---
license: creativeml-openrail-m
task_categories:
- text-to-image
configs:
- config_name: default
  data_files:
  - split: train
    path: data/*.parquet
dataset_info:
  features:
  - name: prompt
    dtype: string
  - name: image
    dtype: image
  splits:
  - name: train
    num_examples: 1005
language:
- en
tags:
- art
- ai-generated
- stable-diffusion
- midjourney
size_categories:
- 1K<n<10K
---

# 1k-AI-Images Dataset

A collection of over 1,000 high-quality AI-generated images paired with their original text prompts. This dataset is ideal for training reward models, analyzing prompt engineering, or evaluating text-to-image consistency.

## Dataset Structure

The dataset is organized into multiple Parquet shards located in the `data/` directory. All shards are automatically merged into a single `train` split.

### Data Fields

- **prompt**: The text string used to generate the image.
- **image**: The generated image (PIL.Image format) and binary images.

## Usage

You can load the dataset and access the prompts/images with the following code:

```python
from datasets import load_dataset

# Load the dataset in streaming mode to save local space
dataset = load_dataset("HyHorX/1k-AI-Images", split="train")

# Check the first entry
first_sample = dataset[0]
print(f"Prompt: {first_sample['prompt']}")
# first_sample['image'].show()