Skip to content

using vmas approach for parallel training #6253

@majid5776

Description

@majid5776

hi.
I'm so excited for using unity and ml agents.
my question is why don't you use vmas approach for parallel training.
it make training faster, but it is developed for 2D env.
this approach in VMAS is in the paper: A Vectorized Multi-Agent Simulator for
Collective Robot Learning.
tnx

Metadata

Metadata

Assignees

No one assigned

    Labels

    requestIssue contains a feature request.

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions