Sketchy: Fashion Image Generation via Sketch-Text Pairing

Federicogirella's Sketchy dataset, adapted from Fashionpedia, provides multi-conditioning data for image generation. It was published for the ICCV25 paper 'LOTS of Fashion! Multi-Conditioning for Image Generation via Sketch-Text Pairing'. The dataset was last updated on September 18, 2025.

Use Cases

• Training sketch-to-image generation models based on the provided sketch-text pairs.
• Researching multi-conditioning techniques for image synthesis based on the dataset's described pairing.
• Fine-tuning generative models for fashion-specific content creation based on the adapted Fashionpedia data.

Strengths

• Dataset is associated with a peer-reviewed ICCV25 publication, suggesting academic rigor.
• Last updated on 2025-09-18, indicating recent maintenance.

Limitations

• Description metadata is limited; actual data quality requires manual inspection after download.
• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment.

Provenance

Source

Adapted from the Fashionpedia Dataset.

Collection Method

Computed data provided by the author; requires merging with the original Fashionpedia data.

Freshness

Last updated 2025-09-18 11:05:41