CompVis
训练问题，ImageNet
训练自动编码器，报一下错误，请问是什么原因呢？ IndexError: Caught IndexError in DataLoader worker process 0. Original Traceback (most recent call last):
IndexError: list index out of range
It seems the indices in imagenet_train_hr_indices.p larger than the number of imagenet training images.
I replaced ImageNetSRTrain with ImageNetTrain.
if i want to train my own datasets，should I train the autoencoder first using my own datasets??
if i want to train my own datasets，should I train the autoencoder first using my own datasets??
nah, it is not a strict requirement although desired. You can also use one of the autoencoders (the checkpoints) on the repo. It would give you an idea of how the model performs. If it is good, then you may go and train your own autoencoder. But for the first check-up, I would go with a trained autoencoder.
