PyVertical
PyVertical copied to clipboard
Make generalised dataset splitter functions (PyTorch)
Feature Description
- Create functions which split PyTorch datasets into separate datasets
- Should work for image and non-image datasets
- Functions should apply random IDs to datapoints
- Unit test the functions
"Splitting" in this context means to split input features into two separate datasets. For images, split them top/bottom (a further issue will look to extend this)
Is your feature request related to a problem?
We should provide utility code to make it easy for people to turn non-vertically federated datasets into vertically federated ones, for experimental purposes.
We currently have some code built for this task, but it is not generaliseable to a wide range of datasets
What alternatives have you considered?
- Don't provide generaliseable code: This is okay for initial experimentation, but PyVertical should be a widely useable package for VFL