You can't train directly on the virtual circuit tracks.
Each race in the circuit has its own new competition track that you cannot directly trained on. This is to prevent people from just overfitting a model on the trainable track to get the best time. The virtual circuit will resemble the matching trainable circuit, but will be changed just enough to ensure your model needs to be generalized.
What is good Action Space and Hyperparameter for Kumo Torakku track?asked 3 years ago
[deepracer] about originally included model 'sample_single_cam'asked 2 months ago
How to increase Speed of a Deepracer Car?asked 9 days ago
How to track the inactive assets in AWSasked 9 months ago
Does Textract have a way to extract a table of specified dimensions?asked 6 months ago
Tracking model artifacts used in machine learningAccepted AnswerEXPERTasked 2 years ago
How does virtual training translate to physical model performance?asked 2 months ago
Is it possible to track reads of SSM Parameter?asked 8 months ago
Different layouts of Shanghai Sudu trackAccepted Answerasked 3 years ago
Is the DeepRacer service down?asked 4 years ago