BroadcastDB PixelArt
Specialized dataset of 2000 pixel art images generated by Stable Diffusion with prompt users and associated hyperparameters.
Description
BroadcastDB PixelArt is a subset of the DiffusionDB dataset comprising 2000 pixel art images generated by Stable Diffusion. Each image is associated with a text prompt and hyperparameters, allowing detailed analyses of text-image relationships.
What is this dataset for?
- Studying the generation of pixel art images via generative AI
- Analyze the correspondence between text prompts and images produced
- Develop human-AI interaction tools for assisted artistic creation
Can it be enriched or improved?
Yes, this dataset can be enriched by adding new pixel art images with prompts, or by adding additional annotation of artistic styles and details.
🔎 In summary
🧠 Recommended for
- AI researchers
- Digital artists
- Generative model developers
🔧 Compatible tools
- Stable Diffusion
- PyTorch
- TensorFlow
- JSON/parquet readers
💡 Tip
Exploit hyperparameters to optimize the targeted generation of pixel art images.
Frequently Asked Questions
How many images does this pixel art dataset contain?
It contains 2000 pixel art images with their prompts and associated metadata.
What format do images and metadata take?
The images are in PNG format, the metadata is available in JSON and Parquet.
Can this dataset be used to train an image generation model?
Yes, it is suitable for training or refining pixel art image generation models based on text prompts.
