Today I landed my Lunar Lander on the moon using reinforcement learning 🌙
Reinforcement Learning is a computational approach of learning from actions.
Instead of manually controlling the inputs and directions of the thrusters, I used a Proximal Policy Optimization agent that is a mix of value-based RL and policy-based RL to predict actions leading to the greatest sum of rewards. https://huggingface.co/AGI-CEO/ppo-LunarLander-v2
I am learning this as part of Hugging Face's Deep RL Course (https://huggingface.co/learn/deep-rl-course/). It's a free, self-paced course.
I am planning on hosting group study sessions and create a free, open cohort to learn Deep Learning over the next 10 weeks.
I will host these study sessions and discussions on my AI Engineering / Entrepreneurship Discord group with 500+ members.
If you have some experience with Python and are interested in joining a group to learn Deep RL free of charge, please reach out to me. I am planning the first group session next Thursday at 10AM Eastern.
PPO Agent playing LunarLander-v2
This is a trained model of a PPO agent playing LunarLander-v2 using the stable-baselines3 library.
Usage (with Stable-baselines3)
TODO: Add your code
from stable_baselines3 import ...
from huggingface_sb3 import load_from_hub
...
- Downloads last month
- 1
Evaluation results
- mean_reward on LunarLander-v2self-reported260.63 +/- 20.34