Metadata-Version: 2.4
Name: pytorch-eo
Version: 2025.4.3
Summary: Deep Learning for Earth Observation
Author-email: earthpulse <it@earthpulse.es>
License-Expression: MIT
License-File: LICENSE
Requires-Python: >=3.12
Requires-Dist: albumentations>=2.0.5
Requires-Dist: einops>=0.8.1
Requires-Dist: lightning>=2.5.1
Requires-Dist: rasterio>=1.4.3
Requires-Dist: scikit-image>=0.25.2
Requires-Dist: scikit-learn>=1.6.1
Requires-Dist: torch>=2.6.0
Requires-Dist: torchmetrics>=1.7.0
Requires-Dist: torchvision>=0.21.0
Description-Content-Type: text/markdown

# Pytorch EO

Deep Learning for Earth Observation applications and research.

## Installation

```
pip install pytorch-eo
```

## Examples

Learn by doing with our [examples](https://github.com/earthpulse/pytorch_eo/tree/main/examples).

- [EuroSAT](examples/eurosat.ipynb).
- [UCMerced](examples/ucmerced.ipynb) Land Use Dataset.
- [BigEarthNet](examples/big_earth_net.ipynb).
- [SEN12FLOODs](examples/sen12floods.ipynb).

### Tutorials

Learn how to build with Pytorch EO with our [tutorials](https://github.com/earthpulse/pytorch_eo/tree/main/tutorials).

- Learn about [data loading](tutorials/00_data_loading.ipynb) is Pytorch EO.
- Learn about [data augmentation](tutorials/00_data_augmentation.ipynb) is Pytorch EO.
- Learn how to [create datasets](tutorials/02_creating_datasets.ipynb) with Pytorch EO.
- Learn about training models with our [tasks](tutorials/03_tasks.ipynb).

## Challenges

PytorchEO has been used in the following challenges:

- [EUROAVIA](./challenges/euroavia_hackathon_21) Mission: European Students Space Hackathon, 2021.
- [On Cloud N](./challenges/OnCloudN): Cloud Cover Detection Challenge (DrivenData, 2021).
- [Solar Panel Detection](./challenges/solar_panel_detection): Solar Panel Detection Using Sentinel-2 (Solafune, 2023).

<!-- ### Build your own Datasets

Using SCAN you can annotate your own data and access it directly through Pytorch EO. -->

<!-- ## Research

Pytorch EO can be a useful tool for research:

- Flexibility: build and experiment with new models for EO applications.
- Reproducibility: use same data splits and random seeds to compare with others.

See the [examples](https://github.com/earthpulse/pytorch_eo/tree/main/examples).

## Production

Pytorch EO was built with production in mind from the beginning:

- Optimize model for production.
- Export models to torchscript.
- Upload models to our Models Universe
- Use models directly through SPAI

See the [examples](https://github.com/earthpulse/pytorch_eo/tree/main/examples). -->

<!-- ## Documentation

Read our [docs](https://earthpulse.github.io/pytorch_eo/). -->

## Contributing

Read the [CONTRIBUTING](https://github.com/earthpulse/pytorch_eo/blob/main/CONTRIBUTING.md) guide.
