harvardnlp / sent-summary

sent-summary

Training and evaluation data for Gigaword is available https://drive.google.com/open?id=0B6N7tANPyVeBNmlSX19Ld2xDU1E

Training and evaluation for CNN/DM is available https://s3.amazonaws.com/opennmt-models/Summary/cnndm.tar.gz

sent-summary's Issues

Gigaword dataset is not available

Gigaword dataset is not available， i cant open the link

Is there any paper about this data?

Is there any paper that report there performance on this dataset?

how to process cnndm article data(train.txt) and clean it for summarization task?

Gigaword: https://data.deepai.org/gigaword.zip

Code for producing the data

Hi there,

is the code that you used to format the data available? Specifically, if still on hand, could you share the script that converts the CNN / DM story files one can extract by means of this repo into the src and tgt files? I would like to use an existing approach that relies on the file format from this repo for a new data source.

Data for Gigaword is not available now

Training and evaluation data for Gigaword https://drive.google.com/open?id=0B6N7tANPyVeBNmlSX19Ld2xDU1E is not available now, can you share this resource again? Thanks a lot!

About testing

Hi! May I ask that do we need to remove the UNK word from the predicted and reference when calculating rouge? Thank you very much!