Fix local run qdq op.

[chinazhangchao]

#182

[DingmaomaoBJTU]

Code review

Found 1 issue:

1. logger.warning for an expected build failure should be logger.debug — this reverses a deliberate downgrade from a prior fix. When shape inference is incomplete (common for large QDQ models), the model-build exception fires for many nodes and floods the log. The equivalent handler on main uses logger.debug.

https://github.com/microsoft/WinML-ModelKit/blob/769a7b60a481dffa976a2ab2cc010475d3315bf8/src/winml/modelkit/analyze/core/runtime_checker_query.py#L1804-L1812

🤖 Generated with Claude Code

_{- If this code review was useful, please react with 👍. Otherwise, react with 👎.}