[Question] Loading a trained PPO model in PyTorch without any SB3 dependencies

[SarvagyaVaish]

I created a custom gym environment and successfully trained a model using PPO. I then saved it using the model.save() method and got back a zip file (as per the documentation here). All good so far!

I now want to do inference using my trained model on a different device (in fact it's a jetson nano). How do I load and create a pytorch model from this zip file? I cannot install SB3 on this machine.

I have tried looking for something that does this but am worried my approach isn't correct. Any recommendations for an RL noobie?