Home > Categories > Robotics > Robot Control > Unity ML-Agents

Unity ML-Agents

Rating:

4.5 / 5.0

Tags

unity, machine learning, ai, reinforcement learning, imitation learning, simulation, game development, open-source, python, tensorflow, pytorch

Categories:
GamingMachine learning and neural networksRoboticsScientific Research
Creator Unity Technologies
Date 2017-09-19
Platforms Windows, macOS, Linux
Status Active
Website unity.com
Price Model Free
Sections:
Model TrainingModeling and SimulationNPC and Enemy BehaviorRobot Control

Pricing Details

Free (open-source). Requires a Unity license (free and paid versions available).

Features

- Support for various machine learning algorithms (PPO, SAC, Imitation Learning, etc.).
- Python API for controlling the Unity environment.
- Ability to train single-agent, multi-agent cooperative, and competitive scenarios.
- Curriculum Learning.
- Environment randomization.
- On-Demand Decision Making.
- Support for visual and vector observations.
- Integration with machine learning frameworks like TensorFlow and PyTorch.

Integrations

Python API, TensorFlow, PyTorch, Hugging Face, OpenAI Gym.

Preview

Open-source toolkit for training intelligent agents in simulations and games using machine learning methods such as reinforcement learning.