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

Can't reproduce results training #23

Open
mallorbc opened this issue Feb 19, 2021 · 2 comments
Open

Can't reproduce results training #23

mallorbc opened this issue Feb 19, 2021 · 2 comments

Comments

@mallorbc
Copy link

With the provided splits, I can not reproduce the results when training from scratch. I get an f1 scores of 61.33% and 47.79% for TVSum and SumMe respectfully. I am thinking that maybe the seed of 1234 is not the correct seed to be used for generating these results from scratch?
Thanks!

@ok1zjf
Copy link
Owner

ok1zjf commented Feb 20, 2021

Hi Mallorbc,
The published results were produces with the the github code as is, that is, including the seed. Later on we
found out that the results slightly differed with later pytorch versions. It you setup you machine with the
package versions documented in the readme.md you should be able to get the published values.
Sorry I'm not able to help more.
Best
Jiri

@mallorbc
Copy link
Author

No, thats plenty of help and makes perfect sense. Thanks

I do have a follow up though:
How did you go about selecting the seed and the splits used for the final result? Using different seeds and or different splits leads to a noninsignificant change in the results.
Thanks again!

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