Ecosyste.ms: Awesome
An open API service indexing awesome lists of open source software.
https://github.com/vowpalwabbit/reinforcement_learning
Interaction-side integration library for Reinforcement Learning loops: Predict, Log, [Learn,] Update
https://github.com/vowpalwabbit/reinforcement_learning
cpp machine-learning reinforcement-learning
Last synced: 5 days ago
JSON representation
Interaction-side integration library for Reinforcement Learning loops: Predict, Log, [Learn,] Update
- Host: GitHub
- URL: https://github.com/vowpalwabbit/reinforcement_learning
- Owner: VowpalWabbit
- License: mit
- Created: 2018-10-03T19:02:54.000Z (about 6 years ago)
- Default Branch: master
- Last Pushed: 2024-11-04T21:08:14.000Z (9 days ago)
- Last Synced: 2024-11-04T22:19:33.827Z (9 days ago)
- Topics: cpp, machine-learning, reinforcement-learning
- Language: C++
- Homepage:
- Size: 15.7 MB
- Stars: 75
- Watchers: 15
- Forks: 40
- Open Issues: 68
-
Metadata Files:
- Readme: README.md
- License: LICENSE.txt
Awesome Lists containing this project
README
[![Linux Build Status)](https://img.shields.io/azure-devops/build/vowpalwabbit/3934113c-9e2b-4dbc-8972-72ab9b9b4342/23/master?label=Linux%20build&logo=Azure%20Devops)](https://dev.azure.com/vowpalwabbit/Vowpal%20Wabbit/_build?definitionId=31)
[![MacOS Build Status](https://img.shields.io/azure-devops/build/vowpalwabbit/3934113c-9e2b-4dbc-8972-72ab9b9b4342/22/master?label=MacOS%20build&logo=Azure%20Devops)](https://dev.azure.com/vowpalwabbit/Vowpal%20Wabbit/_build?definitionId=32)
[![Windows Build status](https://ci.appveyor.com/api/projects/status/57p7o5v34onsqma2/branch/master?svg=true)](https://ci.appveyor.com/project/JohnLangford/reinforcement-learning/branch/master)
[![Integration with latest VW](https://github.com/VowpalWabbit/reinforcement_learning/actions/workflows/daily_integration.yml/badge.svg?event=schedule)](https://github.com/VowpalWabbit/reinforcement_learning/actions/workflows/daily_integration.yml)# RL Client Library
Interaction-side integration library for Reinforcement Learning loops: Predict, Log, [Learn,] Update## Compiling the library
### Getting the source code
```sh
git clone --recursive https://github.com/VowpalWabbit/reinforcement_learning.git
cd reinforcement_learning
```### Building with Vcpkg + CMake Presets
#### Requirements
* CMake 3.20+
* Ninja#### Commands
```sh
cmake --preset=vcpkg-release
cmake --build --preset=vcpkg-release
```## Alternative system-specific build instructions
### Ubuntu
#### Dependencies
Install the dependencies for this project with the following commands. We recommend the use of [vcpkg](https://github.com/microsoft/vcpkg) for installing `cpprestsdk` and `flatbuffers`.
```sh
sudo apt-get install libboost-all-dev libssl-dev
vcpkg install cpprestsdk flatbuffers
```#### Configure + Build
When configuring a CMake project using vcpkg dependencies, we must provide the full path to the `vcpkg.cmake` toolchain file.
```sh
# Configure
cmake -S . -B build -DCMAKE_TOOLCHAIN_FILE=/scripts/buildsystems/vcpkg.cmake# Build
cmake --build build -j `nproc`# Test
cmake --build build --target rltest -j `nproc`
cmake --build build --target test
```### MacOS
#### Dependencies
MacOS dependencies can be managed through homebrew.
```sh
brew install cpprestsdk flatbuffers openssl
```#### Configure + Build
In order to build using homebrew dependencies, you must invoke cmake this way:
```sh
cmake -S . -B build -DOPENSSL_ROOT_DIR=`brew --prefix openssl` -DOPENSSL_LIBRARIES=`brew --prefix openssl`/lib
cmake --build build --target all -j 4
```### Windows
#### Dependencies
Dependencies on Windows should be managed using [vcpkg](https://github.com/microsoft/vcpkg).
```cmd
vcpkg install --triplet x64-windows zlib boost-system boost-program-options boost-test boost-align boost-foreach boost-math boost-uuid cpprestsdk flatbuffers openssl
```If needed, add the flatbuffers executables to your PATH: `\installed\x64-windows\tools\flatbuffers`
#### Configure + Build in Visual Studio
Open the project directory in Visual Studio. Building the library can be easily done in the GUI.
- Edit CMake command line settings with `Project > CMake Settings for `
- Run CMake configuration with `Project > Configure Cache`. Use `Project > Delete Cache and Reconfigure` to force a full reconfiguration starting from a clean working directory.
- Compile with `Build > Build All`
- Run tests with `Test > Run CTests for `This procedure has been verified to work well in Visual Studio 2022.
#### Configure with command line + Build in Visual Studio
Alternatively, CMake configuration can be done in the command line.
```cmd
cmake -S . -B build -DCMAKE_TOOLCHAIN_FILE=\scripts\buildsystems\vcpkg.cmake -DVCPKG_TARGET_TRIPLET=x64-windows -A x64 -G "Visual Studio 16 2019"
rem Generates a solution you can open and use in Visual Studio
.\build\reinforcement_learning.sln
```Set VcpkgIntegration environment variable to `vcpkg.targets` file on your machine. Example:
```
VcpkgIntegration=c:\s\vcpkg\scripts\buildsystems\msbuild\vcpkg.targets
```Open `reinforcement_learning.sln` in Visual Studio. Build Release or Debug x64 configuration.
## Troubleshooting
### OpenSSL on MacOS
If you get an error similar to the following on MacOS when running `cmake ..`, then you may be able to fix it by supplying the OpenSSL path to CMake.
```
Make Error at /usr/local/Cellar/cmake/3.14.4/share/cmake/Modules/FindPackageHandleStandardArgs.cmake:137 (message):
Could NOT find OpenSSL, try to set the path to OpenSSL root folder in the
system variable OPENSSL_ROOT_DIR (missing: OPENSSL_INCLUDE_DIR)
Call Stack (most recent call first):
/usr/local/Cellar/cmake/3.14.4/share/cmake/Modules/FindPackageHandleStandardArgs.cmake:378 (_FPHSA_FAILURE_MESSAGE)
/usr/local/Cellar/cmake/3.14.4/share/cmake/Modules/FindOpenSSL.cmake:413 (find_package_handle_standard_args)
/usr/local/Cellar/cmake/3.14.4/share/cmake/Modules/CMakeFindDependencyMacro.cmake:48 (find_package)
/usr/local/lib/cpprestsdk/cpprestsdk-config.cmake:11 (find_dependency)
CMakeLists.txt:9 (find_package)
```This can be fixed by invoking CMake similar to the following:
```bash
cmake -DOPENSSL_ROOT_DIR=`brew --prefix openssl` -DOPENSSL_LIBRARIES=`brew --prefix openssl`/lib ..
```### System cpprestsdk on Linux
Installing cpprestsdk on Ubuntu18.04 using apt-get may result in cmake failing with:
```
CMake Error at CMakeLists.txt:9 (find_package):
By not providing "Findcpprestsdk.cmake" in CMAKE_MODULE_PATH this project
has asked CMake to find a package configuration file provided by
"cpprestsdk", but CMake did not find one.Could not find a package configuration file provided by "cpprestsdk" with
any of the following names:cpprestsdkConfig.cmake
cpprestsdk-config.cmakeAdd the installation prefix of "cpprestsdk" to CMAKE_PREFIX_PATH or set
"cpprestsdk_DIR" to a directory containing one of the above files. If
"cpprestsdk" provides a separate development package or SDK, be sure it has
been installed.
```The workaround is to specify where to search
```
cmake .. -DCMAKE_PREFIX_PATH=/usr/lib/x86_64-linux-gnu/cmake
```