unit-1

Unit 1: Introduction to Deep Reinforcement Learning

You can find the unit theory here.

HOW TO RUN

First check that you have the requirements listed here.

IMPORTANT: Before runing the following commands ensure that you are inside the unit folder.

1 - Build the Docker image

You only need to run this command once

docker-compose build

2 - Run training

If you want to modify some hyperparameters you can do it in the configuration files ppo.yaml and train.yaml.

docker-compose run train

Each training run will create a folder in "runs/train/<DATETIME>/". You will find there the experiment outputs (like the trained model).

3 - Push the trained model to the Hub

Before pushing the model to the Hub YOU MUST edit the file model_to_hub.yaml. You should at least change the model path (to point to the model in "runs/train/<DATETIME>/<MODEL_FILENAME>.zip") and the repo_id to push the model to you personal repository.

docker-compose run model_to_hub