README.md

Over Paramerization project using VIT

How to use the code

1. install the environment.yml as conda environment
2. run vit_cifar.ipynb

1. Fraction of Recovered Examples

This image showcases various examples of corrupted and recovered image patches, highlighting different corruption patterns. Each subplot shows the corrupted fraction on the top and the corresponding recovered fraction below.

2. Iterative Recovery Process

The image illustrates the iterative recovery process of corrupted images. Starting from the left, the image shows the initial corrupted input followed by multiple iterations leading to a final restored image on the right.

3. Corrupted vs. Recovered Train Examples

This image compares corrupted and recovered training examples. The left column displays corrupted images with various shapes and patterns, while the right column shows their corresponding recovered versions.