Loading Models

[RichardKCook]

Hello everyone,

I've been learning how to use the FinRL library and I recently decided to move on from the demo version of ERL that's in some of their tutorials.

I spent last weekend rebuilding the train and test pipelines to work with the full version of ERL found in this library.

I know that my train function works and it saves to the correct directory, but I'm wondering if I'm loading the models correctly to test/apply them.

I'm using the init_agent() function to load my agent, but I'm afraid this is creating a new instance as opposed to loading a trained model. It's working without throwing errors, but that doesn't mean it's working as intended.

I'm happy to share my work if that will help

Also, when loading an agent for application and not backtesting, should I be setting the environment parameter to None for init_agent()?