Dataset · timm
259 thousand museum exhibits with attribute annotations (Metropolitan)
timm/imet-open-access
A large dataset of artworks from the Metropolitan Museum for multi-label classification by style and materials.
What's inside
The dataset contains 259,559 images of artworks and artifacts from the Metropolitan Museum's open archive (CC0), gathered based on Kaggle iMet competitions. Each object is assigned multiple text labels: thematic tags, culture, country of origin, materials, and execution technique. Suitable for fine-grained classification tasks with a long tail of rare classes.
What you can do with it
- Training multi-label classification and visual attribute recognition models (material, era, culture)
- Fine-tuning multimodal models (VLM) and visual search systems across museum and art collections
- Tasks
- Image classification
- Size
- 100K–1M rows
- Downloads this month
- 1.4K
- Likes
- 3
How to load it
from datasets import load_dataset
ds = load_dataset("timm/imet-open-access")