Fetching the paper…

FastTD3: Simple, Fast, and Capable Reinforcement Learning for Humanoid Control · Around