Roadmap Update (Feb 14, 2020)

Hi all, please see release notes here on what was released as part of the 0.14.0. Our next release will be prior to the GDC conference. We will start making roadmap updates in the Unity forums. Please see our guidelines on where to route issues.

Previous roadmap updates:

Jan 21, 2020

ML-Agents 3 Month Roadmap as of Feb 12, 2020 -

• Repository restructuring to support packages
  • mlagents-envs (Low Level Python API)
  • mlagents-trainers (RL Algorithm development)
• Release of ML-Agents Unity preview package (Object-Oriented, Mono-behavior)
• Refactoring of the Unity SDK (C#) codebase
  • Introduction of Reward Providers (deprioritized)
  • Refactor of DemonstrationRecorder and store
  • Expose actions to users in API
• Refactoring of the Trainers (Python) codebase
  • Support for new LL-Python API, including Multi-Agent/Teams
  • Redesign and refactor of Trainer, TrainerController, and Buffer interfaces to no longer rely on BrainInfo and have a well-defined/documented interface
  • Breakdown of Trainer, Policy, and Model into re-usable bite-sized modules, and re-implementation of SAC and PPO using these modules
    Introduction of distributions
• Sensors and observations
  • Additional support for different data types
• Support for ONNX model export and import
• Explicit definition and documentation of the public APIs
• Updated docker images as part of each ML-Agents release

Thanks and please continue to provide us valuable feedback.

Jeff

 

I believe it's in master, but its not fully tested so "beware". https://github.com/Unity-Technologies/ml-agents/pull/3101

 

we should be shipping as part of the next release (in a week or so). We will still be releasing regardless of GDC.