Troubles installing Tensorflow for R - r

I'm trying to install TensorFlow for R, and up to now I've followed these steps:
devtools::install_github("rstudio/tensorflow" Worked fine
library(tensorflow)
install_tensorflow()
And at this lasts point I get this error
HTTP error 404 while getting https://storage.googleapis.com/tensorflow/linux/cpu/tensorflow-1.9.0-cp37-cp37m-linux_x86_64.whl
Could not install requirement tensorflow==1.9.0 from https://storage.googleapis.com/tensorflow/linux/cpu/tensorflow-1.9.0-cp37-cp37m-linux_x86_64.whl because of error 404 Client Error: Not Found for url: https://storage.googleapis.com/tensorflow/linux/cpu/tensorflow-1.9.0-cp37-cp37m-linux_x86_64.whl
Could not install requirement tensorflow==1.9.0 from https://storage.googleapis.com/tensorflow/linux/cpu/tensorflow-1.9.0-cp37-cp37m-linux_x86_64.whl because of HTTP error 404 Client Error: Not Found for url: https://storage.googleapis.com/tensorflow/linux/cpu/tensorflow-1.9.0-cp37-cp37m-linux_x86_64.whl for URL https://storage.googleapis.com/tensorflow/linux/cpu/tensorflow-1.9.0-cp37-cp37m-linux_x86_64.whl
You are using pip version 10.0.1, however version 18.0 is available.
You should consider upgrading via the 'pip install --upgrade pip' command.
Errore: Error 1 occurred installing packages into conda environment r-tensorflow
Shortly it seems R tries to install Tensorflow with installing python version 3.7 which is not currently supported by this package (only up to 3.6).
Is it possible to force install_tensorflow to use an earlier python version?
P.S. I saw there is the possibility of using virtual envs, but I'm not familiar with them.

Related

Jupyternotebook Internal Server Error 500 due to conflicting pip dependency errors (Jinja2)

I was facing an 'Internal Server Error 500' problem with my Jupyter Notebook, which led me to uninstall and reinstall Anaconda multiple times each hour. However, the problem persisted. I followed the suggestions on stackoverflow to tackle the internal server error problem.
pip install --upgrade "ipython[all]"
An error message appeared in the midst of installing spyder 5.1.5.
spyder 5.1.5 requires pyqt5<5.13, which is not
installed.
spyder 5.1.5 requires pyqtwebengine<5.13, which is
not installed.
I proceeded to install the correct versions of pyqtwebengine and pyqt5. Along the way, Jinja2 and MarkupSafe were also installed.
Successfully installed pyqtwebengine-5.12.1
Successfully installed pyqt5-5.12.3
Successfully installed Jinja2-2.11.3 MarkupSafe-
1.1.1
When I tried to install spyder once more (spyder==5.1.5) thereafter, I was notified that conflicting versions of Jinja2 were required.
flask 2.1.0 requires Jinja2>=3.0, but you have
jinja2 2.11.3 which is incompatible.
cookiecutter 1.7.2 requires Jinja2<3.0.0, but you
have jinja2 3.0.0 which is incompatible.
More details about the conflicting version errors are shown in the picture.
Assuming that error involving conflicting versions of jinja2 is the reason for my 'internal server error 500' problem, may I know how should I go about revsolving this issue?

Dependency issue installing DevStack

I am trying to install Openstack using DevStack. After correctly creating the user stack and running ./stack.sh, the script starts to install the required dependencies until it hits Obtaining file:///opt/stack/cinder. Then I obtain the following errors:
Exception:
ContextualVersionConflict: (grpcio 1.21.1 (/usr/local/lib/python2.7/dist-packages), Requirement.parse('grpcio<1.16.0,>=1.0.0'), set(['tooz']))
I understand that there is a conflict in grpcio, but I don't know how to solve it. I already unsuccessfully tried to install a specific version between 1.0.0 and 1.16.0.
The Devstack version I am using is the latest from the git master.
I think I solved the problem myself.
I had the tooz package installed with version 1.62 which was causing the issue.
After removing that package the installation completed.

Trying to search or install a package doesn't work and gives an error

When i try to search for a package or a theme i get this error message
Searching for “package” failed. connect ETIMEDOUT
xx.xx.xxx.xx:xxx
But When i try to install a package or a theme if it the searched worked without a problem i get this error message
Installing “package#ver” failed
Request for package information failed: connect ETIMEDOUT
xx.xx.xxx.xx:xxx (5 attempts) (ETIMEDOUT)
Compiler tools not found Packages that depend on modules that contain
C/C++ code will fail to install. Read here for instructions on
installing Python and Visual Studio. Run apm install --check after
installing to test compiling a native module.

Error while installing tensorflow on Rstudio

I´m trying to install Tensorflow on Rstudio windows.
I have installed, Anaconda 3, all my R librarys are updated, and load library Keras on R
When I try to install, using:
install_keras()
The installation was not completed and an error message prompt:
Error: Error 2 occurred installing packages into conda environment r-tensorflow
In addition: Warning message:
"running command '"C:/PROGRA~3/ANACON~1/Scripts/activate" r-tensorflow && pip install --upgrade --ignore-installed https://storage.googleapis.com/tensorflow/windows/cpu/tensorflow-1.3.0-cp36-cp36m-win_amd64.whl had status 2"
Can someone help me in this issue?
Thanks!!!
I got the same error. I think it is because tensorflow is already installed through the install of keras. If I try to install keras again, I get the error that keras and tensorflow can only be installed in a fresh R session, where they had not been installed before. Otherwise there might be dll error. So it appears to me that you do not (and should not) install tensorflow if you have already installed keras. When I stopped Rstudio and started it again, I could install keras and re-install tensorflow through.

'curl_version_info_data' has no member named 'libssh_version'

I need to install plotly package for R(3.3.1) in a Redhat server 5.11. It requires curl to be installed first. I downloaded the package from https://cran.r-project.org/web/packages/curl/index.html.
When I try to install the package, it gives below error:
version.c: In function 'R_curl_version':
version.c:15: error: 'curl_version_info_data' has no member named 'libssh_version'
version.c:15: error: 'curl_version_info_data' has no member named 'libssh_version'
make: *** [version.o] Error 1
ERROR: compilation failed for package 'curl'
Has anyone had the same issue before?
Thanks in advance for any suggestion.
You need to install the development version of the curl library before installing the package. Try yum install curl-devel.
One note though - whenever you run into a similar error it is typically because a development version of a library is not installed on your redhat system.
Thanks Jim for your reply.
I found the solution is to use a lower version curl package which has not included the definition for "libssh_version" in the file version.c, e.g. curl_0.9.6.tar.gz with the command:
sudo R CMD INSTALL curl_0.9.6.tar.gz
I've also tried "yum install curl-devel", but it shows the message:
This system is receiving updates from RHN Classic or RHN Satellite.
Setting up Install Process
Package curl-devel-7.15.5-17.el5_9.x86_64 already installed and latest version
Package curl-devel-7.15.5-17.el5_9.i386 already installed and latest version
Nothing to do

Resources