Sign in to view source links and access this dataset
Description
FashionMV is a large-scale dataset for product-level Composed Image Retrieval (CIR) created by yuandaxia. It contains 127,000 products, 472,000 multi-view images, and over 220,000 CIR triplets, built through an automated pipeline leveraging large multimodal models. The dataset was last updated on April 14, 2026.
Use Cases
Train composed image retrieval models based on the 220,000+ CIR triplets.
Benchmark multi-view fashion image understanding based on the 472,000 images.
Develop product search systems based on the 127,000 product entries.
Fine-tune large multimodal models for fashion applications based on the dataset's automated pipeline.
Strengths
Contains 127,000 distinct fashion products.
Includes 472,000 multi-view images.
Provides over 220,000 CIR triplets for training and evaluation.
Limitations
Column-level documentation is absent; field semantics must be inferred after download.
Row count is unknown, which may limit suitability assessment.
Description metadata is limited; actual data quality requires manual inspection after download.
Provenance
Source
yuandaxia on Hugging Face.
Collection Method
Built through a fully automated pipeline leveraging large multimodal models.
Time Range
null
Freshness
Last updated 2026-04-14 14:42:01; freshness should be verified.
Geography
null
License is unknown; terms of use must be verified before application.