-
Notifications
You must be signed in to change notification settings - Fork 554
dataset: Add hebrew v3 #3607
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
dataset: Add hebrew v3 #3607
Conversation
KennethEnevoldsen
left a comment
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Should we just run a model on both datasets to compare
I am kinda unsure how about the differences in the data
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
We need to add superseeded_by
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
V1 is already superseeded_by v2
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
v2 needs a superseeded_by v3
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Then we should clenup it as in #2900, but I can't find script for this
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Maybe. Try to run this too
|
Results for
|
|
hmm would have expected it would have gotten better |
|
Results for
|
|
Do we need to add v3? |
|
Changed |
Close #3578