DESCRIPTION

Package: rlR
Type: Package
Title: Reinforcement Learning in R
Version: 0.1.0
Authors@R: c(
    person("Xudong", "Sun", email = {"smilesun.east@gmail.com"}, role = c("aut", "cre")),
    person("Sebastian", "Gruber", email = {"gruber_sebastian@t-online.de"}, role = c("ctb"))
    )
Maintainer: Xudong Sun <smilesun.east@gmail.com>
Description: Reinforcement Learning with deep Q learning, double deep Q
    learning, frozen target deep Q learning, policy gradient deep learning, policy
    gradient with baseline deep learning, actor-critic deep reinforcement learning.
License: BSD_2_clause + file LICENSE
Encoding: UTF-8
Depends:
    R (>= 3.4.0),
Imports:
    R6,
    checkmate,
    data.table,
    reticulate,
    keras,
    tensorflow,
    logging,
    ggplot2,
    openssl,
    magrittr,
    abind,
    foreach
LazyData: true
RoxygenNote: 6.1.1
BugReports: https://github.com/smilesun/rlR/issues
URL: https://github.com/smilesun/rlR
SystemRequirements: The following python package are needed to use the gym openAI environment. gym >= 0.10.5; At least one deep learning backend which keras requires(tensorflow, cntk, theano) should be installed on your computer, for example tensorflow >= 1.1.0 (tested on Ubuntu 14.04); The backend keras requires could be installed by keras::install_keras(); Both dependencies can also be installed by rlR::installDep() function. It is important to note that the user should run 'reticulate::use_python("/usr/local/bin/python")' to specify the python path and 'reticulate::use_virtualenv("myenv")' to specify which virtual environment to use. By default, the package is using "~/anaconda3/bin/python"  as its python version. For detail, please refer to https://rstudio.github.io/reticulate/articles/versions.html
Suggests:
    imager,
    png,
    devtools,
    testthat,
    knitr,
    covr,
    rmarkdown
VignetteBuilder: knitr