Opened this issue 12 days ago · 0 comments
Hello!I'm encountering a problem when training on a dataset that is not strictly forward-facing. Is there a way to handle this?