I convert CNN data by your code but error
Tường Lâm

I found cause of the issue. The vocabulary file must contain records for sentence start and sentence end tokens. I fixed the code for vocabulary generation: https://github.com/surmenok/TextSum/blob/master/textsum_data_convert.py

One clap, two clap, three clap, forty?

By clapping more or less, you can signal to us which stories really stand out.