The Keras R package on Windows requires a conda environment into which to install the Python components, but this has nothing to do with the RStudio IDE (see: FAQ: What’s the difference between R and RStudio?).
The RStudio IDE does not install R for you — but if you’ve installed R for Windows according to the usual methods, then the RStudio IDE will interact with that installation of R. Similarly, if you have installed Anaconda according to the usual methods, the Keras R package (which you install using R) should be able to find what it needs without any trouble.
So the supported default sequence is:
- Install R for Windows (and RTools) from CRAN
- Install the RStudio IDE from RStudio
- Install Anaconda 3.x for Windows
- From the R console, run
install.packages("devtools")
- From the R console, run
devtools::install_github("rstudio/keras")
- From the R console, run:
library(keras) install_keras()
If you’re concerned about the details of any of this, or want a non-default Keras setup (e.g., GPU installation), take a look at the R Keras installation documentation.
If after that you still have questions, it’s probably best to start a new topic (you can link new topics to old ones: at the bottom of this topic, click the Share button, then click + New Topic in the little popup that appears), and be sure to add the #keras tag.