Refiners is a Python library that simplifies working with foundation model adapters in PyTorch. It offers first-class support for popular adapter techniques including LoRA, ControlNet, IP-Adapter, T2I-Adapter, and textual inversion. The library focuses on diffusion models like Stable Diffusion and SDXL, enabling users to adapt these models for tasks such as text-to-image generation, segmentation (SAM), and image manipulation. It provides unified abstractions for model loading, training loops, and inference across these diverse adapter types.