LeX-10K: 10,000 High-Resolution Images for Text-to-Image Generation

LeX-10K is a curated collection of 10,000 high-resolution images created by X-ART and released on Hugging Face in April 2025. The dataset is designed for text-to-image generation tasks, with a focus on aesthetic quality and text fidelity. It contains 1024x1024 images described as visually diverse and stylistically rich.

Use Cases

• Training text-to-image generation models based on the described high-resolution, aesthetically focused images.
• Benchmarking model performance on text fidelity and aesthetic quality based on the dataset's stated focus.
• Fine-tuning image generation models for stylistic richness based on the described visual diversity.

Strengths

• Contains 10,000 images, providing a substantial sample size.
• Images are high-resolution at 1024x1024 pixels.
• Dataset is curated with a stated focus on aesthetics, text fidelity, and stylistic richness.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is known, but the specific distribution of styles or text prompts is not detailed.
• Data may reflect stylistic or source bias inherent to the curation process.

Provenance

Source

X-ART via Hugging Face.

Collection Method

Curated collection; specific gathering method is not detailed.

Time Range

Creation or release timeframe is not specified.

Freshness

Last updated 2025-04 06 15:25:53.

Geography

Spatial coverage is not specified.