Currently, the dataloaders.py file is 1400 loc and shouldn't be so. We may end up implementing different dataloaders in the future, and this should be taken into account when designing the current ones. Right now, I'm planning on making singular, dataloader and dataset
classes that handle the timestep & sequential batch issue eloquently.
Currently, the dataloaders.py file is 1400 loc and shouldn't be so. We may end up implementing different dataloaders in the future, and this should be taken into account when designing the current ones. Right now, I'm planning on making singular, dataloader and dataset classes that handle the timestep & sequential batch issue eloquently.