AsmRL

Reinforcement learning done in assembly. A simplified implementation of the tabular Q-learning algorithm for an easy 8x8 grid world in ARMv7. Why? Good question.

Requirements

You'll need to have the GCC Arm toolchain to build this project. Also, if your host machine doesn't support ARMv7, you'll need QEMU to run the executable.

Build

Just run make:

make

and run with:

./main

If your host machine doesn't support ARMv7, just run it with QEMU:

qemu-arm ./main

Docker

You can also run using our built Docker image:

docker run unexploredtest/asmrl:latest