---
title: Spinning Up Ant Demo
emoji: 🌍
colorFrom: indigo
colorTo: purple
sdk: gradio
sdk_version: 5.37.0
app_file: app.py
pinned: false
license: mit
---
This Space provides an interactive demonstration of a deep reinforcement learning agent trained using a personally modernized and updated version of OpenAI's classic Spinning Up framework. The primary goal of this project was to revitalize this valuable educational resource, making it fully compatible with the modern Python ecosystem, including gymnasium and the open-source mujoco engine.

The agent employs the Soft Actor-Critic (SAC) algorithm, a state-of-the-art model-free algorithm known for its sample efficiency and stability in continuous control tasks. It was trained on the challenging Ant-v4 benchmark environment, where the objective is to learn a stable and effective locomotion gait for a four-legged robot.

To see the agent in action, simply click the "Run Agent Simulation" button below. The application will load the pre-trained PyTorch model, execute one full episode in the environment, and display the resulting performance as a video.

Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference