Description

A Proximal Policy Optimization (PPO) reinforcement learning experiment achieving successful swing-up and balance in a MuJoCo environment. This is the first entry in a series focused on autonomous rocket recovery, demonstrating highly efficient training (60M steps in 5.5 minutes) for just $0.06 on an NVIDIA L4 GPU via a Grok-integrated Colab setup.

Descriptions, tags, and model credits may be AI-generated or inferred from public sources and can be incomplete or wrong. Learn more.