遇见数据集

3d-collection

收藏
魔搭社区2026-11-01 更新2026-11-01 收录
官方服务:

资源简介:

# Scenario 3D Collection 1,342 selected AI-generated 3D examples, with all 1,500 output files downloaded. About 39 GB including previews and available source inputs. Another 393 examples made with Tencent Hunyuan models are published separately under their own licenses: [Scenario 3D Collection (Hunyuan)](https://huggingface.co/datasets/scenario-labs/3d-collection-hunyuan). ## License [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). Credit "Scenario". All assets are AI-generated. ## Metadata `data/train-00000-of-00001.parquet` is the single metadata source, with one row per example. Original assets remain separate files. - `inputs`: structured prompt, effort, generation settings and source assets (IDs, relative paths and SHA-256 hashes). - `outputs`: structured output formats, roles, IDs, relative paths and SHA-256 hashes. - `preview`: embedded thumbnail for the dataset viewer; `preview_file` retains the original preview path and hash. - `input_files` and `output_files`: convenient download URLs. Top-level prompt and effort columns simplify browsing. `inputs.settings` is JSON text because parameter names and value types vary between models; parse it with `json.loads`. All original parameter values and types are preserved. Missing values appear as null; no defaults are inferred. Paths are relative to the repository root. ## Formats and validation Outputs: 1,271 GLB, 137 Blender, 41 SPZ, 23 PLY, 21 VOX and 7 FBX. Structural checks passed; all 137 Blender outputs loaded with auto-execution disabled. One Blender scene references two missing textures. These checks do not certify rendering or animation quality. ## Browse and download ```python import json from datasets import load_dataset examples = load_dataset("scenario-labs/3d-collection", split="train") example = examples[0] settings = json.loads(example["inputs"]["settings"] or "{}") source_assets = example["inputs"]["assets"] outputs = example["outputs"] ``` Use `huggingface_hub.snapshot_download` with `repo_type="dataset"` to download the original files. The Hunyuan companion requires access approval and authentication.

提供机构:
maas
创建时间:
2026-10-27
二维码
社区交流群
二维码
科研交流群
商业服务