Dataloader

[leonlenk]

• the dataloader file should return a dataloader based on a csv
• It should accept the path to an arbitrary csv for the dataset (you can make assumptions about the format) so that members can use their own datasets
• We should also have a default dataset probably this one
• Make sure to have a parameter that allows it to easly take a subset of the data

[floatingtrees]

I can do this one