The Agibot X1, developed by Agibot, is a modular humanoid robot with high degrees of freedom, utilizing the Agibot open-source framework AimRT as middleware and employing reinforcement learning for motion control. This project includes the reinforcement learning training code used by Agibot X1, which can be combined with the inference software provided by Agibot X1 for real robots and simulation walking debugging, or importing other robot models for training.