AntonioAlgaida/DiffusionTrajectoryPlanner

A PyTorch implementation of a conditional Denoising Diffusion Probabilistic Model (DDPM) for multi-modal trajectory prediction. This project trains a U-Net on the Waymo Open Motion Dataset to generate plausible, human-like driving behaviors by denoising trajectories from pure Gaussian noise.

语言：Jupyter Notebook · GitHub 星标：45 · Fork：4

在 GitHub 上查看原仓库