How to use other GPUs in Keras with TensorFlow support?

Question

This is due to How to enable Keras with Theano to use multiple GPUs , but instead of using multiple GPUs, I am interested in indicating which GPU is using a particular model or running.

My conclusion is nvidia-smias follows:

+------------------------------------------------------+                       
| NVIDIA-SMI 361.42     Driver Version: 361.42         |                       
|-------------------------------+----------------------+----------------------+
| GPU  Name        Persistence-M| Bus-Id        Disp.A | Volatile Uncorr. ECC |
| Fan  Temp  Perf  Pwr:Usage/Cap|         Memory-Usage | GPU-Util  Compute M. |
|===============================+======================+======================|
|   0  Tesla K80           Off  | 0000:03:00.0     Off |                    0 |
| N/A   38C    P0    60W / 149W |  11354MiB / 11519MiB |      0%      Default |
+-------------------------------+----------------------+----------------------+
|   1  Tesla K80           Off  | 0000:04:00.0     Off |                    0 |
| N/A   37C    P0    71W / 149W |    224MiB / 11519MiB |      0%      Default |
+-------------------------------+----------------------+----------------------+
|   2  GeForce GTX 750 Ti  Off  | 0000:06:00.0      On |                  N/A |
| 40%   29C    P8     1W /  38W |    120MiB /  2047MiB |      0%      Default |
+-------------------------------+----------------------+----------------------+

This conclusion, of course, does not work. The problem is that I'm not sure about Keras how to specify which GPU to work on. Of course, with TensorFlow we can just make a paradigm with tf.device('/cpu:1'):, but I'm not sure how this will integrate with Keras.

Thank!

Answer 1

, tensorflow keras.json , / CUDA_VISIBLE_DEVICES (http://acceleware.com/blog/cudavisibledevices-masking-gpus). , .

Answer 2

, :

• tensorflow, . https://www.tensorflow.org/versions/r0.7/get_started/os_setup.html

• Keras Theano . tensorflow

vi ~/.keras/keras.json

: { "epsilon": 1e-07, "floatx": "float32", "backend": "theano" }

"theano" "tensorflow"

, tenorflow gpu. gpu.

. cuda tensorflow http://www.nvidia.com/object/gpu-accelerated-applications-tensorflow-installation.html