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.

Sign up for GitHub

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

Jump to bottom

Corrected Submission URL #5

Open
wants to merge 1 commit into
base: master
Choose a base branch
from
Open

Corrected Submission URL #5

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
4 changes: 2 additions & 2 deletions data/README.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -8,7 +8,7 @@ In `train_full.csv` or `val_full.csv`: we have both the texts of the endings/con

## regular (shuffled)

This could be more interesting for modeling, and it's the way the test data is formatted. For each `startphrase` (also, split into `sent1`,`sent2`) we have 4 endings, and a label which says the correct one. You can use `test.csv` for submission on the leaderboard here: [https://leaderboard.dev.allenai.org/swag/submission/create](https://leaderboard.dev.allenai.org/swag/submission/create). The fields are exactly the same as `val.csv` and `train.csv` except for the label.
This could be more interesting for modeling, and it's the way the test data is formatted. For each `startphrase` (also, split into `sent1`,`sent2`) we have 4 endings, and a label which says the correct one. You can use `test.csv` for submission on the leaderboard here: [https://leaderboard.allenai.org/swag/submission/create](https://leaderboard.allenai.org/swag/submission/create). The fields are exactly the same as `val.csv` and `train.csv` except for the label.



Expand All @@ -22,4 +22,4 @@ If the source starts with `gold`, it comes from the found data (from an actual v

For training, we also have questions marked as `gen-orig`: these are generated answers that are selected as the *best* answer, while the real answer was selected as the second best (`gold1-orig`)

tl;dr you probably don't have to worry about this one. However, during training, some models work better if also shown the `gen-orig` examples, and some work better if not.
tl;dr you probably don't have to worry about this one. However, during training, some models work better if also shown the `gen-orig` examples, and some work better if not.