Search Unity

  1. We are migrating the Unity Forums to Unity Discussions. On July 12, the Unity Forums will become read-only.

    Please, do not make any changes to your username or email addresses at id.unity.com during this transition time.

    It's still possible to reply to existing private message conversations during the migration, but any new replies you post will be missing after the main migration is complete. We'll do our best to migrate these messages in a follow-up step.

    On July 15, Unity Discussions will become read-only until July 18, when the new design and the migrated forum contents will go live.


    Read our full announcement for more information and let us know if you have any questions.

Questions about Multiple “Brains,” SetModel(), and Wall Jump example.

Discussion in 'ML-Agents' started by sarachan, Jun 2, 2022.

  1. sarachan

    sarachan

    Joined:
    Jul 21, 2013
    Posts:
    43
    I would like to create two or more models for the same agent, using the same action and observation spaces. I have seen the Wall Jump example and how SetModel() is used to switch among 3 models.

    However, I am not sure how these 3 models were trained. Were they each trained individually, or were they somehow trained all together? The documentation for SetModel() says, “The model and inferenceDevice parameters are ignored when not using inference.” So this sounds as though SetModel() does not switch among the models while training, implying that the models were trained individually. Is that correct? And if so, are there typically different agent scripts for the different models, or a single script having different cases for the different models? (The rewards would be different, although the actions and observations are the same.)

    Also, there are two different behaviors in the Wall Jump example: SmallWallJump and BigWallJump. I see that these bring up different sets of parameters from the configuration yaml file. Are there any other effects of these different named behaviors?

    Thanks for any insights or clarifications!