Reward fine-tuning for Stable Diffusion models based on stochastic optimal control, including Adjoint Matching - View it on GitHub