dsgiitr / aikavach Goto Github PK

We need to return deployable binaries to our users with all the functionalities that the users opt for on top of the base classifier. The user could then download these from the web platform and use them.

Currently, the pipeline only takes 32x32-sized images only, making it generalizable.

Currently, the AIKavach platform can only be deployed on your local device for making it remotely deployable, we need a login system and more changes.

Currently, the model input and loading method breaks whenever the loaded model requires some different dependencies or needs to be loaded from model class and weight dicts. We need to make a general system to help input and load the models.

Convert the equation to LATEX and insert it into README.md at the same place as the image.

Making changes so that it can also run on cpu

Since we are already computing a large number of samples in a given neighbourhood, it would be more advantageous to use this for more applications that could be useful for the user.

Add multiple options for denoisers with varying computing requirements for different use-cases.