This README provides an overview of my current and future DeepRacer Models (v1 currently). The goal is basically to:
- Define an action space
- Develop a reward function
- Experiment with various hyperparameters
- Modify training times: Optimal turn speeds
- Test on different tracks: AWS DeepRacer Tracks
- Faster turning speed
- Easier handling of sharper turns
Training Optimization Used from deepracer-on-the-spot
- Training URLS :
http://98.80.171.249:8100/menu.htmlandhttp://98.80.171.249:8080 - Action space definitions
- Reward function templates
- Hyperparameter tuning scripts
- Training optimization techniques
- Track testing configurations