search

weidler / semlc

Biologically Inspired Lateral Connectivity in Convoluiontal Neural Networks
MIT License
1 stars 0 forks source link

allow training with augmentation #11

Closed jlehnen closed 5 years ago

jlehnen commented 5 years ago

Some transormations like RandomCrop return PIL Images rather than Tensors, that our current framework does not allow. Probably not hard or time consuming but needs to be done

weidler commented 5 years ago

Was indeed rather trivial. The ToTensor and Normalize transforms need to be in the end! Then the dataloader first augments, then converts to tensor and can return the tensor instead of a PIL image.