Closed 9p15p closed 8 months ago
https://github.com/Henry1iu/TNT-Trajectory-Prediction/blob/4c996808f4f6b9a2a12a62917c7f3647c74ed7a2/core/trainer/vectornet_trainer.py#L100
I think it should be self.criterion = VectorLoss(aux_loss=aux_loss, reduction="sum")
self.criterion = VectorLoss(aux_loss=aux_loss, reduction="sum")
Hi,
You are right. Thanks for identifying this bug.
https://github.com/Henry1iu/TNT-Trajectory-Prediction/blob/4c996808f4f6b9a2a12a62917c7f3647c74ed7a2/core/trainer/vectornet_trainer.py#L100
I think it should be
self.criterion = VectorLoss(aux_loss=aux_loss, reduction="sum")