anothy1/image-gen-instruct-360K

## Overview

This repository contains a subset of the [DiffusionDB](https://huggingface.co/datasets/poloclub/diffusiondb) dataset, enriched with synthetic generated instruction prompts.

## Description

- The original dataset is DiffusionDB.
- Synthetic instruction prompts were generated using the phi-3.8b-mini model on 2x RTX 3090 GPUs.
- The dataset was filtered to include only images and prompts with NSFW scores below 0.15.

## Credits

Special thanks to the curators of DiffusionDB:
- Jay Wang
- Evan Montoya
- David Munechika
- Alex Yang
- Ben Hoover
- Polo Chau

## Licensing

This dataset adheres to the same licensing terms as the DiffusionDB dataset.

- **DiffusionDB License:** [CC0 1.0 License](https://creativecommons.org/publicdomain/zero/1.0/)
- **Repository Code License:** MIT License

Please ensure to credit the original curators and follow the licensing terms.

This dataset is a subset of DiffusionDB, enriched with synthetic instruction prompts generated using the phi-3.8b-mini model on 2x RTX 3090 GPUs. The dataset has been filtered to include only images and prompts with NSFW scores below 0.15.

原始信息汇总

数据集概述

描述

• 原始数据集: DiffusionDB
• 数据处理: 使用phi-3.8b-mini模型在2x RTX 3090 GPUs上生成合成指令提示
• 数据过滤: 仅包含NSFW评分低于0.15的图像和提示

版权

• DiffusionDB许可证: CC0 1.0 License
• 代码许可证: MIT License