Tensorflow后端:无法使用gpu

时间:2017-11-23 09:49:01

标签: tensorflow tensorflow-gpu

以下是from tensorflow.python.client import device_lib print(device_lib.list_local_devices())

的输出
[name: "/device:CPU:0"
device_type: "CPU"
memory_limit: 268435456
locality {
}
incarnation: 8087623604945614369
]

以下是pip list | grep tensorflow的输出:

tensorflow-gpu (1.4.0)
tensorflow-tensorboard (0.4.0rc3)

我可以确认我已经在我的机器上安装了cuda 8.0和cudnn,nvidia-smi的输出显示了GPU以及其他详细信息。有人可以帮我理解为什么print(device_lib.list_local_devices())的输出没有显示GPU?rr

试过这个简单的张量流示例:

 with tf.device('/gpu:0'):

 a = tf.constant([1.0, 2.0, 3.0, 4.0, 5.0, 6.0], shape=[2, 3], name='a')
 b = tf.constant([1.0, 2.0, 3.0, 4.0, 5.0, 6.0], shape=[3, 2], name='b')
 c = tf.matmul(a, b)

 sess = tf.Session(config=tf.ConfigProto(log_device_placement=True))
 print(sess.run(c))

错误:

Operation was explicitly assigned to /device:GPU:0 but available devices are [ /job:localhost/replica:0/task:0/device:CPU:0 ]

0 个答案:

没有答案