Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Question about constructing test set #6

Open
TingchenFu opened this issue Aug 22, 2021 · 1 comment
Open

Question about constructing test set #6

TingchenFu opened this issue Aug 22, 2021 · 1 comment

Comments

@TingchenFu
Copy link

First of all, thanks for your great work!
I'm trying to run the code and model on the DialyDialog dataset to have a better understanding of your work. But I cant figure out how to construct the input to the model from /data/test.json. Also, the file 'test.refs.txt' which apprears in generate.py is not provided in this repository.
I have tried to construct the input to the model myself from the /data/test.json, but i was confused since i couldn't find multi-reference for every examples.
I wonder whether the code to preprocess /data/test.json could be released?

@lizekang
Copy link
Collaborator

Hi, thanks for your attention to our paper.
The "test.refs.txt" is built from the Reddit Dataset which is not allowed for open-source. You can build the test.refs.txt from the instruction of DialoGPT. But I don't recommend to do this.
For preprocessing "/data/test.json", you can refer to the dataset to see how to construct the input. Besides, the "/data/test.json" in this repo is not the multi-reference version. For the multi-reference version of DailyDialog dataset, please refer to "https://github.com/prakharguptaz/multirefeval".
If you have any questions, please feel free to contact me.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

2 participants