Hi, thanks for the high-quality rendering dataset. I've downloaded the dataset via download_objaverse_280k_tar.py, and I counted the number of objects by the following script:
root_dir = "path/to/GObjaverse_280k"
dir_ids = sorted(os.listdir(root_dir)) # 0~159
all_ids = [
f"{dir_id}/{object_id}"
for dir_id in dir_ids
for object_id in sorted(os.listdir(os.path.join(args.root_dir, dir_id)))
]
It turns out to be 258,825 objects by len(all_ids). However, the length of gobjaverse_280k_index_to_objaverse.json and gobjaverse_280k.json are both 265,232.
So I wonder if I missed some objects by downloading via the tar way (download_objaverse_280k_tar.py).
Hi, thanks for the high-quality rendering dataset. I've downloaded the dataset via
download_objaverse_280k_tar.py
, and I counted the number of objects by the following script:It turns out to be 258,825 objects by
len(all_ids)
. However, the length ofgobjaverse_280k_index_to_objaverse.json
andgobjaverse_280k.json
are both 265,232.So I wonder if I missed some objects by downloading via the tar way (
download_objaverse_280k_tar.py
).Thank you very much!