Error loading the dataset

#5
by vpkprasanna - opened

from datasets import load_dataset

dataset = load_dataset("oscar", "unshuffled_original_ta")

Error
[{'expected': SplitInfo(name='train', num_bytes=9934879441, num_examples=1263280, shard_lengths=None, dataset_name=None), 'recorded': SplitInfo(name='train', num_bytes=9933590150, num_examples=1263180, shard_lengths=[59000, 64000, 61000, 63000, 71000, 68000, 63000, 60000, 64000, 67000, 67000, 64000, 64000, 64000, 63000, 67000, 64000, 61000, 66000, 43180], dataset_name='oscar')}]

how to resolve this ?

Hi! I fixed the size info in https://huggingface.co/datasets/oscar/discussions/6.

For future reference, you can also pass verification_mode="no_checks" to load_dataset in such scenarios to ignore the verification checks.

mariosasko changed discussion status to closed

Sign up or log in to comment