Create issues

[lbda1]

what's the format of training data

[mdtux89]

See instructions of https://github.com/sinantie/NeuralAmr for data preprocessing

[lbda1]

Thank you very much. But the format of generated data by https://github.com/sinantie/NeuralAmr is like this. source: n / need :arg0 t / they :arg1 e / explain :time a / always a / add :arg0 h / he :arg1 ( c2 / cause :arg1 ( c / come :arg1 y / you :arg3 s2 / sky :mod t2 / too ) ) :time t / then target: they always need to have things explained . then he added : so you , too , come from the sky ! But the preprocessing code of your AMR-to-text like below, is seperated by "|". The data don't have one. At 2020-02-15 18:31:58, "Marco Damonte" <[email protected]> wrote: See instructions of https://github.com/sinantie/NeuralAmr for data preprocessing — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub, or unsubscribe.

[mdtux89]

I'm not sure what you mean. What is the code that separates by '|'? Can you point me to the file and line?

[lbda1]

Thanks for your help. At 2020-02-17 17:14:57, "Marco Damonte" <[email protected]> wrote: I'm not sure what you mean. What is the code that separates by '|'? Can you point me to the file and line? — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub, or unsubscribe.