I am working on implementing multiple stable diffusion models within a single endpoint. The goal is to dynamically load and unload these models based on the model ID sent through the API. This approach is necessary because processing/loading each model individually consumes all resources.