# on valid_dataloaders for train valid_gt_encoder

[SolicTous]

1. Why you commented valid_gt_encoder validation from valid_dataloaders, and made it on train_dataloaders?
2. And here is commented net.eval(), so what is the reason?

[Maozhentan2024]

2.Not calling net.eval(): The model remains in training mode during validation, which may lead to unstable model outputs because Dropout and Batch Normalization layers operate according to the training mode.