CalliBench: 3,192 Chinese Calligraphy Images with Recognition and VQA Annotations

3,192 image–annotation pairs for evaluating vision-language models on Chinese calligraphy. The dataset supports tasks like full-page recognition and contextual visual question answering, including author identification and bilingual interpretation. It was created by author gtang666 and last updated on Hugging Face in October 2025.

Use Cases

• Benchmarking VLM performance on Chinese calligraphy recognition based on the full-page recognition task.
• Training models for author and style attribution based on metadata annotations for calligraphy works.
• Developing bilingual interpretation systems for calligraphy based on the contextual VQA task for intent analysis.

Strengths

• 3,192 annotated image–text pairs provide a defined scale for evaluation.
• Annotations cover multiple complex tasks like VQA and recognition, suggesting detailed labeling.
• Dataset is hosted on a major platform (Hugging Face) with a specific author and recent update (2025-10-06).

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• The dataset's size in terms of file formats and total storage is unknown, which may limit suitability assessment.
• Data may reflect bias inherent to the specific collection of calligraphy works used.

Provenance

Source

huggingface

Freshness

Last updated 2025-10-06 09:25:48.