Dataset · SII-Monument-Valley
Visual question answering on antique Chinese porcelain
SII-Monument-Valley/CiQi-VQA
Multimodal dataset with images and English/Chinese questions for Chinese porcelain appraisal.
What's inside
A specialized dataset of 10K to 100K records created to train the CiQi-Agent multimodal agent (accepted to ECCV 2026). Contains images and English and Chinese question-answer pairs covering antique Chinese porcelain appraisal, fine-grained visual perception, and cultural heritage knowledge.
What you can do with it
- Fine-tuning VLMs and LoRAs for art appraisal and antique attribution
- Testing visual reasoning of multimodal agents in a specialized cultural domain
- Tasks
- Question answering, Image text → text
- Languages
- English, Chinese
- Size
- 10K–100K rows
- Downloads this month
- 89K
- Likes
- 6
Models trained on it
How to load it
from datasets import load_dataset
ds = load_dataset("SII-Monument-Valley/CiQi-VQA")