Hugging Face has introduced its first custom Deep Reinforcement Learning environment, Snowball Fight 1vs1, built with Unity ML-Agents. In this game, players shoot snowballs against an AI agent trained with reinforcement learning. The game is hosted on Hugging Face Spaces and available to play online.
This release is part of a broader ecosystem Hugging Face is developing for Unity ML-Agents researchers and enthusiasts. The ecosystem includes:
- Custom environments: Open-source environments like Snowball Fight are shared on the Hugging Face Hub for experimentation.
- Model hosting: Users can save and share trained models on the Hub. The Snowball Fight training environment is already available here.
- Spaces demos: Researchers can quickly deploy interactive demos using Hugging Face Spaces.
To foster community, Hugging Face has launched a Discord server with dedicated channels for Deep Reinforcement Learning and ML-Agents.
Upcoming developments include technical tutorials on ML-Agents, a collaborative 2vs2 Snowball Fight version using the MA-POCA algorithm, and additional custom environments on the Hub.
"We're excited to see what you're working on with ML-Agents and how we can build features and tools that help you to empower your work," the team said.
Join the Discord server to stay updated on new features and environments.