I have a custom dataset and I would like to split that dataset into a "training" and "test" set (also potentially a "validation" set if possible). How would I achieve this using Flux.jl or other Julia machine learning packages?
You can import the TrainTestSplit
function from the Lathe package, as in:
using Lathe.preprocess: TrainTestSplit
and then implement it in your code like this for example:
dataset_id = TrainTestSplit(datasetmap[:], 0.8); #datasetmap is your label encoded matrix
Am assuming you're using Pluto notebook but, it should work in any other environment as well i,e jupyter, atom, etc.