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
If one use a textual inversion to describe the style image, how to separate style and content in the reference image? I understand that "content" is part of the "style"(as the paper mentioned, semantics elements are part of style). However, in Fig.4, 2rd row, InST result is totally a robot as the reference image, while the content is a human. So I am confused about the evaluation: which parts of elements of the style image are we really want to transfer to content image...
The text was updated successfully, but these errors were encountered:
If one use a textual inversion to describe the style image, how to separate style and content in the reference image? I understand that "content" is part of the "style"(as the paper mentioned, semantics elements are part of style). However, in Fig.4, 2rd row, InST result is totally a robot as the reference image, while the content is a human. So I am confused about the evaluation: which parts of elements of the style image are we really want to transfer to content image...
The text was updated successfully, but these errors were encountered: