Loading pre-trained model using Python API

Unity Engine
, ,
1

Hello everyone,

I trained an agent using mlagents-learn. I visualized how it’s doing post-training via “inference” mode. However, this was done via command line. I want to be able to use the python API to load in my pre-trained model and do some other stuff with it like collect trajectories for agent.

I am somewhat familiar with “UnityEnvironment” class but only used it to load in environments with fresh agents not pre-trained. How do I go about loading a pre-trained model in my python script?

I wasn’t able to find pertinent information in the ml-agents python API docs. Would someone be able to point me in the right direction?

Thanks in advance!

2

Models trained using their learn command can be run on python. There are showcase projects in the showcase thread that did manage to do this.

I think you have to look up how to run the .nn onyx files in python.