ReinforcementExamples

I am learning reinfocement learning based on pytorch, and will record the projects in this repository.

Environments and Install\

• As pytorch 2.6.0 needs python 3.9 or later, gym needs 3.7-3.10. Use python 3.10 is a good choice. And basic conda env is in pytorchgym.yml
  You can just use conda env create -f pytorchgym.yml to create the same env in your PC.
• Rules of gym.make.reset and render() are updated, so codes in YuxiLiu's cookbook need to be updated

Games Examples

• MountainCar
  This is a basic game. Code is form 肖智清Zhiqing Xiao. In this game, we can learn how to define an agent(without policy reinforcement, just an equation), how to use render() to show the process of each step, the interaction between agent and env.

Pytorch and neural network

• Note can be found in my website Learn with rookie. All code I wrote in this note can be found in the .ipynb files in this repositry.