Add Diffusion Model

This skill provides a comprehensive guide for adding various types of diffusion models, including text-to-image, text-to-video, and more, to the vLLM-Omni framework. It covers the necessary steps for model classification, adaptation, and integration, ensuring support for advanced features like Cache-DiT acceleration and multi-GPU parallelism. Ideal for developers looking to enhance vLLM-Omni with custom or existing models from HuggingFace Diffusers or private repositories.