Open laisimiao opened 1 year ago
The main difference is in the train phrase, we add the correlative decoder based on the MAE.
Hello, I received an error message 'Network output is NAN, stop training!' during the training process. Do you know what could be the cause? My PyTorch version is 1.10.1 and my GPU is 4090
The main difference is in the train phrase, we add the correlative decoder based on the MAE.