gaodrew/midjourney-prompts-highquality
收藏资源简介:
--- license: apache-2.0 task_categories: - text-to-image tags: - midjourney - aiart - art size_categories: - 10K<n<100K --- <img src="https://cloud-3i4ld6u5y-hack-club-bot.vercel.app/0home.png" alt="Akash Network logo" width="400"/> Thank you to the [Akash Network](https://akash.network/) for sponsoring this project and providing A100s/H100s for compute! # About A filtered version of the [vivym/midjourney-prompts](https://huggingface.co/datasets/vivym/midjourney-prompts) dataset # Filtering criteria - top 10% in length (assuming that longer prompts = more effort and higher quality) - used on an image to be upscaled (assuming that users are more likely to upscale an image that is aesthetically pleasing) - used on midjourney version 5.0+ - deduplicated # Run yourself filter.py script provided (go to dataset files) so you can create your own dataset however you like
数据集概述
基本信息
- 许可证: Apache-2.0
- 任务类别: 文本到图像
- 标签:
- midjourney
- aiart
- art
- 大小类别: 10K<n<100K
数据集描述
- 来源: 该数据集是vivym/midjourney-prompts的一个过滤版本。
过滤标准
- 长度: 选取长度排名前10%的提示(假设较长提示代表更多努力和更高品质)
- 使用场景: 用于图像升级(假设用户更可能升级外观上吸引人的图像)
- 版本要求: 使用于midjourney 5.0及以上版本
- 去重处理: 数据集已进行去重
自定义使用
- 脚本提供: 提供
filter.py脚本,用户可根据需要创建自己的数据集。



