Having the R 3.6 support should just be a matter of updating the S3 repo. I think @hetong007 was performing it? @LaranIkal are you looking for the CPU or GPU version? As a quick fix I could make a build tomorrow and post a link to access it.
I’ve just tried it after an update to Cuda 10.1 and still works fine on Windows with R 3.6.0.
It’s a built of MXNet 1.3.1, I’ll look for an update to 1.5.1 shortly.
Hi @lynne, are you looking for a CPU or GPU install? I unfortunately don’t have a direct access to a Mac, but if it’s for CPU, I might be able to give a shot at it in the upcoming week.
In the meatime, perhaps a bit hacky, but if you have a version of R 3.5 installed and make the installation of MXNet for which I think exist for OS X, then you will likely be able to use the library on R 3.6.0 by simply copy paste the mxnet package folder under the R library folder from 3.5 to 3.6.
@hetong007, aware of any official support for for Mac OS X? Also a Windows built on the official aws repo? Let me know if I can push Windows built on it.
Hi @jeremiedb, thanks for your attention and quick reply. I’m looking for the CPU version, and I got the the same problem when trying to install mxnet for both R 3.5 and 3.6. I tried to downloaded mxnet_1.3.0.tar.gz and sucessfully installed it in RStudio environment. But when I library(mxnet), I got the following error message:
Error: package or namespace load failed for ‘mxnet’:
.onLoad failed in loadNamespace() for ‘mxnet’, details:
call: dyn.load(file, DLLpath = DLLpath, …)
error: unable to load shared object ‘/Library/Frameworks/R.framework/Versions/3.5/Resources/library/mxnet/libs/libmxnet.so’:
dlopen(/Library/Frameworks/R.framework/Versions/3.5/Resources/library/mxnet/libs/libmxnet.so, 10): Library not loaded: /usr/local/opt/openblas/lib/libopenblasp-r0.3.1.dylib
Referenced from: /Library/Frameworks/R.framework/Versions/3.5/Resources/library/mxnet/libs/libmxnet.so
Reason: image not found
And then I tried to do as follows:
brew upgrade openblas
brew upgrade opencv
ln -sf /usr/local/opt/openblas/lib/libopenblasp-r0.3.3.dylib /usr/local/opt/openblas/lib/libopenblasp-r0.3.1.dylib
Hello,
I wonder if something similar is available also for linux. I am on debian and I have R 3.6 and I am experiencing problems in building from source according to
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 ‘C:/Progs/R/R-3.6.0/library/mxnet/libs/x64/libmxnet.dll’:
LoadLibrary failure: The specified module could not be found.
which seems weird since I checked the directory path and libmxnet.dll actually IS inside there
Hi @jeremiedb
I met an error after using the installing URL(GPU), while my Cuda version was 10,R version was 3.6.1
Could u please tell me how to deal with it?
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 ‘C:/R/R-3.6.1/library/mxnet/libs/x64/libmxnet.dll’:
LoadLibrary failure: 找不到指定的模組。