Installation failed

[shiyu-wangbyte]

Sorry to bother developers.

As your document guide, I am trying to install this software on my machine (OpenSuse 15.4).

First, I create an anaconda virtual environment with python (verion 3.6.13) and install tensorflow (version 2.2.0) keras (version 2.3.0) with pip install command.

Next, I install R (version 3.5.1) on my machine with source code and trying to install reticulate for R. However, an error occur: package ‘withr’ is not available (for R version 3.5.1). Therefore, I just can update the version to 4.2.0 and install reticulate successfully.

When I trying to run : R/4.2.0/bin/Rscript Off-targetP_ML.R external_test.csv, an new error occur: Error in library(tensorflow) : there is no package called ‘tensorflow’. So, my question is : although i activated the conda environment r-tensorflow, the tensorflow and keras for python still cannot be used for R. Therefore, I install tensorflow and keras using "install.packages("tensorflow")" and "install.packages("keras")". As a result, tensorflow (version 2.16.0) and keras (version 2.15.0) were installed.

When I trying to run : R/4.2.0/bin/Rscript Off-targetP_ML.R external_test.csv, an new error occur: ValueError: Unknown regularizer: L2. That maybe because the version of tensorflow is too high. Therefore, I download the 2.2.0.tar.gz file from tensorflow for R github, and install tensorflow for R from local file. and download the 2.3.0.0.tar.gz file from keras for R github, and install keras for R from local file. However, the above error had not been solved, still reminding me that ValueError: Unknown regularizer: L2.

Do you have any suggestion for me? Thank you!

[DohaNaga]

Hello, It seems that the problems are all within packages version issues. I would suggest two options:

…

-first option is to look into the code and adapt the syntax according to the new versions you are using -second option is to check the code and use the exact versions for all of your packages. Best, Doha

On Thu 10 Oct 2024 at 03:26, shiyu-wangbyte ***@***.***> wrote: Sorry to bother developers. As your document guide, I am trying to install this software on my machine (OpenSuse 15.4). First, I create an anaconda virtual environment with python (verion 3.6.13) and install tensorflow (version 2.2.0) keras (version 2.3.0) with pip install command. Next, I install R (version 3.5.1) on my machine with source code and trying to install reticulate for R. However, an error occur: package ‘withr’ is not available (for R version 3.5.1). Therefore, I just can update the version to 4.2.0 and install reticulate successfully. When I trying to run : R/4.2.0/bin/Rscript Off-targetP_ML.R external_test.csv, an new error occur: Error in library(tensorflow) : there is no package called ‘tensorflow’. So, my question is : although i activated the conda environment r-tensorflow, the tensorflow and keras for python still cannot be used for R. Therefore, I install tensorflow and keras using "install.packages("tensorflow")" and "install.packages("keras")". As a result, tensorflow (version 2.16.0) and keras (version 2.15.0) were installed. When I trying to run : R/4.2.0/bin/Rscript Off-targetP_ML.R external_test.csv, an new error occur: ValueError: Unknown regularizer: L2. That maybe because the version of tensorflow is too high. Therefore, I download the 2.2.0.tar.gz file from tensorflow for R github, and install tensorflow for R from local file. and download the 2.3.0.0.tar.gz file from keras for R github, and install keras for R from local file. However, the above error had not been solved, still reminding me that ValueError: Unknown regularizer: L2. Do you have any suggestion for me? Thank you! — Reply to this email directly, view it on GitHub <#2>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AMQZQGRQTN5JFECBBP7I4RDZ2XJVZAVCNFSM6AAAAABPVWEDFGVHI2DSMVQWIX3LMV43ASLTON2WKOZSGU3TOMRYHE3DQNQ> . You are receiving this because you are subscribed to this thread.Message ID: ***@***.***>

[shiyu-wangbyte]

Thanks for you reply!

The error information follows:
"
The loop is magicalized with print().
| | 0%Error in py_call_impl(callable, call_args$unnamed, call_args$named) :
ValueError: Unknown regularizer: L2
Run reticulate::py_last_error() for details.
Calls: for ... load_model -> do.call -> -> py_call_impl
Execution halted
"
You suggest me to check the version of tensorflow and keras of R or Python?