Evaluation discrepancy with Model Zoo checkpoint: DET_ll = 0.201 (far below paper)

[Frankyezq]

I evaluated TopoLogic using the pre-trained model from the Model Zoo and obtained DET_ll = 0.201, which is much lower than the number reported in the paper. I followed the default evaluation setup without any training or code changes.