Open shuangw98 opened 8 months ago
Hi,
Thanks for your interest!
We're using a fixed focal length so that training and inference would share this same fov setting and impose no extra need for GT camera parameters during inference. We haven't tried using a random focal length during training, but we suspect it to be unnecessary for models trained solely on Objaverse.
Plz feel free to comment.
Thanks for your great work! I noticed that you used a fixed focal length when rendering objaverse, which can cause the focal length of the input image to be inconsistent with the camera model. I’m curious why you didn’t use a random focal length.