Closed xnuohz closed 3 days ago
Hi ! If the dataset owner uses push_to_hub()
instead of save_to_disk()
and upload the local files it will fix the issue.
Right now datasets
sees the train/test/valid pickle files but they are not supported file formats.
Alternatively you can load the arrow file instead:
from datasets import load_dataset
dataset = load_dataset('OpenMol/PubChemSFT', data_files='stage1/*.arrow')
Thanks! I'll have a try.
Describe the bug
Cannot load the dataset https://huggingface.co/datasets/OpenMol/PubChemSFT
Steps to reproduce the bug
Expected behavior
Environment info