A2C on PandaReachDense-v3
This repository contains an Advantage Actor-Critic (A2C) agent trained using Stable-Baselines3.
Environment
- Name: PandaReachDense-v3
- Framework: panda-gym (PyBullet)
- Task: Reach a target position with a robotic arm
Algorithm
- Algorithm: Advantage Actor-Critic (A2C)
- Library: Stable-Baselines3
Training
The agent was trained in a headless environment (Google Colab).
Due to PyBullet rendering limitations, the evaluation video was generated separately.
Video
A replay of the trained agent is available in the videos/replay.mp4 file.
Author
maryna-13
