Datasets:
Languages:
English
Multilinguality:
monolingual
Size Categories:
100B<n<1T
Language Creators:
found
Annotations Creators:
no-annotation
Source Datasets:
original
ArXiv:
License:
Loading validation split only
#9
by
jeffreygwang
- opened
Hello! I am trying to load only the validation of The Pile and am having trouble doing so:
When I try running the below code, I get the following message: WARNING:datasets.builder:No config specified, defaulting to: the_pile/all
dataset = load_dataset("EleutherAI/the_pile", split="validation")
It proceeds to download the entire thing.
I have also tried specifying subsets of the validation set to load (e.g. load_dataset("EleutherAI/the_pile", subsets=subsets_to_load, split="validation")) to no avail. Is there a way to load just the validation split?