在张量流中，我试图计算相对于表示为layer_out 的隐藏层，表示为输出的n层NN模型的输出的梯度i = 1,2，...，n。
的任何一层输出的形状是[128,10]，layer_out的形状是[128,100]，其中128是批量大小。我为此目的使用以下命令。
l_grad = tf.gradients(output, layer_out)[0]
我期望的渐变形状是[10,100]或[100,10]。假设梯度是整个批次的平均值。但是，我得到的形状是[128,100]
我尝试过的一个途径是在梯度计算之前使用tf.reduce_mean（）来平均批次中的两个量。但是，结果是无列表，并显示以下错误
Traceback (most recent call last):
File "Test_Gen_Error_MNIST_EDL.py", line 128, in <module>
Results[i,4], Results[i,5], Results[i,6] = Analyse_custom_Optimizer_EDL_old(X_train,y_train,X_test,y_test, Kappa_s[i])
File "Test_Gen_Error_MNIST_EDL.py", line 60, in Analyse_custom_Optimizer_EDL_old
model = model.init_NN_custom(classes, 1e-03, depth, tf.sigmoid,'EDL')
File "/usr/local/home/krm9c/Documents/Research/ErrorDriven-Learning/MN/Class_gen_Error.py", line 265, in init_NN_custom
Error_Loss, self.classifier['class'], array, dec)
File "/usr/local/home/krm9c/Documents/Research/ErrorDriven-Learning/MN/Class_gen_Error.py", line 192, in direct_feedback_alignement
print_grad_shape = tf.Print(l_grad, [tf.shape(l_grad)], 'L_gradient shape')
File "/usr/local/home/krm9c/anaconda2/envs/deep/lib/python2.7/site-packages/tensorflow/python/ops/array_ops.py", line 276, in shape
return shape_internal(input, name, optimize=True, out_type=out_type)
File "/usr/local/home/krm9c/anaconda2/envs/deep/lib/python2.7/site-packages/tensorflow/python/ops/array_ops.py", line 300, in shape_internal
input_tensor = ops.convert_to_tensor(input)
File "/usr/local/home/krm9c/anaconda2/envs/deep/lib/python2.7/site-packages/tensorflow/python/framework/ops.py", line 932, in convert_to_tensor
as_ref=False)
File "/usr/local/home/krm9c/anaconda2/envs/deep/lib/python2.7/site-packages/tensorflow/python/framework/ops.py", line 1022, in internal_convert_to_tensor
ret = conversion_func(value, dtype=dtype, name=name, as_ref=as_ref)
File "/usr/local/home/krm9c/anaconda2/envs/deep/lib/python2.7/site-packages/tensorflow/python/framework/constant_op.py", line 233, in _constant_tensor_conversion_function
return constant(v, dtype=dtype, name=name)
File "/usr/local/home/krm9c/anaconda2/envs/deep/lib/python2.7/site-packages/tensorflow/python/framework/constant_op.py", line 212, in constant
value, dtype=dtype, shape=shape, verify_shape=verify_shape))
File "/usr/local/home/krm9c/anaconda2/envs/deep/lib/python2.7/site-packages/tensorflow/python/framework/tensor_util.py", line 401, in make_tensor_proto
raise ValueError("None values not supported.")
ValueError: None values not supported.
我在这里遗漏了一些东西，但我无法弄明白。任何方向都非常感谢。 谢谢