Urban Air Mobility (UAM) is an avation-based urban transportation that enables time-efficient passenger and cargo movement within complex low-altitude urban airspace. Vertiports are the ground infrastructes of this novel transporation system. Choosing the best location for vertiports in a city is one of the main challenges in developing this system. Previous studies have suggested different methods to solve this problem. This study takes the strengths of several existing methods and integrates them with a deep learning approach called Generative Modeling to propose a new framework for optimal vertiport placement. The core component of the proposed framework is the conditional Generative Adversarial Network (GAN), which learns complex and implicit spatial relationships between different factors and identifies highly suitable areas for vertiport construction. Subsequently, the refined suitability map created by the GAN is combined with the DBSCAN clustering technique to group feasible sites, and the resulting candidates have been optimized using a Location-Allocation model to choose the best locations with maximum service coverage and minimum travel time to Toronto airports. The results show that the proposed method can successfully integrate the UAM system with the existing transportation network and support the development of an integrated and interconnected urban transportation system. Furthermore, the identification of existing now-closed airport infrastructure as a viable vertiport location indicates that the results are coherent and aligned with real-world conditions.