Worked on the ONNX/onnx repository, focusing on enhancing the Reference Evaluator test suite through targeted maintenance and cleanup. Applied Python and unit testing expertise to remove dead code, correct imports, and update Python constructs, aligning the tests with standard ONNX practices. Improved documentation by fixing docstrings, which increased clarity and maintainability for future contributors. Addressed redundant and commented-out code in STFT-related tests, reducing noise and potential flakiness in the suite. These changes strengthened test reliability and traceability, enabling faster continuous integration feedback and smoother onboarding for new developers, while ensuring the test suite remains robust and easier to maintain.