IndexError in Conv1dToConv2d pass due to incorrect argument count check

[Chtho11y]

🐛 Describe the bug

When converting a model to QNN backends using ExecuTorch, the conversion process fails at the Conv1dToConv2d pass with a tuple index out of range error:

Traceback (most recent call last):
  File "/root/anaconda3/envs/executorch/lib/python3.10/site-packages/torch/fx/passes/infra/pass_manager.py", line 276, in __call__
    res = fn(module)
  File "/root/anaconda3/envs/executorch/lib/python3.10/site-packages/torch/fx/passes/infra/pass_base.py", line 46, in __call__
    res = self.call(graph_module)
  File "/workspace/executorch/backends/qualcomm/_passes/convert_conv1d_to_conv2d.py", line 108, in call
    groups = node.args[6] if num_args > 5 else 1
IndexError: tuple index out of range

The error occurs at executorch/backends/qualcomm/_passes/convert_conv1d_to_conv2d.py:108 where the code checks for group argument:

groups = node.args[6] if num_args > 5 else 1

Obviously, the current condition num_args > 5 is incorrect for accessing node.args[6].
Modify it to num_args > 6 could fix the bug, just like line 122.

Versions

PyTorch version: 2.8.0.dev20250625+cpu
executorch==0.8.0a0+75d4b2e

cc @cccclai @winskuo-quic @shewu-quic @cbilgin

[shewu-quic]

Thanks for your finding. We will fix it in the incoming PR.

[SS-JIA]

Hello @shewu-quic - has this been fixed at this point?

[cccclai]

I think there quite a lot of optimization done for the past few months, can you try it again? Also it will be helpful to have the model definition to help us repro

[nil-is-all]

Hi @Chtho11y, were you able to try this again? Since there were some improvements done over the past few months, do let us know if this bug persists.

[shewu-quic]

Sorry for late reply. We have fixed this error in this PR. If you have any problem, please feel free to let me know.
Thanks