[GitHub] [incubator-mxnet] AgatheBauer opened a new issue #15194: How do I get the smoothest installation experience in R for GPU?

[GitBox <gi...@apache.org>]

AgatheBauer opened a new issue #15194: How do I get the smoothest installation experience in R for GPU?
URL: https://github.com/apache/incubator-mxnet/issues/15194
 
 
   Assume, I'm a total idiot (indeed, I feel like one, at least like a bloody beginner...)
   
   This whole neural net installation for GPU is a total nightmare. I came here after failing to get tensorflow to run. Seems like it's the same disaster with MXNet. People, it's awesome that these tools are free to use for a hobbyist like me, but not everybody is a professional developer! I will keep learning of course, but this hurdle is just too high at the moment.
   
   This is my whole installation Routine:
   
   Windows 10 Version 10.0.18362.145 64bit (clean install)
   R-3.6.0 -> uninstalled and installed R-3.5.2 later
   R-Studio 1.2.1335
   Rtools35
   Nvidia Display Driver 26.21.14.3086 (GTX 1060 3GB)
   cuda_9.0.176_win10 -> update to 9.0.176.1 -> 9.0.176.2 -> 9.0.176.3 (
   cudnn-9.0-windows10-x64-v7.1.zip -> unzip cudnn64_7.dll to C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v9.0\bin
   Anaconda3-2019.03-Windows-x86_64 -> downgrade to Anaconda3-5.2.0-Windows-x86_64 after problems with tensorflow -> re-upgrade back later to Anaconda3-2019.03-Windows-x86_64
   
   In R-Studio, I called:
   ```
   cran <- getOption("repos")
   cran["dmlc"] <- "https://apache-mxnet.s3-accelerate.dualstack.amazonaws.com/R/CRAN/GPU/cu90"
   options(repos = cran)
   install.packages("mxnet")
   ```
   Calling `library('mxnet') ` i get this:
   
   > Error: package or namespace load failed for ‘mxnet’:
   >  .onLoad failed in loadNamespace() for 'mxnet', details:
   >   call: inDL(x, as.logical(local), as.logical(now), ...)
   >   error: unable to load shared object 'D:/Benutzer/MyName/Documents/R/win-library/3.5/mxnet/libs/x64/libmxnet.dll':
   >   LoadLibrary failure:  Die angegebene Prozedur wurde nicht gefunden. (German for "the specified procedure wasn't found")
   > 
   
   Is there any older version that a dumb end user like me can get running by just copying and pasting some lines of R code in the R terminal and execute these?

----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
users@infra.apache.org


With regards,
Apache Git Services