I’m appreciated for ur response, but I encountered the error again:
Warning in install.packages :
unable to access index for repository https://apache-mxnet.s3-accelerate.dualstack.amazonaws.com/R/CRAN/GPU/cu92/bin/windows/contrib/3.6:
cannot open URL ‘https://apache-mxnet.s3-accelerate.dualstack.amazonaws.com/R/CRAN/GPU/cu92/bin/windows/contrib/3.6/PACKAGES’
Warning in install.packages :
cannot open URL : HTTP status was ‘404 Not Found’
Error in download.file(url, destfile, method, mode = “wb”, …) :
cannot open URL
Sorry you encountered that bug from official repo, it may be worth opening an issue on github.
In the meantime, have you tried the other option:
install.packages("https://jeremiedb.s3.ca-central-1.amazonaws.com/share/mxnet/GPU/1.5.1/mxnet.zip", repos = NULL)
Thanks for ur patience. The result showed below:
Warning in install.packages :
cannot open URL ‘https://jeremiedb.s3.ca-central-1.amazonaws.com/share/mxnet/GPU/1.5.1/mxnet.zip’: HTTP status was ‘403 Forbidden’
Error in download.file(p, destfile, method, mode = “wb”, …) :
cannot open URL ‘https://jeremiedb.s3.ca-central-1.amazonaws.com/share/mxnet/GPU/1.5.1/mxnet.zip’
Sorry, my mistake, I had forgot to make the file public for download. It’s now fixed, could you make another try?
It could be downloaded now,thanks for your help.
But i met an error while librarying mxnet:
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
I just tried reinstalling from the S3 storage and have it worked. I’m on Windows 10 with R 3.6.1. My current best guess would be that the issue would come from the CUDA link.
First, did the following for install:
> install.packages("https://jeremiedb.s3.ca-central-1.amazonaws.com/share/mxnet/GPU/1.5.1/mxnet.zip", repos = NULL)
Installing package into ‘C:/Users/Jeremie/Documents/R/win-library/3.6’
(as ‘lib’ is unspecified)
trying URL 'https://jeremiedb.s3.ca-central-1.amazonaws.com/share/mxnet/GPU/1.5.1/mxnet.zip'
Content type 'application/zip' length 244107614 bytes (232.8 MB)
downloaded 232.8 MB
> library(mxnet)
Is your CUDA added to your PATH? Does the following works from the command line:
D:\>where nvcc
C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v10.1\bin\nvcc.exe
Also, has CUDNN been installed? In the above CUDA directory, do you have a cudnn dll?:
Hope these might help!
@jeremiedb Really thanks for your helping.
For the first step, I encountered an error below:
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:/Users/Lin/Documents/R-package/inst/libs/libmxnet.so’:
LoadLibrary failure: 找不到指定的模組。
For the second step, the following command line does work, it could find the nvcc.exe file
However, till the third step, I cannot find a cudnn dll file.
Sorry about this late reply.
For what I seen, it seems like CUDNN had not been installed. To do so, you can follow the steps described here: https://docs.nvidia.com/deeplearning/cudnn/install-guide/index.html#install-windows
In short, you need to register with NVIDIA developer program, download the CUDNN zip and copy-paste the provided files in your CUDA folder.
