This function can be used to create a persistent virtual environment for usage
with Legacy Keras. New code is recommended to use the keras3 package which
works automatically and does any special configuration like this.
install_keras(
method = c("auto", "virtualenv", "conda"),
conda = "auto",
version = "default",
tensorflow = version,
extra_packages = NULL,
...
)py_require_legacy_keras(extra_packages = TRUE)
Installation method. By default, "auto" automatically finds a method that will work in the local environment. Change the default to force a specific installation method. Note that the "virtualenv" method is not available on Windows.
The path to a conda executable. Use "auto" to allow
reticulate to automatically find an appropriate conda binary.
See Finding Conda and conda_binary() for more details.
TensorFlow version to install. Valid values include:
"default" installs 2.20
"release" installs the latest release version of tensorflow (which may
be incompatible with the current version of the R package)
A version specification like "2.4" or "2.4.0". Note that if the patch
version is not supplied, the latest patch release is installed (e.g.,
"2.4" today installs version "2.4.2")
nightly for the latest available nightly build.
To any specification, you can append "-cpu" to install the cpu version
only of the package (e.g., "2.4-cpu")
The full URL or path to a installer binary or python *.whl file.
Synonym for version. Maintained for backwards.
Additional Python packages to install along with TensorFlow.
other arguments passed to reticulate::conda_install() or
reticulate::virtualenv_install(), depending on the method used.
Note that recent versions of reticulate will automatically manage
environments if dependencies are declared with reticulate::py_require().
Instead of using this install_keras() function, new users are first
encouraged to use keras3. If that's not an option, instead of creating a
persistent venv, you can this code at the start of the R session, before
loading keras:
# declare requirements for legacy keras (tf-keras)
reticulate::py_require(c("tensorflow", "tf-keras", "numpy<2"))# also declare optional dependencies
reticulate::py_require(c(
"tensorflow-hub", "tensorflow-datasets", "scipy",
"requests", "Pillow", "h5py", "pandas", "pydot"
))
library(keras)
Or more simply, call py_require_legacy_keras()
library(keras)
py_require_legacy_keras()
This function will install Tensorflow and all Keras dependencies. This is a
thin wrapper around tensorflow::install_tensorflow(), with the only
difference being that this includes by default additional extra packages that
keras expects, and the default version of tensorflow installed by
install_keras() may at times be different from the default installed
install_tensorflow(). The default version of tensorflow installed by
install_keras() is "2.15".
The default additional packages are: tensorflow-hub, tensorflow-datasets, scipy, requests, pyyaml, Pillow, h5py, pandas, pydot, with their versions potentially constrained for compatibility with the requested tensorflow version.
tensorflow::install_tensorflow()