search

Unity-Technologies / ml-agents

The Unity Machine Learning Agents Toolkit (ML-Agents) is an open-source project that enables games and simulations to serve as environments for training intelligent agents using deep reinforcement learning and imitation learning.
https://unity.com/products/machine-learning-agents
Other
16.93k stars 4.14k forks source link

Allow to train at runtime #4360

Closed Fax3D closed 4 years ago

Fax3D commented 4 years ago

A great features, imho, it would be the possibility to perfom an agent training not only offline via Unity-editor, but also during the game execution! There are some ai-oriented game scenario, for example when you want to use imitation learning, in wich would be very useful to run a command to perform a train directly in-game, without rely on an external service.

What do you think about?

Thanks

shihzy commented 4 years ago

hi @Fax3D - can you clarify your request? You should be able to do both imitation learning and training via mlagents-learn. Let me know if there is something I am missing in your request. Thanks!

Fax3D commented 4 years ago

Hi @unityjeffrey, mlagents-learn it's callable during gameplay? I mean... it's possible for an user playing my game start a train of what he played using imitation learning? Or mlagents-learn it's callable only offline? Thanks!

shihzy commented 4 years ago

Hi @Fax3D - ah I think I see what you are asking. If someone playing your game, could the agent learn while the person is playing the game. Unfortunately, the trainer code is run in Python and not C#. We are looking into potentially incorporating C# training in the future.

github-actions[bot] commented 3 years ago

This thread has been automatically locked since there has not been any recent activity after it was closed. Please open a new issue for related bugs.