search

Waikato / wekaDeeplearning4j

Weka package for the Deeplearning4j java library
https://deeplearning.cms.waikato.ac.nz/
GNU General Public License v3.0
184 stars 197 forks source link

Cant install wekaDeeplearning4j #33

Closed chanapong3 closed 6 years ago

chanapong3 commented 6 years ago

Description

Cant install wekaDeeplearning4j when I looked at console when start weka 3.8 console show like this

Skipping package wekaDeeplearning4j the OS/arch (Windows 7 x86) does not meet package OS/arch constraints: 64

It is look like wrong version 32 and 64 bit. I use this file wekaDeeplearning4j-1.5.3.zip How can I fix this issued ? Sorry if i did something wrong I am newbie here.

Expected behavior:

Encountered behavior:

Additional Information

braun-steven commented 6 years ago

Hi chanapong3,

could you please do the following and report if the above console message still shows up:

  1. Find your WEKA_HOME directory. If you haven't specified it yourself via an environment variable this directory usually resides in your home folder and is called wekafiles
  2. Go to wekafiles\packages\wekaDeeplearning4j
  3. Open and edit the File Description.props
  4. Remove OSArch=64 in line 52
  5. Restart weka from the console

Edit: Also, Weka 3.8 is not supported anymore. Please upgrade to 3.8.1 or higher

chanapong3 commented 6 years ago

Hello The problem is gone. 1.But I got new problems. When I run Dl4j classifier. it is show like this " Problem evaluating classifier: NULL " 2.How I know my WEKA is running in GPU base ? Thank a lot for fast response.

braun-steven commented 6 years ago
  1. Could you please elaborate the steps you have taken that lead to "Problem evaluating classifier: NULL"?
  2. Did you install the CUDA libraries? If not, check out our guide.
chanapong3 commented 6 years ago

1) I added my data set to WEKA. 1.1) I click classifies TAB and then Click Start to build model with Default Parameters. (Dlj4 classifier) 1.2)Then the dialog was pop up. " Problem evaluating classifier: NULL "

2)I have CUDA libraries already(was use with tensorflow). But i dont understand this line of code.

./install-cuda.sh ~/Downloads/wekaDeeplearning4j-cuda-9.1-1.5.0-linux-x86_64.zip

i tried to change to my file PATH. But it doesn't work. my cmd show like this The term 'install-cuda-libs.ps1' is not recognized as the name of a cmdlet,fuction,script file

braun-steven commented 6 years ago

1) Does the console output give any additional information? Would you mind sharing that dataset so I can have a look at what goes wrong? 2) You can just start the install-cuda-libs.ps1 powershell script and it automatically downloads and installs the additional libraries that are necessary for the package to use the GPU. You can download it from the latest release.

chanapong3 commented 6 years ago

1.I used Iris data set still have the same problems.

  1. Now I can install CUDA !!! my computer block script for security reason !! 3.Why my WEKA main directory is not WEKA_HOME(That have wekafiles fodler in it) is it ok ? or how to change it.
chanapong3 commented 6 years ago

1

This happen when I start to build model with Iris data

braun-steven commented 6 years ago

Which Weka version do you use?

chanapong3 commented 6 years ago

3.8.2

braun-steven commented 6 years ago

When seeing the error above, could you please scroll up and check for any sophisticated error message? I have the feeling that you either have a 32 bit Java installation or your Windows is 32 bit. Please check that as well.

The error message you got in the OP states that you are running Windows x86 which is 32 bit and not supported.

chanapong3 commented 6 years ago

It is work !! My window is 64 bit .My java is 64 bit too. I reinstall my JAVA and it s work. Thanks a lot. Now it is time to play WEKA. PS.How to know my WEKA is run on GPU.

eibe commented 6 years ago

You can use the nvidia-smi command to check what processes are running on the GPU.

On 19/07/2018, at 5:56 PM, chanapong3 notifications@github.com wrote:

It is work !! My window is 64 bit .My java is 64 bit too. I reinstall my JAVA and it s work. Thanks a lot. Now it is time to play WEKA. PS.How to know my WEKA is run on GPU.

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub, or mute the thread.

braun-steven commented 6 years ago

Nice to hear.

You can either check the console output when starting the classifier it should give some message that JCuBlas has been loaded and that your Graphic card with a certain amount of memory has been found.

On the other side you can also check your GPU stats with GPU-Z.

chanapong3 commented 6 years ago

Thank a lot.

chanapong3 commented 5 years ago

Did you check file is in there ? Or about " \" change to this "/" I am not professional.

From: Cknnne notifications@github.com Sent: Sunday, September 8, 2019 3:00 PM To: Waikato/wekaDeeplearning4j wekaDeeplearning4j@noreply.github.com Cc: chanapong3 chanapong3@hotmail.com; State change state_change@noreply.github.com Subject: Re: [Waikato/wekaDeeplearning4j] Cant install wekaDeeplearning4j (#33)

I have the same problem while adding GPU support :( After entering"./install-cuda-libs.ps1~/ "C:\Users\USER\Downloads\wekaDeeplearning4j-cuda-8.0-1.5.13-windows-x86_64.zip" my cmd shows this"bash: ./install-cuda-libs.ps1~/: No such file or directory"

[image]https://user-images.githubusercontent.com/39508332/64485207-1c911e80-d250-11e9-8e90-8aad46e5652f.png

Additional Information

— You are receiving this because you modified the open/close state. Reply to this email directly, view it on GitHubhttps://github.com/Waikato/wekaDeeplearning4j/issues/33?email_source=notifications&email_token=AEG23HJCVHQZNMIXBFXJ4FLQISWLHA5CNFSM4FKRHDXKYY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOD6FKJJY#issuecomment-529179815, or mute the threadhttps://github.com/notifications/unsubscribe-auth/AEG23HLU6U7AJYOT5HHZO73QISWLHANCNFSM4FKRHDXA.

GeethaSushama123 commented 2 years ago

I have weka 3.8.5 . I tried to install wekaDeeplearning4j:1.7.2 through GUI of weka Package Manager. Showing that it is installed. But when GPU is checked using "Is GPU Avaliable" showing false. I have cuda version showing 11.0 by " nvcc-V" command and Driver Version: 512.15 CUDA Version: 11.6 by "nvdia-smi" command. what should i do to get the GPU available for wekaDeeplearning4j

eibe commented 2 years ago

The current version of wekaDeeplearning4j requires CUDA Toolkit 10.0, 10.1, or 10.2, so you will need to use 10.2 instead of a later version.

On Sat, 27 Aug 2022 at 18:01, GeethaSushama123 @.***> wrote:

I have weka 3.8.5 . I tried to install wekaDeeplearning4j:1.7.2 through GUI of weka Package Manager. Showing that it is installed. But when GPU is checked using "Is GPU Avaliable" showing false. I have cuda version showing 11.0 by " nvcc-V" command and Driver Version: 512.15 CUDA Version: 11.6 by "nvdia-smi" command. what should i do to get the GPU available for wekaDeeplearning4j

— Reply to this email directly, view it on GitHub https://github.com/Waikato/wekaDeeplearning4j/issues/33#issuecomment-1229131122, or unsubscribe https://github.com/notifications/unsubscribe-auth/ACS2HVPJWXSK43HYQVW4B3DV3GVKFANCNFSM4FKRHDXA . You are receiving this because you commented.Message ID: @.***>