沒有OpKernel登記支持歐普 'Conv2D' 這些ATTRS

Question

新的這可能是一些愚蠢的，但不能讓conv2d運行

---

窗口10

蟒蛇4.2.13

python 3.5.2

C：\ windows \ system32> nvcc --version nvcc：NVIDIA（R）Cuda編譯器驅動程序 Copyright（c）2005-2016 NV IDIA公司 建立在Sat_Sep__3_19：05：48_CDT_2016 Cuda的彙編工具，8.0版，V8.0.44

cudnn 5.1

TensorFlow 0.12

---

import numpy as np 
import tensorflow as tf 

graph1 = tf.Graph() 
with graph1.as_default(): 
    f=tf.constant( np.ones(10).reshape(1,1,-1,1) ) 
    g=tf.constant( np.ones(3).reshape(1,-1,1,1) ) 
    conv1=tf.nn.conv2d(f,g, strides=[1,1,1,1] , padding="SAME",name="conv1") 

with tf.Session(graph=graph1) as sess: 
    sess.run(tf.global_variables_initializer()) 
    print(sess.run(f)) 
    print(sess.run(g)) 
    print(sess.run(conv1)) 
    sess.close() 

結果：

InvalidArgumentError (see above for traceback): No OpKernel was registered to support Op 'Conv2D' with these attrs. Registered devices: [CPU,GPU], Registered kernels: 
    device='CPU'; T in [DT_HALF] 
    device='CPU'; T in [DT_FLOAT] 
    device='GPU'; T in [DT_HALF] 
    device='GPU'; T in [DT_FLOAT] 

    [[Node: conv = Conv2D[T=DT_DOUBLE, data_format="NHWC", padding="SAME", strides=[1, 1, 1, 1], use_cudnn_on_gpu=true](Const, Const_1)]] 

Answer 1

您應該將兩行改爲

f=tf.constant( np.ones(10).reshape(1,1,-1,1).astype(np.float32) ) 
g=tf.constant( np.ones(3).reshape(1,-1,1,1).astype(np.float32) ) 

否則這些節點採取默認numpy的類型float64，不具有Conv2D內核