Open Suho0922 opened 5 months ago
In the paper, the pose parameter is fixed to I_4 during the content personalization. However, in the code, the pose parameter is not I_4. I think it uses the camera parameter of sample views. is it correct? or the bug?
I_4 for uncond embedding, camera parameters for cond embedding.
In the paper, the pose parameter is fixed to I_4 during the content personalization. However, in the code, the pose parameter is not I_4. I think it uses the camera parameter of sample views. is it correct? or the bug?