Contributing to Yachay.ai
Are you interested in enhancing the geotagging models of Yachay.ai? We appreciate your consideration! Participating in Open Source Projects can be an enjoyable experience and can enhance your resume. Text-to-location language models have a potential for numerous applications in various projects. Feel free to explore our use-cases for inspiration.
There are multiple ways to contribute, such as improving the project's documentation to improving initial architecture. To get some inspiration on how to begin, take a look at our open call challenges and datasets issues.
To assist you in getting started, we have provided some guidelines.
About Yachay issue-challenges
You can see some issues in our repository are tagged challeges. This is an open call for improving architecture on a particular scenario/methodology that we think has a room for bright ideas. We are giving the curated annotated datasets on which we believe the model architecture can be improved with new design solutions.
There are many approaches you can take, and Yachay.ai core developers will be happy to discuss them in the issues and pull requests. You have the flexibility to consider or disregard input fields, predict coordinates, or forecast cluster regions, aligning with our team's validation metrics.
As we're building a community of developers, engineers and scientists of any level are welcome to share their solutions: let us know if you think there are specific dataset issues, want to suggest new SOTA models, ByT5 architecture improvements, fine-tuning suggestions, infrastructure ideas, or simply hit us up on discord with questions on the existing models!
Note: If you just want to share your results from model training we encourage you to make a comment in our issue tracker to get feedback from core developers. We are happy to discuss your approach, validation metrics, and advise on improvement!
Follow these steps to start contributing code:
- Create your own fork of the repository and clone it to your local machine.
git clone https://github.com/<YOUR-USERNAME>/byt5-geotagging.git cd byt5-geotagging git remote add upstream https://github.com/Yachay-AI/byt5-geotagging.git
- Create a new branch for your changes, and give it a concise name that reflects your contribution.
git checkout -b <BRANCH-NAME>
- Install the development dependencies in a Python environment and train your model
- Apply your modifications by creating small, separate commits that are thoroughly documented throughout the process.
- Commit your final changes.
- Push the changes to your fork.
- Create a pull request to the Yachay.ai repository! Make sure to include a description of your changes and link to a relevant issue.
Have a question? Feel free to opening an issue, you can also chat with the core team on our Discord server.
This project adheres to the BlockShop's Code of Conduct. By participating, you are expected to follow the Code Of Conduct.
By contributing, you agree that your contributions will be licensed under its MIT License.