Can't load dataset as language_pair and year.
#3
by
FredDYyy
- opened
I'm working at an NMT task for my School project and yesterday I can still use dataset = load_dataset("ted_talks_iwslt", language_pair=("en", "vi"), year="2016"), but today when I try to reload it, it fails and stated that "Config name is missing", anyone having the same error or is there any way around to load the dataset as language pair and year? (only language pair matters, any year would by fine)
Thanks for reporting, @FredDYyy .
This is a known issue: https://github.com/huggingface/datasets/issues/6533
And we have recently fixed it and made a patch release:
- https://github.com/huggingface/datasets/pull/6544
- https://github.com/huggingface/datasets/releases/tag/2.16.1
Could you please try again after having updated datasets?
pip install -U datasets