Search Unity

  1. Megacity Metro Demo now available. Download now.
    Dismiss Notice

Discussion Potential error in the documentation of max_step in the Training Configuration

Discussion in 'ML-Agents' started by aashwinsnambiar12, Feb 13, 2024.

  1. aashwinsnambiar12


    Feb 7, 2024
    In the documentation page of the Training Configuration File (, the definition of max_steps seems to be misleading. Currently, the documentation states the following:
    "max_steps (default = 500000) Total number of steps (i.e., observation collected and action taken) that must be taken in the environment (or across all environments if using multiple in parallel) before ending the training process. If you have multiple agents with the same behavior name within your environment, all steps taken by those agents will contribute to the same max_steps count."

    The max_steps mentioned here should actually be the number of academy steps taken divided by the decision period in the decision requestor. Only the steps in which decision is requested counts towards the max_steps count.

    Is my understanding correct or did I go wrong somewhere?
    Last edited: Feb 15, 2024