README.md

metadata

license: mit
pretty_name: Churches
tags:
  - image
  - computer-vision
  - buildings
  - architecture
task_categories:
  - image-classification
language:
  - en
configs:
  - config_name: default
    data_files: train/**/*.arrow
    features:
      - name: image
        dtype: image
      - name: unique_id
        dtype: string
      - name: width
        dtype: int32
      - name: height
        dtype: int32
      - name: image_mode_on_disk
        dtype: string
      - name: original_file_format
        dtype: string
  - config_name: preview
    data_files: preview/**/*.arrow
    features:
      - name: image
        dtype: image
      - name: unique_id
        dtype: string
      - name: width
        dtype: int32
      - name: height
        dtype: int32
      - name: original_file_format
        dtype: string
      - name: image_mode_on_disk
        dtype: string

Churches

High resolution image subset from the Aesthetic-Train-V2 dataset, a collection of facades, interior shots and landscapes.

Dataset Details

• Curator: Roscosmos
• Version: 1.0.0
• Total Images: 780
• Average Image Size (on disk): ~5.8 MB compressed
• Primary Content: Church buildings
• Standardization: All images are standardized to RGB mode and saved at 95% quality for consistency.

Dataset Creation & Provenance

1. Original Master Dataset

This dataset is a subset derived from: zhang0jhon/Aesthetic-Train-V2

• Link: https://huggingface.co/datasets/zhang0jhon/Aesthetic-Train-V2
• Providence: Large-scale, high-resolution image dataset, refer to its original dataset card for full details.
• Original License: MIT

2. Iterative Curation Methodology

CLIP retrieval / manual curation.

Who are the source data producers?

• Original Dataset Creators: Refer to the original dataset card.
• Curator and Refiner: Roscosmos

Dataset Structure & Content

This dataset offers the following configurations/subsets:

• Default (Full train data) configuration: Contains the full, high-resolution image data and associated metadata. This is the recommended configuration for model training and full data analysis. The default split for this configuration is train.

• preview configuration: Contains a small, random subset of images from the train data. The images in this configuration are downsampled and re-compressed to be viewer-compatible on the Hugging Face Hub. The default split for this configuration is train (if no other split is specified). Each example (row) in the dataset contains the following fields:

• image: The actual image data. In the default (full) configuration, this is full-resolution. In the preview configuration, this is a viewer-compatible version.

• unique_id: A unique identifier assigned to each image.

• width: The width of the image in pixels (from the full-resolution image).

• height: The height of the image in pixels (from the full-resolution image).

Usage

To download and load this dataset from the Hugging Face Hub:

from datasets import load_dataset, Dataset, DatasetDict

# Login using e.g. `huggingface-cli login` to access this dataset

# To load the full, high-resolution dataset (recommended for training):
# This will load the 'default' configuration's 'train' split.
ds_main = load_dataset("ROSCOSMOS/Churches", "default")

print("Main Dataset (default config) loaded successfully!")
print(ds_main)
print(f"Type of loaded object: {type(ds_main)}")

if isinstance(ds_main, Dataset):
    print(f"Number of samples: {len(ds_main)}")
    print(f"Features: {ds_main.features}")
elif isinstance(ds_main, DatasetDict):
    print(f"Available splits: {list(ds_main.keys())}")
    for split_name, dataset_obj in ds_main.items():
        print(f"  Split '{split_name}': {len(dataset_obj)} samples")
        print(f"  Features of '{split_name}': {dataset_obj.features}")

# To load the smaller, viewer-compatible preview data (if available):
# This will load the 'preview' configuration's default split (often also 'train').
# Check your dataset card for exact config and split names.
# try:
#     ds_preview = load_dataset("{push_to_hub_id}", "preview") 
#     print("\nPreview Dataset (preview config):")
#     print(ds_preview)
#     print(f"Number of samples in the preview dataset: {len(ds_preview) if isinstance(ds_preview, Dataset) else 'N/A'}") 
# except ValueError as e:
#     print(f"\nPreview config not found or failed to load: {e}") 

# To access specific splits from a DatasetDict:
# my_train_data = ds_main['train']
# my_preview_data = ds_preview['train'] # if preview loads as DatasetDict

# The 'image' column will contain PIL Image objects.

Citation

@inproceedings{zhang2025diffusion4k,
    title={Diffusion-4K: Ultra-High-Resolution Image Synthesis with Latent Diffusion Models},
    author={Zhang, Jinjin and Huang, Qiuyu and Liu, Junjie and Guo, Xiefan and Huang, Di},
    year={2025},
    booktitle={IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)},
}
@misc{zhang2025ultrahighresolutionimagesynthesis,
    title={Ultra-High-Resolution Image Synthesis: Data, Method and Evaluation},
    author={Zhang, Jinjin and Huang, Qiuyu and Liu, Junjie and Guo, Xiefan and Huang, Di},
    year={2025},
    note={arXiv:2506.01331},
}

Disclaimer and Bias Considerations

Please consider any inherent biases from the original dataset and those potentially introduced by the automated filtering (e.g., CLIP's biases) and manual curation process.

Contact

N/A