Training a Bipedal Walker using Deep Deterministic Policy Gradient algorithm

This is a project of DeepRL , I have used the DDPG algorithm on the gymnassium environment "bipredal walker" of box2D environment . (visit gymnassium here)

---

📂 Contents :

• DDPG/ddpg_bipedal.py contains the training code.
• DDPG/new_models contains the trained models on different episodes.
• DDPG/evaluation.py evaluates the saved models on the agent.

---

Setup

• Windows configeration🪟 :
  • Install miniconda from here .
  • Install visual studio Build tools 2022 (Desktop Development with C++) from here.
  • Get into base environment and create a new environment
    • conda install swig
    • pip install gymnasium[box2d]

📹 Demo