bentoml/BentoDiffusion
📦 Open Source Projectbentoml
Production-ready diffusion model serving with BentoML, including fine-tuning and Kubernetes deployment.
BentoDiffusion is a comprehensive toolkit for serving diffusion models in production, built on top of BentoML. It includes ready-to-use services for popular models like Stable Diffusion and SDXL, with support for LoRA adapters and text-to-image generation. The project also provides fine-tuning scripts for customizing models on user datasets. It integrates seamlessly with Kubernetes, allowing scalable and reliable deployment. Key features include automatic model caching, dynamic batching, and REST API endpoints. The code is structured to be modular, enabling easy extension to new diffusion models. BentoDiffusion is designed for ML engineers and DevOps teams looking to deploy generative AI models efficiently.
💡Highlights
- ├─Serves SD, SDXL, LoRA
- ├─BentoML + Kubernetes
- └─Fine-tuning scripts included
🎯For
- ├─ML Engineers
- ├─DevOps Practitioners
- └─AI Researchers