Polyvore Outfits Dataset

Overview

The Polyvore Outfits Dataset is a large-scale fashion dataset curated by real users on the Polyvore platform. Each outfit consists of manually constructed combinations of items that go well together, reflecting authentic human styling choices rather than synthetically or algorithmically generated sets.

This human-driven curation captures implicit knowledge of color harmony, stylistic coherence, and garment compatibility, making the dataset well suited for research in similarity and compatibility learning, visual search and retrieval, recommendation systems, and multimodal reasoning. As a result, the dataset has been widely adopted in research across computer vision, recommendation systems, and multimodal learning.

Dataset Description

• 68,306 user-created outfits composed of 261,058 heterogeneous fashion items
• Rich metadata consisting of item descriptions, categories, and outfit groupings
• Benchmarks for compatibility prediction, fill-in-the-blank (FITB) outfit completion, and multimodal retrieval
• Supports research in both similarity and compatibility reasoning

Two types of evaluation splits are provided to benchmark performance:

• Non-disjoint (outfit-level): an easier setting where no outfits overlap between training and test sets, though individual items may appear in both.
• Disjoint (item-level): a more challenging and recommended split with no overlap of individual items between training and test sets; it evaluates the model's ability to generalize to completely unseen fashion items.

Intended Use

This dataset is intended for research in:

• Fashion similarity and compatibility learning
• Outfit modeling and generation
• Visual recommendation systems
• Visual search and retrieval
• Metric learning and representation learning
• Contrastive learning
• Vision and language learning
• Multimodal reasoning

License

BibTeX

@inproceedings{vasileva2018learning,
            title={Learning Type-Aware Embeddings for Fashion Compatibility},
            author={Vasileva, Mariya I. and Plummer, Bryan A. and Dusad, Krishna and Rajpal, Shreya and Kumar, Ranjitha and Forsyth, David A.},
            booktitle={European Conference on Computer Vision (ECCV)}, 
            year={2018}
        }