Performance testing

[LeTrungLinh]

Thank you for your research. i also retrain your model with your config on GTX3090ti on COVID-QU-Ex dataset (3728:932:1166 for Train:Val:Test), but it's give a really high performance than the table in your paper. Can you help me to verify my confusion:

dice0.9807 _miou0.9629 _pre0.9718 _recall0.9702 _f1_score0.9704 _pa0.9866

[FDG2801]

hi, can you tell me how did you manage to run the covid dataset? I'm having problems with the type of files. Should I use the txt for validation? what about training?

[LeTrungLinh]

Hi friend, I just use the model file of CoTrFuse. For manage the covid dataset, i just put it in separate folder and process it by Dataset code that use can implement by yourself.

[BinYCn]

It sounds like the performance discrepancy might stem from differences in how evaluation metrics are being computed.