Navigation

Training Code

The repository (or zip file) includes functional, well-documented, and organized code for training the agent.

Framework

The code is written in PyTorch and Python 3.

Saved Model Weights

The submission includes the saved model weights of the successful agent.

README.md

The GitHub (or zip file) submission includes a README.md file in the root of the repository.

Project Details

The README describes the the project environment details (i.e., the state and action spaces, and when the environment is considered solved).

Getting Started

The README has instructions for installing dependencies or downloading needed files.

Instructions

The README describes how to run the code in the repository, to train the agent. For additional resources on creating READMEs or using Markdown, see here and here.

Report

The submission includes a file in the root of the GitHub repository or zip file (one of Report.md, Report.ipynb, or Report.pdf) that provides a description of the implementation.

Learning Algorithm

The report clearly describes the learning algorithm, along with the chosen hyperparameters. It also describes the model architectures for any neural networks.

Plot of Rewards

A plot of rewards per episode is included to illustrate that the agent is able to receive an average reward (over 100 episodes) of at least +13. The submission reports the number of episodes needed to solve the environment.

Ideas for Future Work

The submission has concrete future ideas for improving the agent's performance.