The results are not the same each time

[CumtStudentZhanghua]

Over the past six months, I have been working with MMDetection, and I would like to seek advice from the community. I have observed that, despite using the same dataset and code, the results vary with each run. To address this, I have explicitly set randomness = dict(seed=0, deterministic=True) in my configuration. I have also experimented with multiple versions of MMDetection, including 3.0.0 and 2.28.2, but the issue persists: the results remain inconsistent across runs with identical datasets and code. Could anyone provide insights or suggestions to resolve this problem?