Do you have a plan to pre-train Uni3D on Objaverser-xl, a 10M-scale 3D dataset?

[auniquesun]

Thanks for sharing the paper and code. It's a great work.

Since Uni3D has scaled the point encoder to 1B parameters, that are rather big. But the scale of Objaverse 1.0 is only 800K 3D objects, I think it is still relatively small to support the 1B-scale point encoder pre-training and the generalization is still far behind the counterparts in image and text field.

Now Objaverse-xl is released, which contains 10M+ 3D objects, does your team have such a plan to pre-train on larger Objaverse-xl? I think BAAI has the computing source to finish such a task. How do you think about it?

[QiopeWallt]

I also think scaling up pre-training using the Objaverse-xl dataset could greatly improve the generalization of the point encoder. I’m curious—has your team processed the Objaverse-xl dataset for Uni3D adaptation yet? If not, perhaps there’s an opportunity for collaboration.