This repository presents an in-depth analysis of various Open Source Diffusion models for generating photorealistic images based on given prompts. The evaluation criteria include execution time, CPU and RAM usage, photorealism, steerability, and resource optimization. The goal is to identify the most effective pipeline for producing high-quality images efficiently.
Diffusion models are a type of artificial intelligence (AI) model used for generating realistic images from textual prompts. These models work by gradually adding details to an image, starting from a low-quality version and refining it over multiple steps. By doing so, diffusion models can create high-quality, photorealistic images that closely match the given text description. They are particularly useful in various applications, such as generating images for storytelling, creating visual content for marketing, and enhancing image search capabilities.
To run this project, you need an Hugging Face Hub Token. To generate that, follow the instructions in the Hugging Face API documentation.
Make sure you have installed all the dependencies mentioned in the beginning of the IPython notebook.
For detailed analysis, please check the Open Source Diffusion Models Analysis Report.pdf available in this repository.