Flexible Loader

[dipta007]

Feature request

Can we have a utility function that will use load_from_disk when given the local path and load_dataset if given an HF dataset?

It can be something as simple as this one:

def load_hf_dataset(path_or_name):
    if os.path.exists(path_or_name):
        return load_from_disk(path_or_name)
    else:
        return load_dataset(path_or_name)

Motivation

This can be done inside the user codebase, too, but in my experience, it becomes repetitive code.

Your contribution

I can open a pull request.

[lhoestq]

Ideally save_to_disk should save in a format compatible with load_dataset, wdyt ?

[dipta007]

Ideally save_to_disk should save in a format compatible with load_dataset, wdyt ?

That would be perfect if not at least a flexible loader.