I think it is definitely doable, but we do not have plan for that. You can train your own.
BTW, I think our model is possible to generate monocular camera by only changing the inference pipeline (e.g., changing cross-view attn to self-attn). You may have a try.
I think it is definitely doable, but we do not have plan for that. You can train your own.
BTW, I think our model is possible to generate monocular camera by only changing the inference pipeline (e.g., changing cross-view attn to self-attn). You may have a try.