You need to agree to share your Hugging Face information to access this dataset
This repository is publicly accessible, but you have to accept the conditions to access its files and content.
Full 200 access costs CNY 100 one-time, with manual payment confirmation and account approval. Contact allen@datumwell.io before requesting access.
Log in or Sign Up to review the conditions and access this dataset content.
Datumwell 3D-full
中文
200 件可检查、可编辑的通用3D资产，覆盖杯、瓶、收纳盒、凳子、花盆；完整集每类40件。每件保留几何参数、seed、米制单位、稳定部件标识与生成源码。
完整集含20个factory，每个10组几何参数，共200件，包含免费20件样本。价格为¥100人民币一次性；提交申请不收费，也不会自动开通访问。请联系确认用途、授权范围和付款后，由Datumwell人工审核并开通Hugging Face账号。
每件包含 GLB、可编辑 Blender 工程、Blender Python 源码、JSON参数/相机/元数据，以及6张320×320 PNG预览。本仓库共有1200张预览。颜色、材质或相机视角不会增加资产计数。
更多数据或定制范围，请联系 allen@datumwell.io。免费样本 · 完整集 · Dataset
English
200 inspectable and editable generic 3D assets across cups, bottles, storage boxes, stools and planters; the full collection contains 40 assets per category. Each asset retains geometric parameters, a seed, metric units, stable part IDs and editable source.
The full collection contains 20 factories with 10 geometric parameter sets each: 200 assets, including the free 20-asset sample. The price is CNY 100 one-time. Submitting a request does not charge you or grant automatic access. Contact Datumwell to confirm intended use, usage scope and payment; account access is reviewed manually.
Each asset includes GLB, an editable Blender scene, Blender Python source, JSON parameters/cameras/metadata and six 320×320 PNG views. This repository contains 1200 previews. Colors, materials and camera views do not increase the asset count.
For more data or a tailored scope, contact allen@datumwell.io. Free sample · Full collection · Dataset
Loading / 加载
The JSONL rows contain metadata and repository-relative asset paths. parameters is a JSON-encoded string to preserve a stable schema across factory-disjoint splits; decode it with json.loads(row["parameters"]). Original parameter objects remain in metadata/instances.json and each source/request.json. For the gated full collection, use the same approved Hugging Face account when downloading files.
from datasets import load_dataset
from huggingface_hub import hf_hub_download
ds = load_dataset("allen-datumwell/3D-full", token=True)
row = ds["train"][0]
glb = hf_hub_download("allen-datumwell/3D-full", row["glb_path"], repo_type="dataset", token=True)
- Downloads last month
- -