Dataset · k-mktr
Cleaned collection of prompts for the FLUX.1 generator
k-mktr/improved-flux-prompts
A filtered set of English prompts for creating detailed images in the FLUX.1 model.
What's inside
The dataset contains 10 to 100k English text prompts created specifically for the FLUX.1 image generation model. The description includes prompting best practices for obtaining high-quality and creative images. The authors performed additional cleaning and filtering of records, removing extra noise.
What you can do with it
- Training language models and LoRA for automatic prompt improvement
- Using as a ready-made query database for image generation in FLUX.1
- Tasks
- Text classification
- Languages
- English
- Size
- 10K–100K rows
- Downloads this month
- 434
- Likes
- 64
Models trained on it
- trollek/ImagePromptHelper-danube3-500M-GGUF 357
- mradermacher/ImagePromptHelper-danube3-500M-GGUF 315
- pentacoxian-dev/iris-3b-turbo-loraText → image 63
- trollek/ImagePromptHelper-danube3-500MText generation 60
- trollek/SmolImagePromptHelper-135M-GGUF 42
- trollek/SmolImagePromptHelper-135MText generation 24
- Plugiloinc/plugilo2.1Text → image 0
- Plugiloinc/plugiloimage3Text → image 0
How to load it
from datasets import load_dataset
ds = load_dataset("k-mktr/improved-flux-prompts")