Tirthanker/TextAtlas5M

---
language:
- en
license: mit
size_categories:
- 1M<n<10M
task_categories:
- text-to-image
configs:
- config_name: CleanTextSynth
data_files:
- split: train
path: CleanTextSynth/train-*
- config_name: CoverBook
data_files:
- split: train
path: CoverBook/train-*
- config_name: LongWordsSubset-A
data_files:
- split: train
path: LongWordsSubset-A/train-*
- config_name: LongWordsSubset-M
data_files:
- split: train
path: LongWordsSubset-M/train-*
- config_name: PPT2Details
data_files:
- split: train
path: PPT2Details/train-*
- config_name: PPT2Structured
data_files:
- split: train
path: PPT2Structured/train-*
- config_name: Paper2Text
data_files:
- split: train
path: Paper2Text/train-*
- config_name: StyledTextSynth
data_files:
- split: train
path: StyledTextSynth/train-*
- config_name: TextScenesHQ
data_files:
- split: train
path: TextScenesHQ/train-*
- config_name: TextVisionBlend
data_files:
- split: train
path: TextVisionBlend/train-*
dataset_info:
- config_name: CleanTextSynth
features:
- name: image
dtype: image
- name: image_path
dtype: string
- name: annotation
dtype: string
splits:
- name: train
num_bytes: 133604812540.416
num_examples: 1907721
download_size: 138418775112
dataset_size: 133604812540.416
- config_name: CoverBook
features:
- name: image
dtype: image
- name: image_path
dtype: string
- name: annotation
dtype: string
splits:
- name: train
num_bytes: 8961075399.568
num_examples: 207566
download_size: 9142089037
dataset_size: 8961075399.568
- config_name: LongWordsSubset-A
features:
- name: image
dtype: image
- name: image_path
dtype: string
- name: annotation
dtype: string
splits:
- name: train
num_bytes: 20045725466.166
num_examples: 259897
download_size: 19707636636
dataset_size: 20045725466.166
- config_name: LongWordsSubset-M
features:
- name: image
dtype: image
- name: image_path
dtype: string
- name: annotation
dtype: string
splits:
- name: train
num_bytes: 66957433594.44
num_examples: 1250428
download_size: 74099567659
dataset_size: 66957433594.44
- config_name: PPT2Details
features:
- name: image
dtype: image
- name: image_path
dtype: string
- name: annotation
dtype: string
splits:
- name: train
num_bytes: 26226098275.87
num_examples: 298565
download_size: 25513899065
dataset_size: 26226098275.87
- config_name: PPT2Structured
features:
- name: image
dtype: image
- name: image_path
dtype: string
- name: annotation
dtype: string
splits:
- name: train
num_bytes: 71897207190.861
num_examples: 96401
download_size: 61182676048
dataset_size: 71897207190.861
- config_name: Paper2Text
features:
- name: image
dtype: image
- name: image_path
dtype: string
- name: annotation
dtype: string
splits:
- name: train
num_bytes: 231020768860.128
num_examples: 356658
download_size: 224999838265
dataset_size: 231020768860.128
- config_name: StyledTextSynth
features:
- name: image
dtype: image
- name: image_path
dtype: string
- name: annotation
dtype: string
splits:
- name: train
num_bytes: 578001343720.47
num_examples: 425826
download_size: 596967675485
dataset_size: 578001343720.47
- config_name: TextScenesHQ
features:
- name: image_path
dtype: string
- name: image
dtype: image
- name: annotation
dtype: string
- name: raw_text
dtype: string
splits:
- name: train
num_bytes: 9719924619.87
num_examples: 48935
download_size: 10296042626
dataset_size: 9719924619.87
- config_name: TextVisionBlend
features:
- name: image
dtype: image
- name: image_path
dtype: string
- name: annotation
dtype: string
splits:
- name: train
num_bytes: 43174392465.408
num_examples: 546829
download_size: 42595172061
dataset_size: 43174392465.408
---

# TextAtlas5M

This dataset is a training set for [TextAtlas](https://textatlas5m.github.io/).

Paper: https://huggingface.co/papers/2502.07870

**(All the data in this repo is uploaded :>)**
# Dataset subsets

Subsets in this dataset are CleanTextSynth, PPT2Details, PPT2Structured,LongWordsSubset-A,LongWordsSubset-M,Cover Book,Paper2Text,TextVisionBlend,StyledTextSynth and TextScenesHQ. The dataset features are as follows:

### Dataset Features

* `image (img)`: The GT image.
* `annotation (string)`: The input prompt used to generate the text.
* `image_path (string)`: The image name.

## CleanTextSynth
To load the dataset
```python
from datasets import load_dataset
ds = load_dataset("CSU-JPG/TextAtlas5M", "CleanTextSynth", split="train")
```

## PPT2Details
To load the dataset
```python
from datasets import load_dataset
ds = load_dataset("CSU-JPG/TextAtlas5M", "PPT2Details", split="train")
```

## PPT2Structured
To load the dataset
```python
from datasets import load_dataset
ds = load_dataset("CSU-JPG/TextAtlas5M", "PPT2Structured", split="train")
```

## LongWordsSubset-A
To load the dataset
```python
from datasets import load_dataset
ds = load_dataset("CSU-JPG/TextAtlas5M", "LongWordsSubset-A", split="train")
```

## LongWordsSubset-M
To load the dataset
```python
from datasets import load_dataset
ds = load_dataset("CSU-JPG/TextAtlas5M", "LongWordsSubset-M", split="train")
```

## Cover Book
To load the dataset
```python
from datasets import load_dataset
ds = load_dataset("CSU-JPG/TextAtlas5M", "CoverBook", split="train")
```

## Paper2Text
To load the dataset
```python
from datasets import load_dataset
ds = load_dataset("CSU-JPG/TextAtlas5M", "Paper2Text", split="train")
```

## TextVisionBlend
To load the dataset
```python
from datasets import load_dataset
ds = load_dataset("CSU-JPG/TextAtlas5M", "TextVisionBlend", split="train")
```

## StyledTextSynth
To load the dataset
```python
from datasets import load_dataset
ds = load_dataset("CSU-JPG/TextAtlas5M", "StyledTextSynth", split="train")
```

## TextScenesHQ
To load the dataset
```python
from datasets import load_dataset
ds = load_dataset("CSU-JPG/TextAtlas5M", "TextScenesHQ", split="train")
```

## Citation

If you found our work useful, please consider citing:
```
@article{wang2025textatlas5m,
title={TextAtlas5M: A Large-scale Dataset for Dense Text Image Generation},
author={Wang, Alex Jinpeng and Mao, Dongxing and Zhang, Jiawei and Han, Weiming and Dong, Zhuobai and Li, Linjie and Lin, Yiqi and Yang, Zhengyuan and Qin, Libo and Zhang, Fuwei and others},
journal={arXiv preprint arXiv:2502.07870},
year={2025}
}
```