Reduce AWS DeepRacer costs with these tips
Autonomous 1/18th scale race car, driven by ML
AWS DeepRacer cost saving tips
Optimize your AWS DeepRacer costs with these tips
Use Free League Events
Participate in free AWS DeepRacer League events to hone your autonomous racing skills. This is a cost-effective way to train your reinforcement learning models.
Optimized Training Regime
Optimize reinforcement learning hyperparameters and training regimen to decrease iterations. Each iteration incurs cost, thus fewer iterations mean lower costs.
Reuse Pre-Trained Models
Using pre-existing models from the AWS DeepRacer model marketplace can be cheaper than training new ones. Only create new models if the existing ones don't meet your needs.
Service information & pricing
About AWS DeepRacer
- About AWS DeepRacer
-
AWS DeepRacer is a 3D racing simulator leveraging reinforcement learning for autonomous race cars training. It offers an educational gaming platform for developers to explore machine learning using Amazon SageMaker for model building and testing. The models can be implemented in real cars or simulated races. Global competitions for prizes are also available. AWS DeepRacer Evo, an advanced version, comes with stereo cameras and LiDAR sensor for better racing outcomes.
- AWS DeepRacer pricing
-
AWS DeepRacer's Free Tier provides 10 hours for model training, evaluation, and 5GB storage free in the first month. Following this, costs are $3.50 per hour for training or evaluation and $0.023/GB-month for storage. AWS DeepRacer car, Evo, and Sensor Kit are priced at $399, $598, and $249 respectively. Monthly model storage and submission cost stands at approximately $0.18.