Reinforcement Learning - Full Course Using Python

This practical introduction teaches reinforcement learning with Python, Stable Baselines, and OpenAI Gym. Learners train and evaluate agents for environments such as CartPole, Atari, and CarRacing, and build a custom environment.

Syllabus

- Start
- Introduction
- Gameplan
- RL in a Nutshell
- 1. Setup Stable Baselines
- 2. Environments
- Loading OpenAI Gym Environments
- Understanding OpenAI Gym Environments
- 3. Training
- Train a Reinforcement Learning Model
- Saving and Reloading Environments
- 4. Testing and Evaluation
- Evaluating RL Models
- Testing the Agent
- Viewing Logs in Tensorboard
- Performance Tuning
- 5. Callbacks, Alternate Algorithms, Neural Networks
- Adding Training Callbacks
- Changing Policies
- Changing Algorithms
- 6. Projects
- Project 1 Atari
- Importing Dependencies
- Applying GPU Acceleration with PyTorch
- Testing Atari Environments
- Vectorizing Environments
- Save and Reload Atari Model
- Evaluate and Test Atari RL Model
- Updated Performance
- Project 2 Autonomous Driving
- Installing Dependencies
- Test CarRacing-v0 Environment
- Train Autonomous Driving Agent
- Save and Reload Self Driving model
- Updated Self Driving Performance
- Project 3 Custom Open AI Gym Environments
- Import Dependencies for Custom Environment
- Types of OpenAI Gym Spaces
- Building a Custom Open AI Environment
- Testing a Custom Environment
- Train a RL Model for a Custom Environment
- Save a Custom Environment Model
- 7. Wrap Up