Giters

xinyadu / nqg

neural question generation for reading comprehension

Home Page:https://arxiv.org/abs/1705.00106

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

the output file

j92415 opened this issue · comments

commented

Hello!
I had finished the above commands before the last evaluation.

I have two question:
1.I don't know how to key the last command
"./eval.py --out_file "
the output file must be created first?
and what the output file type should be created?
Can u give me a example?

I have train 15 epochs.t7 in Paragraph and Sentence
the command "th translate.lua -model model/ -config config-trans"
means i must to execute translate.lua in 30 times?

I am a rookie in the kind of model
I will be very grateful If you can answer me !

commented

The output file is created when you run the translate command which creates a pred.txt file. (or using -output when running the translate.lua to chose own file name)

If you want to evaluate all 15 epochs for the paragraph and sentence model, you will need to translate each model and generate an output file (make sure you name the output files as it will override the file when you arent using the -output option). It's more likely you want the best results, choose the model with the lowest perplexity and translate with that model.