However, the most interested of you will not suffer from this expectation considering all the options that this product will offer them.
Indeed, the most experienced developers will have plenty of fun with DeepRacer, and this for the simple reason that this car will allow them to better understand the operation of reinforcement learning. The latter is a type of machine learning that is usually found only on autonomous cars.
Looking closely at DeepRacer, we see that the product is basically inspired by DeepLens, a combination of computer and video camera that re: Invent introduced in 2017. In short, DeepRacer is a bit like a complete Linux computer on wheels.
It also works from an Intel Atom processor with 4 GB of RAM.
An annual competition for developers
Amazon has even planned a competition system for DeepRacer developers. This is the AWS DeepRacer League. Each year, this competition will end with a championship re: Invent.
During the last re: Invent conference, competitors only had 24 hours to schedule DeepRacer cars and compete.
During this competition, the DeepRacer cars, AWS accounts and the entire MGM Grand Arena were made available to competitors. Similarly, they were also able to use two labs provided in Amazon's Github account.
Players will coach their DeepRacer cars
To prepare for this annual competition, players can complete training and evaluate their Reinforcement Learning Model online through the AWS Management Console.
To win the competition, players will also need to upgrade a rudimentary reward feature provided in Python in the AWS Management Console.
This reward feature is used to train the model to keep the DeepRacer vehicle on the road. The player will have to devote some of his time to improving this reward function while he will also have to use another part of his time to find a way to keep his car DeepRacer on the road while driving at high speed.
Once the Python code has been modified by the developer, it will just be enough to press the learning button. A virtual machine will be launched to execute the code and automatically cause its DeepRacer.
