Welcome to the official GitHub repository for the Pixiol project! The project aims to showcase the power of integrating LLMs and (Text -> Image) Diffusion Models for mass content creation. 🤖🎨
- Docker / Docker Compose: For containerization, ensuring that the app runs consistently across all environments. 🐳📦
- Huggingface Diffusers: For generating images from text, allowing us to create content dynamically and at scale. 🤖🎨
- Huggingface Hub: For miscellaneous models like the current NSFW classifier providing better content safety. 🤖🔍
- Contentful: As the headless CMS, providing a robust backend for content management that integrates seamlessly with our frontend. 📚🔗
- LangChain: For easy transfer between LLMs allowing better modularity and flexibility. 🤖🔗
- Ollama: For easily running LLMs locally, allowing us to generate content without having to rely on external services. 🤖🏠
- OpenAI API: Implemented, but not used due to cost (and Ollama + Mistral being free and similar in performance). 🤖💸
- Integration of LLMs and Text-to-Image Diffusion Models: This is the heart of Pixiol. The app aims to demonstrate how these AI models can be used to create and manage content dynamically and at scale. 🤖🎨
- Containerization: The app is containerized using Docker and Docker Compose, ensuring that it runs consistently across all environments. 🐳📦
- Headless CMS: The app uses Contentful as the headless CMS, providing a robust backend for content management that integrates seamlessly with our frontend. 📚🔗
- Modularity: The app is designed to be modular, allowing for easy transfer between LLMs and other models. 🤖🔗
- Local LLMs: The app uses Ollama to run LLMs locally, allowing us to generate content without having to rely on external services. 🤖🏠
- NSFW Classifier: The app uses a NSFW classifier to ensure that the generated content is safe for all audiences. 🤖🔍
👀 Looking for the frontend part? Check out Pixiol Front End (1/2) repository for more details on how we manage our site.
🙌 Stay tuned for updates and feel free to contribute or suggest improvements!
✨ Happy exploring! ✨