1.0.0

UnrealMLAgents 1.0.0 (2025-02-11)

Features

• Introduced core Unreal ML Agents Plugin to enable reinforcement learning (RL) in Unreal Engine.
• Implemented Python API integration for seamless communication with training workflows.
• Added vector-based observations and RaySensor components for improved perception.
• Provided Blueprint-friendly design, allowing customization without C++.
• Enabled Multi-Agent support for scalable and parallelized training environments.

Known Issues

• Inference is not yet implemented; currently, training is the primary focus.
• SideChannels are not yet supported.
• Imitation Learning is not yet supported.
• Limited sensor support compared to Unity ML-Agents; more sensors will be added in future updates.
• Not all features have been thoroughly tested, and some areas may still contain unresolved issues.
• Lacking proper unit tests; additional validation and automated testing are needed to improve stability.

v0.1.0

First pre-release of the plugin. This release has the only goal to provide the ThirdPartyLibraries package that is required to use the plugin.