$1.$discuss the various metrics and methods used for evaluating the performance and output quality of GANs including cycleGAN and C$-$GAN.

How do these metrics address the challenges inherent in assessing generative models $?$

$2.$examine the common issues faced during the training of GANs, such as mode collapse and non$-$convergence.Propose strategies and modifications to the training process that can help mitigate the issues with examples from cycle GAN auxiliary classisfier GAN $($ACGAN$)$ and C$-$GAN

$3.$ explore the architectural variations in transformer models that have been proposed to improve performance in specific tasks such as large scale language models or fine grained classification tasks. How do these variations modify the original transformer architecture and what benefits do they offer $?$