datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
VIPS-v2xreal-assets
VIPS — V2X-Real Evaluation Assets
Generated assets for the VIPS benchmark — a NAVSIM-style two-stage PDM
(Predictive Driver Model) evaluation on the V2X-Real
dataset. Use them together with the code at
github.com/mickeykang16/VIPS.
These are generated assets only. The official V2X-Real raw sensor images
are not redistributed here — download them from UCLA (step 1 below).
Contents
Path
Description
meta/spd_infos_temporal_test.pkl
Temporal test-split… See the full description on the dataset page: https://huggingface.co/datasets/mickeykang/VIPS-v2xreal-assets.roam_assets_gen_data3d-1k-assets
3D-1K Assets
A curated 3D rigid-object asset library with a 10-category taxonomy and 1,000 GLB assets (one per base object class). Generated by a text → image → RGBA → 3D pipeline (FLUX.2-klein → RMBG-2.0 → TRELLIS.2).
This is the base class layer of a planned 1,000 × 1,000 = 1,000,000 instance dataset for embodied AI / robotic manipulation.
Files
file
size
content
glbs.tar
6.5 GB
1000 GLB assets: 0001.glb … 1000.glb
rgba_images.tar
700 MB
1000 RGBA… See the full description on the dataset page: https://huggingface.co/datasets/Dennis0626/3d-1k-assets.baseline_3_v4_collection_assets
baseline_3 v4 — sim collection assets
2026-05-26 — STATUS: ARCHIVED / REFERENCE-ONLY.
The plan to run sim collection on the A6000 was abandoned this round
due to a glibc incompatibility: the A6000 system glibc is 2.31 (Ubuntu
20.04) but IsaacSim 5.1.0 requires glibc 2.35 (Ubuntu 24.04). All sim
collection (DexYCB + OakInk) was completed on the dev box (RTX 5090).
The contents of this dataset (tarballs of source episodes, USDs, scripts,
partial results log) are kept for reference… See the full description on the dataset page: https://huggingface.co/datasets/UCBProject/baseline_3_v4_collection_assets.assets_