The load_batches generator iterates through the entire dataset, but it's often required to split into multiple sets (e.g. train/validation). We should have a method that creates multiple generators, each with distinct subsets of the full data but with all other parameters kept the same.
The
load_batchesgenerator iterates through the entire dataset, but it's often required to split into multiple sets (e.g. train/validation). We should have a method that creates multiple generators, each with distinct subsets of the full data but with all other parameters kept the same.