experiments/STT-tensorflow
1
0
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity

Update cuda instructions to be more specific about versions (#2065)

Browse Source
This commit is contained in:
Vijay Vasudevan 2016-04-22 13:51:21 -07:00
parent dc19800ee1
commit 4b7bc3174e
1 changed files with 13 additions and 6 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

19
tensorflow/g3doc/get_started/os_setup.md
View File

@ -7,8 +7,10 @@ github source.
The TensorFlow Python API supports Python 2.7 and Python 3.3+. The TensorFlow Python API supports Python 2.7 and Python 3.3+.
The GPU version (Linux only) requires the Cuda Toolkit >= 7.0 and cuDNN >= The GPU version (Linux only) works best with Cuda Toolkit 7.5 and
v2. Please see [Cuda installation](#optional-install-cuda-gpus-on-linux) cuDNN v4. other versions are supported (Cuda toolkit >= 7.0 and
cuDNN 6.5(v2), 7.0(v3), v5) only when installing from sources.
Please see [Cuda installation](#optional-install-cuda-gpus-on-linux)
for details. for details.
## Overview ## Overview
@ -325,7 +327,7 @@ You can now [test your installation](#test-the-tensorflow-installation) within t
### (Optional, Linux) Enable GPU Support ### (Optional, Linux) Enable GPU Support
If you installed the GPU version of TensorFlow, you must also install the Cuda If you installed the GPU version of TensorFlow, you must also install the Cuda
Toolkit 7.0 and cuDNN v2. Please see [Cuda installation](#optional-install-cuda-gpus-on-linux). Toolkit 7.5 and cuDNN v4. Please see [Cuda installation](#optional-install-cuda-gpus-on-linux).
You also need to set the `LD_LIBRARY_PATH` and `CUDA_HOME` environment You also need to set the `LD_LIBRARY_PATH` and `CUDA_HOME` environment
variables. Consider adding the commands below to your `~/.bash_profile`. These variables. Consider adding the commands below to your `~/.bash_profile`. These
@ -466,20 +468,25 @@ Supported cards include but are not limited to:
https://developer.nvidia.com/cuda-downloads https://developer.nvidia.com/cuda-downloads
Install version 7.5 if using our binary releases.
Install the toolkit into e.g. `/usr/local/cuda` Install the toolkit into e.g. `/usr/local/cuda`
##### Download and install cuDNN ##### Download and install cuDNN
https://developer.nvidia.com/cudnn https://developer.nvidia.com/cudnn
Download cuDNN v4 (v5 is currently a release candidate and is only supported when
installing TensorFlow from sources).
Uncompress and copy the cuDNN files into the toolkit directory. Assuming the Uncompress and copy the cuDNN files into the toolkit directory. Assuming the
toolkit is installed in `/usr/local/cuda`, run the following commands (edited toolkit is installed in `/usr/local/cuda`, run the following commands (edited
to reflect the cuDNN version you downloaded): to reflect the cuDNN version you downloaded):
``` bash ``` bash
tar xvzf cudnn-6.5-linux-x64-v2.tgz tar xvzf cudnn-7.5-linux-x64-v4.tgz
sudo cp cudnn-6.5-linux-x64-v2/cudnn.h /usr/local/cuda/include sudo cp cudnn-7.5-linux-x64-v4/cudnn.h /usr/local/cuda/include
sudo cp cudnn-6.5-linux-x64-v2/libcudnn* /usr/local/cuda/lib64 sudo cp cudnn-7.5-linux-x64-v4/libcudnn* /usr/local/cuda/lib64
sudo chmod a+r /usr/local/cuda/lib64/libcudnn* sudo chmod a+r /usr/local/cuda/lib64/libcudnn*
``` ```