Package: nlmixr2nn
Version: 0.0.0.9000
Title: Neural-Network ODEs for 'rxode2' and 'nlmixr2'
Description: Embed neural networks inside ordinary differential
        equation right-hand sides for 'rxode2' and 'nlmixr2'. Networks
        are exposed as compiled 'rxode2' custom functions that supply
        the network value and its analytic input derivatives, allowing
        larger networks than the per-weight code-generation approach of
        'pmxNODE'.
Authors@R: person("Matthew", "Fidler", role = c("aut", "cre"),
    email = "matthew.fidler@gmail.com")
License: GPL (>= 3)
URL: https://nlmixr2.github.io/nlmixr2nn/,
        https://github.com/nlmixr2/nlmixr2nn
BugReports: https://github.com/nlmixr2/nlmixr2nn/issues
Encoding: UTF-8
Imports: rxode2, checkmate, stats, symengine
LinkingTo: rxode2, Rcpp, nlmixr2est
SystemRequirements: C++17. Optionally libtorch, via the 'torch' R
        package: when the 'torch' package and its libtorch binaries
        (torch::install_torch()) are present at build time, they are
        linked as the training backend. When they are not, a
        self-contained Adam/SGD optimizer is built instead, so the
        package installs and trains with no external library.
Suggests: torch, nlmixr2, nlmixr2est, numDeriv, testthat (>= 3.0.0),
        knitr, rmarkdown, minqa, lbfgsb3c, n1qn1, ggplot2, callr, withr
VignetteBuilder: knitr
NeedsCompilation: yes
Config/testthat/edition: 3
Config/roxygen2/version: 8.1.0
Config/pak/sysreqs: cmake libgmp3-dev make libicu-dev
Repository: https://nlmixr2.r-universe.dev
Date/Publication: 2026-09-20 21:07:42 UTC
RemoteUrl: https://github.com/nlmixr2/nlmixr2nn
RemoteRef: HEAD
RemoteSha: d2b9e8bbfe5528fa25d2eebd8aad7a624e997ade
Packaged: 2026-10-20 20:12:19 UTC; root
Author: Matthew Fidler [aut, cre]
Maintainer: Matthew Fidler <matthew.fidler@gmail.com>