Reinforcement learning pipeline for the Unitree G1 humanoid in Unity using ML-Agents. Includes URDF-based simulation, PPO/SAC training, configs, checkpoints, and experiment evidence for humanoid locomotion tasks.