Nahrawy/FAID-Depth-ControlNet

---
dataset_info:
features:
- name: image
dtype: image
- name: depth_map
dtype: image
- name: scene
dtype: string
- name: caption
dtype: string
- name: state
dtype: string
splits:
- name: train
num_bytes: 11835627985.25
num_examples: 5550
download_size: 12139477164
dataset_size: 11835627985.25
---
# A Dataset of Flash and Ambient Illumination Pairs from the Crowd

This is a version of the [A Dataset of Flash and Ambient Illumination Pairs from the Crowd](http://yaksoy.github.io/flashambient/) dataset equipped for training ControlNet using depth maps conditioning.
The dataset includes 2775 pairs of flash light and ambient light images. It includes images of people, shelves, plants, toys, rooms and objects.

Captions were generated using the [BLIP-2, Flan T5-xxl](https://huggingface.co/Salesforce/blip2-flan-t5-xxl) model.
Depth maps were generated using the [GLPN fine-tuned on NYUv2 ](https://huggingface.co/vinvino02/glpn-nyu) model.

## Examples
## Disclaimer
I do not own any of this data.

原始信息汇总

数据集概述

数据集特征

• image: 图像类型
• depth_map: 图像类型
• scene: 字符串类型
• caption: 字符串类型
• state: 字符串类型

数据集划分

• train:
  • 示例数量: 5550
  • 数据大小: 11835627985.25字节

数据集大小

• 下载大小: 12139477164字节
• 数据集大小: 11835627985.25字节