kind inquiry about the uids of G-objaverse

[Moondok]

Hi, thank you for the high-quality dataset!
It was mentioned in README that there are 10 categories of objects, and the total number of those objects is around 580k.
Could you please provide the uids of the holistic g-objaverse dataset ? (not the 280k subset selected with clip score)
Thanks in advance.

[fdyuandong]

Hi, if you want to download a specific category in gobjaverse280k:

Step1: download the index file of the specified category.

wget https://virutalbuy-public.oss-cn-hangzhou.aliyuncs.com/share/aigc3d/gobjaverse_280k_split/gobjaverse_280k_{category_name}.json # category_name: Human-Shape, ...

Step2: download using script.

Example: python ./scripts/data/download_gobjaverse_280k.py ./gobjaverse_280k_Human-Shape ./gobjaverse_280k_Human-Shape.json 10

python ./download_gobjaverse_280k.py /path/to/savedata /path/to/gobjaverse_280k_{category_name}.json nthreads(eg. 10)

[Moondok]

Thank you for the reply!
But I noticed that the number of objects in gobjaverse_280k_{category}.json is quite less that it claimed in README.

In some early issues #13 , you clarified that the g-280k is a subset selected with clip score, could you please provide the origin version without any selection ( i.e. 10 categories with around 580k ) ?