tensorflow 错误: Cannot assign a device for operation 'Tile'

Question

tf.tile 操作无法在 GPU 设备上执行。这段代码:

import tensorflow as tf

with tf.device('/gpu:0'):
    tt = tf.tile(tf.range(4), [3])

with tf.Session() as sess:
    print(sess.run(tt))

结果:

Caused by op u'Tile', defined at:
  File "test_tf3.py", line 6, in <module>
    tt = tf.tile(tf.range(4), [3])
  File "/media/awork/home/astepochkin/drecs/repo/env/local/lib/python2.7/site-packages/tensorflow/python/ops/gen_array_ops.py", line 3847, in tile
    name=name)
  File "/media/awork/home/astepochkin/drecs/repo/env/local/lib/python2.7/site-packages/tensorflow/python/framework/op_def_library.py", line 767, in apply_op
    op_def=op_def)
  File "/media/awork/home/astepochkin/drecs/repo/env/local/lib/python2.7/site-packages/tensorflow/python/framework/ops.py", line 2630, in create_op
    original_op=self._default_original_op, op_def=op_def)
  File "/media/awork/home/astepochkin/drecs/repo/env/local/lib/python2.7/site-packages/tensorflow/python/framework/ops.py", line 1204, in __init__
    self._traceback = self._graph._extract_stack()  # pylint: disable=protected-access

InvalidArgumentError (see above for traceback): Cannot assign a device for operation 'Tile': Could not satisfy explicit device specification '/device:GPU:0' because no supported kernel for GPU devices is available.
     [[Node: Tile = Tile[T=DT_INT32, Tmultiples=DT_INT32, _device="/device:GPU:0"](range, Tile/multiples)]]

我研究了 tensorflow 源代码并找到了为 GPU 实现图 block 的文件。所以我不明白为什么它不起作用。有人可以帮我吗？

Answer 1

这是因为你的类型是int32，并且没有为该类型注册GPU内核。