Nunuvv/HumanEdit
收藏Hugging Face2026-04-16 更新2026-04-26 收录
下载链接:
https://hf-mirror.com/datasets/Nunuvv/HumanEdit
下载链接
链接失效反馈官方服务:
资源简介:
---
language:
- en
license: cc-by-4.0
size_categories:
- 1K<n<10K
task_categories:
- text-to-image
- image-to-image
pretty_name: HumanEdit
dataset_info:
features:
- name: IMAGE_ID
dtype: string
- name: EDITING_TYPE
dtype: string
- name: CORE
dtype: int32
- name: MASK
dtype: int32
- name: EDITING_INSTRUCTION
dtype: string
- name: OUTPUT_DESCRIPTION
dtype: string
- name: INPUT_CAPTION_BY_LLAMA
dtype: string
- name: OUTPUT_CAPTION_BY_LLAMA
dtype: string
- name: INPUT_IMG
dtype: image
- name: MASK_IMG
dtype: image
- name: OUTPUT_IMG
dtype: image
splits:
- name: train
num_bytes: 16682224174.369
num_examples: 5751
download_size: 16377096205
dataset_size: 16682224174.369
---
# Dataset Card for HumanEdit
[Paper](https://huggingface.co/papers/2412.04280) (CVPR 2025 AI for Content Creation (AI4CC) Workshop)
## Dataset Description
- **Homepage:** https://viiika.github.io/HumanEdit
- **Repository:** https://github.com/viiika/HumanEdit
- **Point of Contact:** [Jinbin Bai](mailto:jinbin.bai@u.nus.edu)
## Usage
```python
from datasets import load_dataset
from PIL import Image
# Load the dataset
ds = load_dataset("BryanW/HumanEdit")
# Print the total number of samples and show the first sample
print(f"Total number of samples: {len(ds['train'])}")
print("First sample in the dataset:", ds['train'][0])
# Retrieve the first sample's data
data_dict = ds['train'][0]
# Save the input image (INPUT_IMG)
input_img = data_dict['INPUT_IMG']
input_img.save('input_image.jpg')
print("Saved input image as 'input_image.jpg'.")
# Save the mask image (MASK_IMG)
mask_img = data_dict['MASK_IMG']
mask_img.save('mask_image.png') # Note that the format of the mask image may need to be adjusted. Refer to https://github.com/viiika/HumanEdit/mask_convert.py for more details.
print("Saved mask image as 'mask_image.png'.")
# Save the output image (OUTPUT_IMG)
output_img = data_dict['OUTPUT_IMG']
output_img.save('output_image.jpg')
print("Saved output image as 'output_image.jpg'.")
```
## License
Creative Commons License
This work is licensed under a Creative Commons Attribution 4.0 International License.
## Citation
If you find this work helpful, please consider citing:
```bibtex
@article{bai2024humanedit,
title={HumanEdit: A High-Quality Human-Rewarded Dataset for Instruction-based Image Editing},
author={Bai, Jinbin and Chow, Wei and Yang, Ling and Li, Xiangtai and Li, Juncheng and Zhang, Hanwang and Yan, Shuicheng},
journal={arXiv preprint arXiv:2412.04280},
year={2024}
}
```
---
语言:
- en(英语)
许可协议:cc-by-4.0
样本规模类别:
- 1K<n<10K
任务类别:
- 文本到图像(text-to-image)
- 图像到图像(image-to-image)
数据集展示名:HumanEdit
数据集信息:
特征:
- 名称:IMAGE_ID(图像ID),数据类型:string(字符串)
- 名称:EDITING_TYPE(编辑类型),数据类型:string(字符串)
- 名称:CORE(核心指标),数据类型:int32(32位整数)
- 名称:MASK(掩码标记),数据类型:int32(32位整数)
- 名称:EDITING_INSTRUCTION(编辑指令),数据类型:string(字符串)
- 名称:OUTPUT_DESCRIPTION(输出描述),数据类型:string(字符串)
- 名称:INPUT_CAPTION_BY_LLAMA(由LLAMA生成的输入标题),数据类型:string(字符串)
- 名称:OUTPUT_CAPTION_BY_LLAMA(由LLAMA生成的输出标题),数据类型:string(字符串)
- 名称:INPUT_IMG(输入图像),数据类型:image(图像)
- 名称:MASK_IMG(掩码图像),数据类型:image(图像)
- 名称:OUTPUT_IMG(输出图像),数据类型:image(图像)
划分:
- 名称:train(训练集),字节数:16682224174.369,样本数:5751
下载大小:16377096205
数据集总大小:16682224174.369
---
# HumanEdit 数据集卡片
[论文](https://huggingface.co/papers/2412.04280)(CVPR 2025 内容创作AI(AI4CC)研讨会)
## 数据集说明
- **主页:** https://viiika.github.io/HumanEdit
- **代码仓库:** https://github.com/viiika/HumanEdit
- **联系人:** [Jinbin Bai](mailto:jinbin.bai@u.nus.edu)
## 使用方法
python
from datasets import load_dataset
from PIL import Image
# 加载数据集
ds = load_dataset("BryanW/HumanEdit")
# 打印总样本数并展示首个样本
print(f"训练集总样本数:{len(ds['train'])}")
print("数据集中的首个样本:", ds['train'][0])
# 获取首个样本的数据
data_dict = ds['train'][0]
# 保存输入图像(INPUT_IMG)
input_img = data_dict['INPUT_IMG']
input_img.save('input_image.jpg')
print("已将输入图像保存为'input_image.jpg'。")
# 保存掩码图像(MASK_IMG)
mask_img = data_dict['MASK_IMG']
mask_img.save('mask_image.png') # 请注意掩码图像的格式可能需要调整,详情请参考https://github.com/viiika/HumanEdit/mask_convert.py。
print("已将掩码图像保存为'mask_image.png'。")
# 保存输出图像(OUTPUT_IMG)
output_img = data_dict['OUTPUT_IMG']
output_img.save('output_image.jpg')
print("已将输出图像保存为'output_image.jpg'。")
## 许可协议
知识共享许可协议
本作品采用知识共享署名4.0国际许可协议进行许可。
## 引用声明
若您认为本工作对您有所帮助,请引用以下文献:
bibtex
@article{bai2024humanedit,
title={HumanEdit: A High-Quality Human-Rewarded Dataset for Instruction-based Image Editing},
author={Bai, Jinbin and Chow, Wei and Yang, Ling and Li, Xiangtai and Li, Juncheng and Zhang, Hanwang and Yan, Shuicheng},
journal={arXiv preprint arXiv:2412.04280},
year={2024}
}
提供机构:
Nunuvv



