Datasets:

ArXiv:
License:

Cannot stream dataset Yodas2

#5
by martinsipka - opened

Hi,

Trying to stream the dataset with minimal code

from datasets import load_dataset

ds = load_dataset('espnet/yodas2', "es000", streaming=True)
print(next(iter(ds['train'])))

Unfortunatelly I am getting errors:

Failed with encoding: utf-8
JSON decode error with encoding: latin-1
Failed with encoding: cp1252
JSON decode error with encoding: iso-8859-1
Could not read file with any encoding: https://huggingface.co/datasets/espnet/yodas2/resolve/c9674490249665d658f527e2684848377108d82c/data/es000/text/00000000.json

Streaming Yodas dataset (Instead of Yodas2) seems to work fine.

Using

datasets           3.3.2
huggingface-hub    0.34.4

Thank you for any ideas what I may be doing wrong.

Sign up or log in to comment