I installed Annaconda, tensorflow ans Keras and Python version is 3.8.5.
But after installing R and the 2 packages Tensorflow and Keras I always got Python 3.6. So Keras/TensorFlow does not work and I cannot run models as illustrated after. I uninstalled, reinstalled, changed the way to perform and always I got 3.6 in libpython. How to do to get Python 3.8 in R ?
I installed in the following order:
install.packages("devtools") library(devtools) install.packages("reticulate") install.packages("tensorflow") install.packages("keras") library(tensorflow) install_tensorflow() library(keras) install_keras() reticulate::py_config() > reticulate::py_config() python: C:/Users/PC/AppData/Local/r-miniconda/envs/r-reticulate/python.exe libpython: C:/Users/PC/AppData/Local/r-miniconda/envs/r-reticulate/python36.dll pythonhome: C:/Users/PC/AppData/Local/r-miniconda/envs/r-reticulate version: 3.6.10 |Anaconda, Inc.| (default, May 7 2020, 19:46:08) [MSC v.1916 64 bit (AMD64)] Architecture: 64bit numpy: C:/Users/PC/AppData/Local/r-miniconda/envs/r-reticulate/Lib/site-packages/numpy numpy_version: 1.19.1
If I run a model I got the following message:
> setwd("C:/Users/PC/Desktop/R/") > model <- load_model_hdf5("my_model_2c2.h5")
Error: Installation of TensorFlow not found.
Python environments searched for 'tensorflow' package: C:\Users\PC\AppData\Local\r-miniconda\envs\r-reticulate\python.exe You can install TensorFlow using the install_tensorflow() function.
I tried also the next solution. I wrote in a .Renviron file the following line into the user's home directory does work. `Sys.setenv(RETICULATE_PYTHON = "C:/Users/PC/AppData/Local/r-miniconda/")` But Numpy and Tensorflow are lost after restarting R as illsutrated:
version: 3.8.3 (default, May 7 2020, 19:46:08) [MSC v.1916 64 bit (AMD64)]
numpy: [NOT FOUND]
Tensorflow: [NOT FOUND]