._export /usr/local/lib/python3.10/site-packages/torch/_export/__init__.py(25)<module>() -> import torch._dynamo /usr/local/lib/python3.10/site-packages/torch/_dynamo/__init__.py(2)<module>() -> from . import convert_frame, eval_frame, resume_execution /usr/local/lib/python3.10/site-packages/torch/_dynamo/convert_frame.py(48)<module>() -> from . import config, exc, trace_rules /usr/local/lib/python3.10/site-packages/torch/_dynamo/trace_rules.py(52)<module>() -> from .variables import ( /usr/local/lib/python3.10/site-packages/torch/_dynamo/variables/__init__.py(38)<module>() -> from .higher_order_ops import ( /usr/local/lib/python3.10/site-packages/torch/_dynamo/variables/higher_order_ops.py(14)<module>() -> import torch.onnx.operators /usr/local/lib/python3.10/site-packages/torch/onnx/__init__.py(62)<module>() -> from ._internal.onnxruntime import ( /usr/local/lib/python3.10/site-packages/torch/onnx/_internal/onnxruntime.py(37)<module>() -> import onnxruntime # type: ignore[import] ``` This issue breaks generated triton kernel because it imported torch, and unexpected runtime libraries as well. I've also added a test for this specific case under `test/onnx`, perhaps we should add more somewhere else? Related issue: huggingface/accelerate#3056 Pull Request resolved: pytorch#134662 Approved by: https://github.com/justinchuby
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
What does this PR do?
Fix a circular import when multiple package involves, including
deepspeed,apex,transformer_engine,onnx.A simple repro can be built based on ngc:
full `pip list`
Before submitting
Pull Request section?
to it if that's the case.
documentation guidelines, and
here are tips on formatting docstrings.
Who can review?
Anyone in the community is free to review the PR once the tests have passed. Feel free to tag
members/contributors who may be interested in your PR.
Maybe @muellerzr ?