ygorg/kptimes

Repository for KPTimes corpus

kptimes's Issues

Could you please provide the processed data？

Could you please provide the processed data, for example, put it on Google Drive. Because my network here is not good, I cannot use the shell script to download the dataset. Thanks.

what's the size of test set

Table 1 in the paper shows test set as 10k but when we downloaded and checked, it amounted to 20k. Could you please double check the figures?

Thanks.

Broken Links

Hi,
Thanks for making this dataset available. Would it be possible to provide the full jsonl files, instead of having to download the html files one-by-one?

I tried downloading the datasets using the scripts. However, the news articles for certain links have been removed by the publisher and hence are broken links. For example:

1. https://www.japantimes.co.jp/news/2014/10/05/asia-pacific/offbeat-asia-pacific/south-korea-rumor-crackdown-jolts-social-media-users
2. https://www.japantimes.co.jp/news/2014/10/20/asia-pacific/offbeat-asia-pacific/china-zoo-bear-bites-boys-arm-off

• With JPTimes test set, there were 1576 missing articles.

Asking INLG 2019 to change paper title on the accepted papers page ?

Wrong paper title at https://www.inlg2019.com/acceptedpaperlist

Include results from baselines (neural & non-neural) models for assigning keyphrases

It would be useful to include the results from baselines (neural & non-neural) models for assigning keyphrases.