Open
Description
Really like this work @hbutsuak95 @krrish94.
It seems that your implementation doesn't contain the "--type both" training mode (more specifically, can not training in this mode). I think its a strong prior and wanna see an ablation study in your paper. But I didn't find it. If you have tried related experiments, plz share the results to me. Thanks a lot!
Metadata
Assignees
Labels
No labels