Building a generative AI showroom for foundation models with different modalities

Multimodal foundation models represent a novel field of artificial intelligence that is gaining a lot attention and has a lot of potential in many application domains. In this work, we present how to build a generative AI showroom for single-or multimodal foundation models, which allows for the exploration and demonstration of the capabilities of these models. It provides a platform for users to interact with and understand the potential applications of generative AI technology across different modalities such as text, image, and audio. By showcasing the capabilities of these models, it helps in fostering innovation, collaboration, and understanding within the AI community. Additionally, the generative AI showroom serves as a valuable educational resource for both researchers and non-technical individuals, helping to bridge the gap between academia and industry. We provide details on the architecture and design decisions and present the showroom's features, performance and extensibility.