Datasets:
| dataset_info: | |
| features: | |
| - name: key | |
| dtype: string | |
| - name: image | |
| dtype: image | |
| - name: mask | |
| dtype: image | |
| - name: label | |
| dtype: | |
| class_label: | |
| names: | |
| '0': real | |
| '1': fake | |
| splits: | |
| - name: sd15 | |
| num_bytes: 34648812123.24868 | |
| num_examples: 200824 | |
| download_size: 34792882391 | |
| dataset_size: 34648812123.24868 | |
| configs: | |
| - config_name: default | |
| data_files: | |
| - split: sd15 | |
| path: data/sd15-* | |
| task_categories: | |
| - image-classification | |
| license: cc-by-sa-4.0 | |
| This repository contains the OpenSDI training dataset, presented in the paper [OpenSDI: Spotting Diffusion-Generated Images in the Open World](https://huggingface.co/papers/2503.19653). | |
| Code: https://github.com/iamwangyabin/OpenSDI |