How do I download the datasets?
Use the Python API for subsets or torrents for full mirrors. Verify checksums where available.
What formats are supported?
GLB and USDZ are common; Objaverse++ may include neural fields such as NeRF/3DGS.
How big is Objaverse‑XL?
Expect multi‑terabyte storage for full mirrors; most teams start with small targeted pulls.
Is there a Python API?
Yes — try objaverse and objaverse‑plus on PyPI.
Recommended hardware?
XL: 2TB+ SSD, 32–64GB RAM, 16GB+ VRAM GPU. ++: 3TB+ SSD, 64–128GB RAM, 24GB+ VRAM.