Explanations of models for Hyperspectral data
Project description
☄️🛰️ Meteors
🛰️ Introduction
Meteors is an open-source package for creating explanations of hyperspectral and multispectral images. Developed primarily for Pytorch models, Meteors was inspired by the Captum library. Our goal is to provide not only the ability to create explanations for hyperspectral images but also to visualize them in a user-friendly way.
Please note that this package is still in the development phase, and we welcome any feedback and suggestions to help improve the library.
Meteors emerged from a research grant project between the Warsaw University of Technology research group MI2.ai and Kp Labs, financially supported by the European Space Agency (ESA).
🎯 Target Audience
Meteors is designed for:
- Researchers, data scientists, and developers who work with hyperspectral and multispectral images and want to understand the decisions made by their models.
- Engineers who build models for production and want to troubleshoot through improved model interpretability.
- Developers seeking to deliver better explanations to end users on why they're seeing specific content.
📦 Installation
Requirements
- Python >= 3.9
- PyTorch >= 1.10
- Captum >= 0.7.0
Install with pip
:
pip install meteors
With conda: Coming soon
📚 Documentation
Please refer to the documentation for more information on how to use Meteors.
🤝 Contributing
As an open-source project in a rapidly developing field, we are extremely open to contributions, whether it be in the form of a new feature, improved infrastructure, or better documentation.
We use rye as our project and package management tool. To start developing, follow these steps:
curl -sSf https://rye.astral.sh/get | bash # Install Rye
rye pin <python version >=3.9> # Pin the python version
rye sync # Sync the environment
Before pushing your changes, please run the tests and the linter:
rye test
rye run pre-commit run --all-files
For more information on how to contribute, please refer to our Contributing Guide.
Thank you for considering contributing to Meteors!
💫 Contributors
Project details
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
Built Distribution
File details
Details for the file meteors-0.0.2.tar.gz
.
File metadata
- Download URL: meteors-0.0.2.tar.gz
- Upload date:
- Size: 308.8 kB
- Tags: Source
- Uploaded using Trusted Publishing? No
- Uploaded via: twine/5.1.1 CPython/3.12.5
File hashes
Algorithm | Hash digest | |
---|---|---|
SHA256 | 38520b64ab4269552af37839fd9bbfdd1db49066ff844d0fc69cff33cf9963f1 |
|
MD5 | 7a43c83a8fb7df6ca5dcff94c2314557 |
|
BLAKE2b-256 | 5b8dd9f4b684758ddebd47a91112fe31f25fd6c801f65aa3e7c7db5703570529 |
File details
Details for the file meteors-0.0.2-py3-none-any.whl
.
File metadata
- Download URL: meteors-0.0.2-py3-none-any.whl
- Upload date:
- Size: 52.8 kB
- Tags: Python 3
- Uploaded using Trusted Publishing? No
- Uploaded via: twine/5.1.1 CPython/3.12.5
File hashes
Algorithm | Hash digest | |
---|---|---|
SHA256 | e1520d159912257ea074911751b50faaadef0fcd94b6543bc4d2568df190fdb3 |
|
MD5 | 20997485d2378f5f7f131ad385e6d28d |
|
BLAKE2b-256 | 816df7a2fe8cb58d8fb591e463dc27139b6fd6d227cac2fbd53a5e99d835a8c0 |