Objaverse XL Shape Annotations: 537,841 3D Objects with Geometry-Focused Descriptions

537,841 shape-based textual annotations for objects from the Objaverse-XL benchmark. The dataset was created by valira-ai to provide geometry-focused descriptions for each 3D object, which were previously missing. It was last updated on June 18, 2026.

Use Cases

• Text-to-3D retrieval tasks based on shape-focused textual descriptions.
• Contrastive representation learning for 3D objects based on their geometry-focused annotations.
• Training or fine-tuning vision-language models for 3D shape understanding based on structured textual labels.
• Benchmarking 3D object classification and description generation models based on the provided class labels and descriptions.

Strengths

• Annotations for 537,841 objects provide a substantial scale for model training.
• Descriptions focus strictly on shape and structure, offering a consistent annotation style.
• Created to address a specific gap (lack of text) in the large Objaverse-XL benchmark.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment.
• Last updated 2026-06-18 09:42:22; freshness should be verified.

Provenance

Source

Annotations for objects from the Objaverse-XL benchmark, created by valira-ai.

Collection Method

Built to add textual descriptions focusing on shape and structure to the existing 3D objects.

Freshness

Last updated 2026-06-18 09:42:22.