Dataset · m1guelpf
Nouns pixel characters with auto-descriptions
m1guelpf/nouns
Image collection of Nouns project characters with automatically generated English text descriptions.
What's inside
Contains 10k to 100k image-text pairs (train split only). Each row includes a PIL JPEG image of varying size and an English description automatically generated based on character attributes, colors, and items.
What you can do with it
- Training LoRA for generating Nouns pixel-style images
- Fine-tuning text-to-image models to understand composition from given attributes
- Testing automatic image captioning models
- Tasks
- Text → image
- Languages
- English
- Size
- 10K–100K rows
- Downloads this month
- 343
- Likes
- 37
Models trained on it
How to load it
from datasets import load_dataset
ds = load_dataset("m1guelpf/nouns")