Welcome everyone to Sanskriti Bench. This is the open source version of our deployed application. The purpose of open sourcing the application (without the database) are the following:
- Make the processes as much transparent as possible.
- Welcome contributions from the community.
- We support the building and growing in public culture.
Please give it a star if you like this project. Thanks
First of it would be great if you could contribute in some languages in the main application. However if you want to contribute on technical terms, here is how you can contribute
- Add issues suggesting code level enhancements and improvements. The code is super early stage (main code hacked in one day) so, we expect more future improvements.
- Adding PR for the issues. If you feel you can contribute to the issue and enhance a feature or add a feature or fix a bug, that would be great.
- Suggest enhancement in typos. This means, if you feel there are some typo errors, feel free to add a PR.
- if you feel you can add video tutorials (in some native language) on how to use the app on phone or desktop, with different tools like multilingual keyboard or from word etc etc in the specific native language, let us know, we will pin the video in our Tutorial section. Currently there is one video but in English.
Running this app is simple. Just fork this app and then clone it and install the requirements and run the app.
pip install -r requirements.txt
streamlit run sanskriti_bench/app.py
Here is our simple roadmap:
- Release very first version of the app with users using it.
- Improvement cycle with more features or fixes in existing (if any), till we get enough amount of contributions and data point
- Start the developement cycle and making the tools for benchmarking using different LLMs, open and closed source.
- Publish results