bitrl is an effort to provide implementations and wrappers of environments suitable for training reinforcement learning agents using C++. You can find a very basic example here BitRL Example 1 Using Gymnasium environments (Part 1).

cuberl is an extension of bitrl that provides implementations of various reinforcement learning algorithms.

Gymnasium environments exposed over a REST like API can be found at: bitrl-rest-api. Various RL algorithms using the environments can be found at cuberl.

Dependencies

Below are the dependencies for the common dependencies for both libraries

CMake
Boost
Eigen3
Blas

bitrl has the following dependencies:

OpenCV (optional)
PahoMqttCpp (optional)
Chrono (optional)

In addition the library uses

json (extern/nlhmann)
HTTPRequest (extern/)

cuberl has the following dependencies:

OpenCV (optional)
PahoMqttCpp (optional)
Pytorch (optional)
Chrono (optional)

Installation

The usual cmake installation/build can be used:

mkdir build && cd build
cmake -DCMAKE_INSTALL_PREFIX=/path/where/bitrl/should/be/installed/to ..
make install -j4

Build the documentation

The documentation is maintained on a different repository. However, you can build the documentation locally. You will need Doxygen and graphviz installed:

sudo apt install doxygen graphviz

Clone the repository that hosts the documentation:

git clone https://github.com/pockerman/bitrl_cuberl_docs

Update the submodules for the two libraries:

git submodule init

git submodule update

Execute doxygen

doxygen Doxyfile

The documentation will be installed under ./docs/html under the project's source directory

Examples

➡ Examples