AIKavach Pre-Release v0.1.0

This is a pre-release version of Kavach.AI, an application for improving model certified radius with denoisers.

Features

Better Integration of DeNoising algorithms
Calculation of improved certified radius for models
Web interface for managing and analysing models
Improving the project problems and further Ideations and options

Usage

To run AIKavach, use the following command:

python app.py

Once the application is running, you can access it through your web browser at http://localhost:5000.

Known Issues

Issue 1: Currently a lot of model storage formats don’t work or are not supported.
Issue 2: The UI of the app can be further improved and there are some bugs in the same.
Issue 3: Currently supports only GPU runs and tests

Please note that this is a pre-release version of the app, and there may be some known issues and limitations.

Roadmap

The following are planned features and improvements for future releases:

Enhance DeNoising algorithms and adding support for the training and inference
Add support for different model types
Implement user authentication and access control
Adding a login system so that the app can be deployed on GPU hardware
Planning to sample deploy on Heroku or other platforms

Feedback and Contributions

We welcome your feedback and contributions to improve MyApp. If you encounter any issues or have suggestions, please create an issue in the GitHub repository.

Thanks to all the contributors !!! (Thanks in alphabetical order:)
@AK1405
@guptaaryan16
@FireShadow05
@apophis30
@Pradnya2203
@PriyanshuPansari
@mr-sarthakgupta