Xavier is an open-source toolbox that would test the quality of test suites of Deep Neural Networks by using a range of varied mutation operators increasing the overall accuracy of test suites. In the process of elaborative study on Mutation Testing of DL systems, Xavier will cater multiple model level mutation operators mentioned in literature. Xavier will be able to visualize deep neural networks along with their mutants and wil be capable to give extensive results drawn from mutation testing and their effects and results in depth. It would aid researchers in drawing a comparison between the original and mutated model.
This repository includes the backend devolped using FastAPI. To get the front-end app: https://github.com/iem-saad/xavier-frontend