---
license: cc-by-nc-sa-4.0
task_categories:
- image-classification
tags:
- synthetic-image-detection
- arrow
configs:
- config_name: default
  default: true
  data_files:
  - split: train
    path: "data/train/*/*.arrow"
  - split: validation
    path: "data/test/*/*.arrow"
  - split: test
    path: "data/test/*/*.arrow"
- config_name: all-test
  data_files:
  - split: test
    path: "data/test/*/*.arrow"
- config_name: all-train
  data_files:
  - split: train
    path: "data/train/*/*.arrow"
- config_name: all-validation
  data_files:
  - split: validation
    path: "data/test/*/*.arrow"
- config_name: adm
  data_files:
  - split: train
    path: "data/train/ADM/*.arrow"
  - split: validation
    path: "data/test/ADM/*.arrow"
  - split: test
    path: "data/test/ADM/*.arrow"
- config_name: adm-train
  data_files:
  - split: train
    path: "data/train/ADM/*.arrow"
- config_name: adm-test
  data_files:
  - split: test
    path: "data/test/ADM/*.arrow"
- config_name: adm-validation
  data_files:
  - split: validation
    path: "data/test/ADM/*.arrow"
- config_name: biggan
  data_files:
  - split: train
    path: "data/train/BigGAN/*.arrow"
  - split: validation
    path: "data/test/BigGAN/*.arrow"
  - split: test
    path: "data/test/BigGAN/*.arrow"
- config_name: biggan-train
  data_files:
  - split: train
    path: "data/train/BigGAN/*.arrow"
- config_name: biggan-test
  data_files:
  - split: test
    path: "data/test/BigGAN/*.arrow"
- config_name: biggan-validation
  data_files:
  - split: validation
    path: "data/test/BigGAN/*.arrow"
- config_name: glide
  data_files:
  - split: train
    path: "data/train/glide/*.arrow"
  - split: validation
    path: "data/test/glide/*.arrow"
  - split: test
    path: "data/test/glide/*.arrow"
- config_name: glide-train
  data_files:
  - split: train
    path: "data/train/glide/*.arrow"
- config_name: glide-test
  data_files:
  - split: test
    path: "data/test/glide/*.arrow"
- config_name: glide-validation
  data_files:
  - split: validation
    path: "data/test/glide/*.arrow"
- config_name: midjourney
  data_files:
  - split: train
    path: "data/train/Midjourney/*.arrow"
  - split: validation
    path: "data/test/Midjourney/*.arrow"
  - split: test
    path: "data/test/Midjourney/*.arrow"
- config_name: midjourney-train
  data_files:
  - split: train
    path: "data/train/Midjourney/*.arrow"
- config_name: midjourney-test
  data_files:
  - split: test
    path: "data/test/Midjourney/*.arrow"
- config_name: midjourney-validation
  data_files:
  - split: validation
    path: "data/test/Midjourney/*.arrow"
- config_name: sd14
  data_files:
  - split: train
    path: "data/train/stable_diffusion_v_1_4/*.arrow"
  - split: validation
    path: "data/test/stable_diffusion_v_1_4/*.arrow"
  - split: test
    path: "data/test/stable_diffusion_v_1_4/*.arrow"
- config_name: sd14-train
  data_files:
  - split: train
    path: "data/train/stable_diffusion_v_1_4/*.arrow"
- config_name: sd14-test
  data_files:
  - split: test
    path: "data/test/stable_diffusion_v_1_4/*.arrow"
- config_name: sd14-validation
  data_files:
  - split: validation
    path: "data/test/stable_diffusion_v_1_4/*.arrow"
- config_name: sd15
  data_files:
  - split: train
    path: "data/train/stable_diffusion_v_1_5/*.arrow"
  - split: validation
    path: "data/test/stable_diffusion_v_1_5/*.arrow"
  - split: test
    path: "data/test/stable_diffusion_v_1_5/*.arrow"
- config_name: sd15-train
  data_files:
  - split: train
    path: "data/train/stable_diffusion_v_1_5/*.arrow"
- config_name: sd15-test
  data_files:
  - split: test
    path: "data/test/stable_diffusion_v_1_5/*.arrow"
- config_name: sd15-validation
  data_files:
  - split: validation
    path: "data/test/stable_diffusion_v_1_5/*.arrow"
- config_name: vqdm
  data_files:
  - split: train
    path: "data/train/VQDM/*.arrow"
  - split: validation
    path: "data/test/VQDM/*.arrow"
  - split: test
    path: "data/test/VQDM/*.arrow"
- config_name: vqdm-train
  data_files:
  - split: train
    path: "data/train/VQDM/*.arrow"
- config_name: vqdm-test
  data_files:
  - split: test
    path: "data/test/VQDM/*.arrow"
- config_name: vqdm-validation
  data_files:
  - split: validation
    path: "data/test/VQDM/*.arrow"
- config_name: wukong
  data_files:
  - split: train
    path: "data/train/wukong/*.arrow"
  - split: validation
    path: "data/test/wukong/*.arrow"
  - split: test
    path: "data/test/wukong/*.arrow"
- config_name: wukong-train
  data_files:
  - split: train
    path: "data/train/wukong/*.arrow"
- config_name: wukong-test
  data_files:
  - split: test
    path: "data/test/wukong/*.arrow"
- config_name: wukong-validation
  data_files:
  - split: validation
    path: "data/test/wukong/*.arrow"
---

# GenImage Arrow

Generator- and split-partitioned Arrow release of the GenImage benchmark.
Each leaf directory is also a standalone Hugging Face `save_to_disk` bundle.

## Contents

- Train: 2,581,150 valid images across eight generators.
- Test: 100,000 images across eight generators.
- Validation is an alias of test because the official GenImage `val` directory
  is the benchmark test set. It is not a third independent split.
- Seventeen unavailable or zero-byte upstream train entries are documented in
  `manifest.json`.

## Quick start

```python
from datasets import load_dataset

test = load_dataset("nebula/GenImage-arrow", "all-test", split="test")
train = load_dataset("nebula/GenImage-arrow", "all-train", split="train")
biggan = load_dataset("nebula/GenImage-arrow", "biggan-train", split="train")
adm_test = load_dataset("nebula/GenImage-arrow", "adm-test", split="test")
biggan_all = load_dataset("nebula/GenImage-arrow", "biggan")
```

Configuration names follow these rules:

- `all-train`, `all-test`, and `all-validation` load every generator for one
  split.
- `adm`, `biggan`, `glide`, `midjourney`, `sd14`, `sd15`, `vqdm`, and
  `wukong` expose all available splits for one generator.
- Add `-train`, `-test`, or `-validation` to download only one split of one
  generator, for example `sd15-test`.

The `*-validation` configurations point to the same physical Arrow shards as
their `*-test` counterparts and therefore do not duplicate storage.

## Download a standalone Arrow bundle

Every `data/<split>/<generator>` directory is a complete Hugging Face
`Dataset.save_to_disk` bundle. Download only the leaf directory you need, then
open it with `load_from_disk`:

```python
from pathlib import Path

from datasets import load_from_disk
from huggingface_hub import snapshot_download

snapshot = Path(
    snapshot_download(
        repo_id="nebula/GenImage-arrow",
        repo_type="dataset",
        allow_patterns="data/test/ADM/*",
    )
)
adm_test = load_from_disk(snapshot / "data/test/ADM")
print(len(adm_test))  # 12000
```

Change both `test` and `ADM` in the pattern to select a different physical
split or generator. Exact generator directory names are:

```text
ADM
BigGAN
glide
Midjourney
stable_diffusion_v_1_4
stable_diffusion_v_1_5
VQDM
wukong
```

## Recommended toolbox: AID

For synthetic-image detector training and evaluation, use our
[iamwangyabin/AID](https://github.com/iamwangyabin/AID) toolbox:

```bash
git clone https://github.com/iamwangyabin/AID.git
cd AID
```

This dataset release includes AID compatibility indexes. Each train leaf
contains `train.json`; each test leaf contains `test.json`. These are small
indexes over the existing Arrow rows and do not contain duplicate image data.

Download a leaf bundle as shown above, then configure AID's existing
`data.ArrowDatasets` loader. No AID loader code changes are required:

```yaml
datasets:
  source:
    - target: data.ArrowDatasets
      data_root: /path/to/downloaded-snapshot/data/test/ADM
      sub_sets: [ADM]
      split: test
      benchmark_name: GenImage
```

The `data_root` must point to a generator leaf, not the repository root. The
single value in `sub_sets` must exactly match that leaf's generator name. To
evaluate multiple generators, add one source entry per generator and change
both fields together. For training, use `data/train/<generator>`, its exact
generator name, and `split: train`.

For AID, use physical `split: test` for the official GenImage validation/test
partition. The Hub's `validation` split is only an alias of those same test
shards, so there is intentionally no separate `validation.json` compatibility
index.

This compatibility path applies to AID configurations that use
`data.ArrowDatasets`. The detector named AIDE uses its own
`AIDEBinaryJsonDatasets` preprocessing contract and is not covered by this
drop-in configuration.

## Schema

Every row contains `image_path`, embedded `image` bytes, binary `label`, and
`generator`. Train rows additionally retain source `md5`, `width`, and `height`
metadata. Binary labels use `0` for real images and `1` for generated images.

## License and source

The original GenImage dataset is released under CC BY-NC-SA 4.0 with
additional dataset terms restricting use to non-commercial purposes. See the
upstream license and project for the authoritative terms:

- https://github.com/GenImage-Dataset/GenImage/blob/main/License
- https://github.com/GenImage-Dataset/GenImage