Search Unity

  1. Welcome to the Unity Forums! Please take the time to read our Code of Conduct to familiarize yourself with the forum rules and how to post constructively.
  2. Unite 2023 Registrations are now LIVE! We are thrilled to announce that Unite 2023 Registration is now live and open for all!
    Dismiss Notice
New Forum User Notice Unite 2023 Registrations are now LIVE!

Question Decision Requester

Discussion in 'ML-Agents' started by Atilli, Jul 29, 2023.

  1. Atilli

    Atilli

    Joined:
    Aug 31, 2022
    Posts:
    11
    For training I have multiple copies of the same area, the only difference being each agent is configured with a different decision period in the Decision Requester.
    I understand that TensorBoard averages the results of all the agents.
    What I am interested in, is determining which agent's decision period lead to the best training result.
    How can I get this data, if possible?
    Thanks
     
    Atilli, Jul 29, 2023
    #1

  2. Thorce

    Thorce

    Joined:
    Jul 3, 2019
    Posts:
    38
    Instead of training them all at once you should train them after each other. Then you can compare the different runs to each other and continue with the best setting
     
    Thorce, Aug 2, 2023
    #2

  3. Atilli

    Atilli

    Joined:
    Aug 31, 2022
    Posts:
    11
    yep, that's what I've had to do
     
    Atilli, Aug 4, 2023
    #3