pytorch - 💡(How to fix) Fix DISABLED test_modes_return_notimplemented (__main__.TestTorchFunctionMode) [1 comments, 1 participants]

Traceback (most recent call last): File "/var/lib/jenkins/workspace/test/test_overrides.py", line 1479, in test_modes_return_notimplemented class MyMode(TorchFunctionMode): File "/var/lib/jenkins/workspace/test/test_overrides.py", line 1483, in torch_dynamo_resume_in_test_modes_return_notimplemented_at_1479 x = SubTensor() File "/opt/conda/envs/py_3.10/lib/python3.10/site-packages/torch/_dynamo/convert_frame.py", line 2396, in call result = self._torchdynamo_orig_backend( File "/opt/conda/envs/py_3.10/lib/python3.10/site-packages/torch/_dynamo/convert_frame.py", line 2119, in call result = self._inner_convert( File "/opt/conda/envs/py_3.10/lib/python3.10/site-packages/torch/_dynamo/convert_frame.py", line 726, in call result = _compile( File "/opt/conda/envs/py_3.10/lib/python3.10/site-packages/torch/_dynamo/convert_frame.py", line 1984, in _compile graph_node_shapes = output.get_graph_sizes_structured() File "/opt/conda/envs/py_3.10/lib/python3.10/site-packages/torch/_dynamo/output_graph.py", line 2350, in get_graph_sizes_structured ret[node.name] = [s if isinstance(s, int) else repr(s) for s in size] TypeError: 'int' object is not iterable

To execute this test, run the following from the base repo dir: PYTORCH_TEST_WITH_DYNAMO=1 python test/test_overrides.py TestTorchFunctionMode.test_modes_return_notimplemented

This message can be suppressed by setting PYTORCH_PRINT_REPRO_ON_FAILURE=0

TL;DR

The most likely fix involves addressing the 'int' object is not iterable error in the get_graph_sizes_structured method of output_graph.py.

Guidance

• Investigate the get_graph_sizes_structured method in output_graph.py to determine why an integer is being passed where an iterable is expected.
• Verify that the size variable in the get_graph_sizes_structured method is correctly formatted and contains the expected data type.
• Check the test_modes_return_notimplemented test in test_overrides.py to see if there are any specific test cases or inputs that could be causing the error.
• Run the test with the provided command PYTORCH_TEST_WITH_DYNAMO=1 python test/test_overrides.py TestTorchFunctionMode.test_modes_return_notimplemented to reproduce the error and gather more information.

Example

No code snippet is provided as the issue is related to a specific error in the PyTorch library, and modifying the code without more context could be potentially incorrect.

Notes

The error message suggests that there is an issue with the data type being passed to the get_graph_sizes_structured method. Without more information about the expected input and output of this method, it is difficult to provide a more specific solution.

Recommendation

Apply a workaround by modifying the get_graph_sizes_structured method to handle the case where an integer is passed instead of an iterable, or investigate and fix the root cause of why an integer is being passed in the first place.