Development of a WEB Application for Generating Tattoo Images Based on a Promt

Generative models represent a class of neural models that have gained significant popularity today. These models open new possibilities for the creation and transformation of data in various fields. By training on large volumes of data, they can create new, previously non-existent samples, whether they be images, texts, music, or other types of data. Due to the rapid development of this area in neural networks, there has been a growing interest in understanding the architectures of generative models and developing useful web applications based on these architectures. The task in this work was to create a service for generating tattoo images based on a user's text request. It was necessary to study models capable of producing high-quality results that corresponded to the user's text requests without losing the given semantics and that could be trained in a short time. Therefore, a review of existing generative models that work with text requests was conducted. As a result, several models that met the specified criteria were selected, forming the basis for the web application. The developed application needed to possess important characteristics such as scalability, distribution, and asynchronous operation. This work describes the technologies and libraries used to create the service. As a result, a service was created that generates tattoos based on user requests and can deliver the necessary high-quality results in a short time.