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!

max_step is not working for --inference mode

Discussion in 'ML-Agents' started by Farshad22, Feb 26, 2022.

  1. Farshad22

    Farshad22

    Joined:
    Dec 21, 2021
    Posts:
    1
    Hi,
    I have trained a model using following command:

    mlagents-learn config/ppo/unit.yaml --run-id=myid --env=cs_window/Build --force


    that in config file(unit.yaml), max_step is set to 10000. So training process is stopped at step 10000 and save the .onnx file.

    So far so good!

    But now I want to inference the trained model. I have used the following command:

    mlagents-learn config/ppo/unit.yaml --run-id=myid --env=cs_window/Build --resume --inference


    Again, max_step is set to 10000 in unti.yaml config file but the inference process does not stop at step 10000.

    What is the problem? Why the max_step is not working for inference mode?

    Thanks
     
    Last edited: Feb 26, 2022
    Farshad22, Feb 26, 2022
    #1

  2. weight_theta

    weight_theta

    Joined:
    Aug 23, 2020
    Posts:
    65
    When does the inference process stop ?
    Note: A clever work around to this would be to simply continue training using a learning rate of 0, as you dont need to continue training when doing inference in RL (usually).
     
    weight_theta, Mar 1, 2022
    #2