---
license: cc-by-4.0
tags:
- earth-observation
- remote-sensing
- copernicus
- sentinel
- multimodal
- major-tom
task_categories:
- image-to-image
---

![copgen-banner-github](https://cdn-uploads.huggingface.co/production/uploads/63ea69a55c837d9968ebecc0/rjWUAVt39Ewcmk0gv7D8u.png)

# COP-GEN Edinburgh Subset

[![arXiv](https://img.shields.io/badge/arXiv-2603.03239-D12424)](https://arxiv.org/abs/2603.03239)
[![GitHub](https://img.shields.io/badge/GitHub-COP--GEN-black?logo=github)](https://github.com/miquel-espinosa/COP-GEN)
[![Website](https://img.shields.io/badge/🌐-Website-grey)](https://miquel-espinosa.github.io/cop-gen/)
[![HF Model](https://img.shields.io/badge/🤗-Model-yellow)](https://huggingface.co/mespinosami/copgen-base)
[![HF VAEs](https://img.shields.io/badge/🤗-VAEs-yellow)](https://huggingface.co/mespinosami/copgen-vaes)
[![HF Benchmark](https://img.shields.io/badge/🤗-Benchmark-yellow)](https://huggingface.co/datasets/Major-TOM/COP-GEN-Benchmark)
[![HF Collection](https://img.shields.io/badge/🤗-Collection-yellow)](https://huggingface.co/collections/mespinosami/copgen)

A small pre-compiled subset of [Major-TOM](https://huggingface.co/Major-TOM) Copernicus data covering the Edinburgh region. Intended as a ready-to-use development and demonstration dataset for COP-GEN — no download scripting or preprocessing required. All modalities are spatially aligned and patchified.

## Modalities

S1RTC (VV, VH), S2L1C (all bands + cloud mask), S2L2A (all bands), DEM, LULC.

## Download

```bash
git clone https://huggingface.co/datasets/mespinosami/copgen-edinburgh-subset ./data/majorTOM/edinburgh
for f in ./data/majorTOM/edinburgh/*.zip; do
    unzip -q "$f" -d ./data/majorTOM/edinburgh/
done
rm ./data/majorTOM/edinburgh/*.zip
rm -rf ./data/majorTOM/edinburgh/.git ./data/majorTOM/edinburgh/.gitattributes
```

Once downloaded, you can run conditional or unconditional generation immediately using the [COP-GEN examples](https://github.com/miquel-espinosa/COP-GEN/tree/main/examples).

## Relationship to Other Resources

This subset is for local development and quick-start use only. For rigorous evaluation, use the [COP-GEN Benchmark](https://huggingface.co/datasets/Major-TOM/COP-GEN-Benchmark). For the full training pipeline, see the [GitHub README](https://github.com/miquel-espinosa/COP-GEN).

## Citation

```bibtex
@article{copgen2026,
    title   = {COP-GEN: Latent Diffusion Transformer for Copernicus Earth
               Observation Data},
    author  = {Espinosa, Miguel and Gmelich Meijling, Eva and Marsocci,
               Valerio and Crowley, Elliot J. and Czerkawski, Mikolaj},
    year    = {2026},
    journal = {arXiv preprint arXiv:2603.03239},
    url     = {https://arxiv.org/abs/2603.03239},
}
```