You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I faced an issue when working with another STE project DiffSTU.
If resolution of masked word is high and a target word (a word I want to generate) has the same length as a source word, everything is OK. But when length of the target word longer or shorter (mainly longer), I only see artifacts and poor generation quality.
Have you faced such behavior in your model or it can handle such case somehow?
Thanks.
The text was updated successfully, but these errors were encountered:
I faced an issue when working with another STE project DiffSTU.
If resolution of masked word is high and a target word (a word I want to generate) has the same length as a source word, everything is OK. But when length of the target word longer or shorter (mainly longer), I only see artifacts and poor generation quality.
Have you faced such behavior in your model or it can handle such case somehow?
Thanks.
The text was updated successfully, but these errors were encountered: