search

srvk / eesen

The official repository of the Eesen project
http://arxiv.org/abs/1507.08240
Apache License 2.0
822 stars 342 forks source link

Error training ctc parallel (Cuda problem) #119

Closed sanakhamekhem closed 7 years ago

sanakhamekhem commented 7 years ago

Hi Alls, I have installed Eesen on ubuntu 16.04 and cuda 7.5 At the configure step , I used: ./configure --use-cuda=yes --cudatk-dir=/usr/local/cuda-7.5 --mathlib=OPENBLAS --openblas-root=/opt/OpenBLAS/ The nvidia settings is : 01:00.0 VGA compatible controller: NVIDIA Corporation GF119 [GeForce GT 610] (rev a1) 01:00.1 Audio device: NVIDIA Corporation GF119 HDMI Audio Controller (rev a1) nvcc: NVIDIA (R) Cuda compiler driver Copyright (c) 2005-2015 NVIDIA Corporation Built on Tue_Aug_11_14:27:32_CDT_2015 Cuda compilation tools, release 7.5, V7.5.17 When training, I have this error, How can I correct this??

ERROR (train-ctc-parallel:SelectGpuId():cuda-device.cc:86) No CUDA GPU detected!

ramonsanabria commented 7 years ago

Hi,

Can you run nvidia-smi command? This problem sometimes comes from a bad installation of nvidia drivers/cuda.

sanakhamekhem commented 7 years ago

Hi, nvidia-smi is not found, How can I install the appropriate nvidia driver?, I have installed it many times with variant versions, and I have got troubling like login loop,

2017-01-12 22:23 GMT+01:00 ramonsanabria notifications@github.com:

Hi,

Can you run nvidia-smi command? This problem sometimes comes from a bad installation of nvidia drivers/cuda.

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/srvk/eesen/issues/119#issuecomment-272287711, or mute the thread https://github.com/notifications/unsubscribe-auth/AO5KDD8MuiP7VjspSWnWda5Wny5XtVN8ks5rRpnGgaJpZM4LiE4a .

-- Best regards, Sana Khamekhem Jemni

PhD Student MIRACL Laboratory ENIS

ramonsanabria commented 7 years ago

Here you can find all nvidia drivers: http://www.nvidia.com/download/find.aspx

ramonsanabria commented 7 years ago

issue related with nvidia installation.

riebling commented 7 years ago

May I add: the GT 610 has only 1 GB of RAM. This may not be sufficient to do GPU training without modifying the recipe. If you run out of RAM, reduce --frame-limit=25000

sanakhamekhem commented 7 years ago

Thank you for alls, the problem is resolved

2017-01-13 6:39 GMT-08:00 riebling notifications@github.com:

May I add: the GT 610 has only 1 GB of RAM. This may not be sufficient to do GPU training without modifying the recipe. If you run out of RAM, reduce --frame-limit=25000

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/srvk/eesen/issues/119#issuecomment-272457866, or mute the thread https://github.com/notifications/unsubscribe-auth/AO5KDAyMCINkG4D_rwHY0hBBIRtCtCn-ks5rR4ycgaJpZM4LiE4a .

-- Best regards, Sana Khamekhem Jemni

PhD Student MIRACL Laboratory ENIS