datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
Stable-diffusion-configsstable_diffusion_instructional_dataset
Stable Diffusion Dataset
Description:
This dataset is in Jsonl format and is based on the MadVoyager/stable_diffusion_instructional_dataset.
Overview:
The Stable Diffusion Dataset comprises approximately 80,000 meticulously curated prompts sourced from the image finder of Stable Diffusion: "Lexica.art". The dataset is intended to facilitate training and fine-tuning of various language models, including LLaMa2.
Key Features:
◉ Jsonl format for… See the full description on the dataset page: https://huggingface.co/datasets/lusstta/stable_diffusion_instructional_dataset.stable-diffusion-prompt-pairsWork in progress. A dataset for creating image generation tags from natural language descriptions.
Uses https://huggingface.co/Gustavosta/MagicPrompt-Stable-Diffusion for tags. Descriptions generated by chronos-hermes-13b-v2.
Please note that the dataset is generated in two batches, with different system prompts. The first is ~2000 rows. The second ~1000 rows.
stable-diffusion-trainingStable_Diffusion_Prompt_MultiFilesStable-diffusion-3.5-Batchfine-tune_Stablediffusion_SandsofAI_MA