pytorch - 💡(How to fix) Fix DISABLED test_torchvision_models_efficientnet_b2 (__main__.TestVisionTracing) [1 comments, 1 participants]

Traceback (most recent call last): File "/__w/pytorch/pytorch/test/test_fx.py", line 5426, in run_test script = torch.jit.script(graph) File "/opt/conda/envs/py_3.10/lib/python3.10/site-packages/torch/jit/_script.py", line 1498, in script ret = _script_impl( File "/opt/conda/envs/py_3.10/lib/python3.10/site-packages/torch/jit/_script.py", line 1196, in _script_impl return torch.jit._recursive.create_script_module( File "/opt/conda/envs/py_3.10/lib/python3.10/site-packages/torch/jit/_recursive.py", line 566, in create_script_module return create_script_module_impl(nn_module, concrete_type, stubs_fn) File "/opt/conda/envs/py_3.10/lib/python3.10/site-packages/torch/jit/_recursive.py", line 634, in create_script_module_impl script_module = torch.jit.RecursiveScriptModule._construct(cpp_module, init_fn) File "/opt/conda/envs/py_3.10/lib/python3.10/site-packages/torch/jit/_script.py", line 673, in _construct init_fn(script_module) File "/opt/conda/envs/py_3.10/lib/python3.10/site-packages/torch/jit/_recursive.py", line 610, in init_fn scripted = create_script_module_impl( File "/opt/conda/envs/py_3.10/lib/python3.10/site-packages/torch/jit/_recursive.py", line 634, in create_script_module_impl script_module = torch.jit.RecursiveScriptModule._construct(cpp_module, init_fn) File "/opt/conda/envs/py_3.10/lib/python3.10/site-packages/torch/jit/_script.py", line 673, in _construct init_fn(script_module) File "/opt/conda/envs/py_3.10/lib/python3.10/site-packages/torch/jit/_recursive.py", line 610, in init_fn scripted = create_script_module_impl( File "/opt/conda/envs/py_3.10/lib/python3.10/site-packages/torch/jit/_recursive.py", line 634, in create_script_module_impl script_module = torch.jit.RecursiveScriptModule._construct(cpp_module, init_fn) File "/opt/conda/envs/py_3.10/lib/python3.10/site-packages/torch/jit/_script.py", line 673, in _construct init_fn(script_module) File "/opt/conda/envs/py_3.10/lib/python3.10/site-packages/torch/jit/_recursive.py", line 610, in init_fn scripted = create_script_module_impl( File "/opt/conda/envs/py_3.10/lib/python3.10/site-packages/torch/jit/_recursive.py", line 634, in create_script_module_impl script_module = torch.jit.RecursiveScriptModule._construct(cpp_module, init_fn) File "/opt/conda/envs/py_3.10/lib/python3.10/site-packages/torch/jit/_script.py", line 673, in _construct init_fn(script_module) File "/opt/conda/envs/py_3.10/lib/python3.10/site-packages/torch/jit/_recursive.py", line 610, in init_fn scripted = create_script_module_impl( File "/opt/conda/envs/py_3.10/lib/python3.10/site-packages/torch/jit/_recursive.py", line 634, in create_script_module_impl script_module = torch.jit.RecursiveScriptModule._construct(cpp_module, init_fn) File "/opt/conda/envs/py_3.10/lib/python3.10/site-packages/torch/jit/_script.py", line 673, in _construct init_fn(script_module) File "/opt/conda/envs/py_3.10/lib/python3.10/site-packages/torch/jit/_recursive.py", line 610, in init_fn scripted = create_script_module_impl( File "/opt/conda/envs/py_3.10/lib/python3.10/site-packages/torch/jit/_recursive.py", line 634, in create_script_module_impl script_module = torch.jit.RecursiveScriptModule._construct(cpp_module, init_fn) File "/opt/conda/envs/py_3.10/lib/python3.10/site-packages/torch/jit/_script.py", line 673, in _construct init_fn(script_module) File "/opt/conda/envs/py_3.10/lib/python3.10/site-packages/torch/jit/_recursive.py", line 610, in init_fn scripted = create_script_module_impl( File "/opt/conda/envs/py_3.10/lib/python3.10/site-packages/torch/jit/_recursive.py", line 638, in create_script_module_impl create_methods_and_properties_from_stubs( File "/opt/conda/envs/py_3.10/lib/python3.10/site-packages/torch/jit/_recursive.py", line 473, in create_methods_and_properties_from_stubs concrete_type._create_methods_and_properties( File "/opt/conda/envs/py_3.10/lib/python3.10/site-packages/torch/jit/_recursive.py", line 1037, in compile_unbound_method create_methods_and_properties_from_stubs(concrete_type, (stub,), ()) File "/opt/conda/envs/py_3.10/lib/python3.10/site-packages/torch/jit/_recursive.py", line 473, in create_methods_and_properties_from_stubs concrete_type._create_methods_and_properties( RuntimeError: builtin cannot be used as a value: File "/opt/conda/envs/py_3.10/lib/python3.10/site-packages/torch/nn/modules/conv.py", line 550 return F.conv2d( F.pad( input, self._reversed_padding_repeated_twice, mode=self.padding_mode ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ <--- HERE ), weight, 'Conv2d._conv_forward' is being compiled since it was called from 'Conv2d.forward' File "/opt/conda/envs/py_3.10/lib/python3.10/site-packages/torch/nn/modules/conv.py", line 565 def forward(self, input: Tensor) -> Tensor: return self._conv_forward(input, self.weight, self.bias) ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ <--- HERE

To execute this test, run the following from the base repo dir: python test/test_fx.py TestVisionTracing.test_torchvision_models_efficientnet_b2

This message can be suppressed by setting PYTORCH_PRINT_REPRO_ON_FAILURE=0

TL;DR

The test test_torchvision_models_efficientnet_b2 is failing due to a RuntimeError caused by the builtin function being used as a value in the Conv2d module.

Guidance

• Investigate the Conv2d module in torch/nn/modules/conv.py to understand why builtin is being used as a value.
• Check the PyTorch version being used, as this issue may be related to a known bug that has been fixed in a later version.
• Run the test with the PYTORCH_PRINT_REPRO_ON_FAILURE=1 environment variable set to get more detailed error messages.
• Try to reproduce the issue by running the test locally using the command python test/test_fx.py TestVisionTracing.test_torchvision_models_efficientnet_b2.

Example

No code snippet is provided as the issue is related to a specific PyTorch module and requires further investigation.

Notes

The issue is marked as flaky, which means it may not always reproduce. The test has been disabled in CI to prevent false negatives.

Recommendation

Apply workaround: Investigate and fix the Conv2d module issue or upgrade to a later PyTorch version if the issue is related to a known bug.

FAIL-SAFE: Given the information, the best course of action is to investigate the Conv2d module and try to reproduce the issue locally. If the issue persists, consider upgrading to a later PyTorch version or seeking further assistance from the PyTorch community.