I am working in the following environment: Windows 11, R 4.4.3 and RStudio, Python 3.13.2, and Anaconda 25.1.1. Keras 3 is installed along with Tensorflow and reticulate

I am trying to install and run Keras on R for the past month. After facing numerous errors along the way, I seem to be stuck at the following error:

Error in stop_no_virtualenv_starter(version = version, python = python) : 
  Suitable Python installation for creating a venv not found.
  Requested version constraint: >=3.9,<=3.11

I have a hard time believing that I need to downgrade python

I am using the following simple code:

library(keras3)
library(tensorflow)
library(reticulate)
keras3::install_keras(backend = "tensorflow")

this triggers the aforementioned error.

I am working in the following environment: Windows 11, R 4.4.3 and RStudio, Python 3.13.2, and Anaconda 25.1.1. Keras 3 is installed along with Tensorflow and reticulate

I am trying to install and run Keras on R for the past month. After facing numerous errors along the way, I seem to be stuck at the following error:

Error in stop_no_virtualenv_starter(version = version, python = python) : 
  Suitable Python installation for creating a venv not found.
  Requested version constraint: >=3.9,<=3.11

I have a hard time believing that I need to downgrade python

I am using the following simple code:

library(keras3)
library(tensorflow)
library(reticulate)
keras3::install_keras(backend = "tensorflow")

this triggers the aforementioned error.

• r
• keras

1 Answer 1

First of all, you should update your Reticulate package to the latest version, which is 1.41 because a new function for handling Python dependencies has been included, py_require(). You can read more about it here

After updating the package, in your R console run:

py_require("keras")

When you do this, Reticulate will automatically handle the installation and loading of the specified Python package when the Python session is initialized.

It will create an ephemeral virtual environment that will be used to manage the Python dependencies declared via py_require() and resolve any conflicts behind the scenes.

If you then want to import the installed package, run:

kr <- import("keras")