Search Unity

  1. Unity support for visionOS is now available. Learn more in our blog post.
    Dismiss Notice

Train Model with another Algorithm

Discussion in 'ML-Agents' started by dani_kal, Jul 22, 2020.

  1. dani_kal

    dani_kal

    Joined:
    Mar 25, 2020
    Posts:
    52
    Hello!!!!
    I would like to ask how can I train an agent with my own algorithm? I dont want to use PPO algorithm but I want to add another trainer in the ml-agents/ trainers. I have search a lot to find something like that but with no success.
    I have to create these three scripts, models, trainer, policy . as I saw in PPO and SAC?

    I have stack to this point a lot.......

    If someone could give some advice.....

    Thank you in advance!
     
  2. awjuliani

    awjuliani

    Unity Technologies

    Joined:
    Mar 1, 2017
    Posts:
    69
    Hi,

    Could you please provide more details about where you are getting stuck in creating your own trainer?
     
  3. dani_kal

    dani_kal

    Joined:
    Mar 25, 2020
    Posts:
    52
    Hello!!!
    Thank you very much for your interest!
    I dont have implement my own trainer yet.
    I try to understand what I have to include in my scripts.
    So, for a new trainer we need three scripts. Models, trainer and policy.
    Of what each should have to consist of?
    Forgive me if my question sounds very simple but I tried to study the code of PPO to understand when and what happens in the algorithm but I have been confused.