Callitrain

3,192 image–annotation pairs form the CalliBench dataset for evaluating Vision Language Models on Chinese calligraphy. The dataset, created by author gtang666, includes tasks for full-page recognition and contextual visual question answering. It was last updated on Hugging Face in July 2025.

Use Cases

• Benchmarking Vision Language Models on Chinese calligraphy recognition based on the full-page recognition task.
• Training models for author attribution based on metadata annotations for calligraphy authors.
• Developing bilingual interpretation systems based on the bilingual interpretation VQA task.
• Analyzing artistic intent and style in historical documents based on the intent analysis VQA task.

Strengths

• 3,192 annotated image–text pairs provide a defined scale for model evaluation.
• Supports multiple specific tasks including recognition, author choice, layout/style analysis, and bilingual interpretation.
• Includes metadata annotations for calligraphy authors, as mentioned in the description.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is known, but specific file formats, data splits, and license details are unknown.
• Data may reflect bias inherent to the specific collection of calligraphy samples used.

Provenance

Source

Hugging Face dataset by author gtang666.

Collection Method

Method of gathering is not described in the provided input.

Time Range

Temporal coverage is not specified.

Freshness

Last updated 2025-07-22 11:22:32; freshness should be verified.

Geography

Spatial coverage is not specified, but the subject is Chinese calligraphy.