Skip to content

Dataset · zxAi

60-class 3D object dataset in Gaussian Splatting format

zxAi/GSModel60

13,847 PLY files for 3D point cloud classification based on Gaussian Splatting with a fixed split.

What's inside

The GSModel60 dataset contains 13,847 3D models in PLY format distributed across 60 classes for classification tasks. It preserves the original split from the GAPrompt++ paper: 7,757 training examples and 6,090 test examples, fixed in train.json and test.json. Additionally, metadata (jsonl) specifies class labels, model IDs, paths, and file sizes.

What you can do with it

  • Training and testing 3D scene classification models in Gaussian Splatting format
  • Comparing 3D point cloud analysis algorithms on the standard GAPrompt++ split
Tasks
Other
Size
10K–100K rows
Downloads this month
31K
Likes
1

How to load it

from datasets import load_dataset

ds = load_dataset("zxAi/GSModel60")