This is the code repository for Hands-on Reinforcement Learning with PyTorch [Video], published by Packt. It contains all the supporting project files necessary to work through the video course from start to finish.
PyTorch, Facebook's deep learning framework, is clear, easy to code and easy to debug, thus providing a straightforward and simple experience for developers. This course is your hands-on guide to the core concepts of deep reinforcement learning and its implementation in PyTorch. We will help you get your PyTorch environment ready before moving on to the core concepts that encompass deep reinforcement learning. Following a practical approach, you will build reinforcement learning algorithms and develop/train agents in simulated OpenAI Gym environments. You'll learn the skills you need to implement deep reinforcement learning concepts so you can get started building smart systems that learn from their own experiences. By the end of this course, you will have enhanced your knowledge of deep reinforcement learning algorithms and will be confident enough to effectively use PyTorch to build your RL projects.
- Build key algorithms using PyTorch
- Implement self-learning agents using PyTorch
- Combine and modify Deep Q Networks and policy gradients to form more powerful algorithms
- Create actor-critic and deep deterministic policy gradients, and apply proximal policy
- Optimization in PyTorch and its extensions to improve performance
- Explore the importance of Q learning, sample efficiency, and the on/off policy in deep reinforcement learning
- Use function approximators, trust regions, and advanced value functions to build upon RL methods and drive new results
This hands-on course focuses on developing reinforcement learning concepts and algorithms with PyTorch and evaluating them with OpenAI gym environments. The course relates sections to each other, showing how the various key topics in deep RL build upon each other. The course material is practical; we elaborate on key concepts and show examples in code.
Minimum Hardware Requirements
For successful completion of this course, students will require the computer systems with at least the following:
OS: Ubuntu, Windows, macOS Processor: i3 Memory: 4GB Storage: 100 GB
Recommended Hardware Requirements
For an optimal experience with hands-on labs and other practical activities, we recommend the following configuration:
OS: Ubuntu, Windows, macOS Processor: i5 Memory: 8GB Storage: 500 GB/ 1 TB GPU: Any Nvidia GPU which has a compute capability greater than 8 GPU Memory: Greater than equal to 2GB Software Requirements Operating system: Ubuntu, Windows, macOS Browser: Any browser such as IE, Mozilla Firefox, Google Chrome Python version: 3.5+ (recommended: 3.6) PyTorch: 1.0.0 (https://pytorch.org/) Anaconda distribution (https://conda.io/docs/user-guide/install/download.html Python IDE: Spyder or Pycharm or any IDE which has iPython. Alternatively jupyter notebooks, jupyter lab, or Colab can be used. OpenAI Gym: latest available version (http://gym.openai.com/docs/)