Stuck on an issue?

Lightrun Answers was designed to reduce the constant googling that comes with debugging 3rd party libraries. It collects links to all the places you might be looking at while hunting down a tough bug.

And, if you’re still stuck at the end, we’re happy to hop on a call to see how we can help out.

Cannot download IWSLT datasets from wit3.fbk.eu

See original GitHub issue

When trying translation examples, I found that prepare-iwslt14.sh and prepare-iwslt17-multilingual.sh both failed. After that, I tried to open wit3.fbk.eu but failed again. I want to train a model from scratch, so how can I download the IWSLT datasets?

Issue Analytics

State:
Created 3 years ago
Reactions:11
Comments:10 (1 by maintainers)

Top GitHub Comments

4reactions

myleottcommented, Jan 7, 2021

I had a backup of the original data archive and uploaded it here: http://dl.fbaipublicfiles.com/fairseq/data/iwslt14/de-en.tgz

I’ll update the docs with the new link.

4reactions

garyhlaicommented, Dec 30, 2020

For those of you still stuck, @zml24 , there is a new url available for downloading IWSLT at

https://drive.google.com/uc?id=1l5y6Giag9aRPwGtuZHswh3w5v3qEz8D8

Note that the link will download an IWSLT tgz containing the tgzs for all the languages. You still need to extract the specific language tgz you want. Once you extract the language-specific tgz, everything else should be the same as before.

I fixed this for Torchtext in a similar issue (https://github.com/pytorch/text/pull/1115/files) as a reference.

Top Results From Across the Web

Cannot download IWSLT datasets from wit3.fbk.eu · Issue #2984

After that, I tried to open wit3.fbk.eu but failed again. I want to train a model from scratch, so how can I download...

Unable to download IWSLT '14 datasets - Google Groups

https://wit3.fbk.eu/archive/2014-01/texts/de/en/de-en.tgz. Any advise on how I can download the data to complete this tutorial? Cheers,

Update files from the datasets library (from 1.3.0) · ted_talks_iwslt at ...

Update files from the datasets library (from 1.3.0) ; 227. + class TedTalksIWSLT(datasets.GeneratorBasedBuilder): ; 228. + """TED TALKS IWSLT: Web Inventory of ...

WIT3

The site is currently under maintenance. If you need some data listed above. whose link is not working yet, please contact. Mauro Cettolo...

torchnlp.datasets package — PyTorch-NLP 0.5.0 documentation

The torchnlp.datasets package introduces modules capable of downloading, ... {source}'], url='https://wit3.fbk.eu/archive/2016-01/texts/{source}/{target}/{ ...