What is the difference between the 280k dataset, the 10 classification dataset, and the Objaverse-XL alignment dataset?

[834810269]

Hello, thank you for your great work. I would like to know the difference between the 280k dataset, the 10 classification dataset, and the Objaverse-XL alignment dataset. Is it due to different cleaning standards or different cleaned data? When I sum up the available 8 classes in the 10 classification dataset, the total does not equal 280k.

[834810269]

And I found that the numbers in 'gobjaverse_alignment.json' (779325), 'gobjaverse_index_to_objaverse.json' (730053), and 'text_captions_cap3d.json' (659657) do not match up

[lingtengqiu]

I will check it , ASAP;

[834810269]

I will check it , ASAP;

Thanks a lot! And another question is why the data of 10 categories add up to 500K. Shouldn't objaverse have more than 700K pieces of data?

[hitsz-zuoqi]

I will check it , ASAP;

Thanks a lot! And another question is why the data of 10 categories add up to 500K. Shouldn't objaverse have more than 700K pieces of data?

Some data are dirty pointcloud or too big files for a multi-thread rendering pipeline or ... So we do not label these data~