我用tensorflow实现CNN，但是遇到了tensorflow.python.framework.errors_impl.InvalidArgumentError：答案

【问题标题】：I used tensorflow to implement CNN,but I met tensorflow.python.framework.errors_impl.InvalidArgumentError:我用tensorflow实现CNN，但是遇到了tensorflow.python.framework.errors_impl.InvalidArgumentError：
【发布时间】：2017-04-23 05:54:59
【问题描述】：

Traceback (most recent call last):
  File "C:/Users/ljz/PycharmProjects/TFConvLSTM/LSTMnetwork.py", line 136, in <module>
    sess.run(train_step, feed_dict={x: X_train[start:end], y_: y_train[start:end]})
  File "C:\Users\ljz\AppData\Local\Programs\Python\Python35\lib\site-packages\tensorflow\python\client\session.py", line 767, in run
    run_metadata_ptr)
  File "C:\Users\ljz\AppData\Local\Programs\Python\Python35\lib\site-packages\tensorflow\python\client\session.py", line 965, in _run
    feed_dict_string, options, run_metadata)
  File "C:\Users\ljz\AppData\Local\Programs\Python\Python35\lib\site-packages\tensorflow\python\client\session.py", line 1015, in _do_run
    target_list, options, run_metadata)
  File "C:\Users\ljz\AppData\Local\Programs\Python\Python35\lib\site-packages\tensorflow\python\client\session.py", line 1035, in _do_call
    raise type(e)(node_def, op, message)
tensorflow.python.framework.errors_impl.InvalidArgumentError: Incompatible shapes: [1500,18] vs. [22500,18]
     [[Node: mul = Mul[T=DT_FLOAT, _device="/job:localhost/replica:0/task:0/cpu:0"](_recv_Placeholder_1_0, Log)]]

Caused by op 'mul', defined at:
  File "C:/Users/ljz/PycharmProjects/TFConvLSTM/LSTMnetwork.py", line 113, in <module>
    cross_entropy=tf.reduce_mean(-tf.reduce_sum(y_*tf.log(y_conv),reduction_indices=[1]))
  File "C:\Users\ljz\AppData\Local\Programs\Python\Python35\lib\site-packages\tensorflow\python\ops\math_ops.py", line 884, in binary_op_wrapper
    return func(x, y, name=name)
  File "C:\Users\ljz\AppData\Local\Programs\Python\Python35\lib\site-packages\tensorflow\python\ops\math_ops.py", line 1105, in _mul_dispatch
    return gen_math_ops._mul(x, y, name=name)
  File "C:\Users\ljz\AppData\Local\Programs\Python\Python35\lib\site-packages\tensorflow\python\ops\gen_math_ops.py", line 1625, in _mul
    result = _op_def_lib.apply_op("Mul", x=x, y=y, name=name)
  File "C:\Users\ljz\AppData\Local\Programs\Python\Python35\lib\site-packages\tensorflow\python\framework\op_def_library.py", line 763, in apply_op
    op_def=op_def)
  File "C:\Users\ljz\AppData\Local\Programs\Python\Python35\lib\site-packages\tensorflow\python\framework\ops.py", line 2395, in create_op
    original_op=self._default_original_op, op_def=op_def)
  File "C:\Users\ljz\AppData\Local\Programs\Python\Python35\lib\site-packages\tensorflow\python\framework\ops.py", line 1264, in __init__
    self._traceback = _extract_stack()

InvalidArgumentError (see above for traceback): Incompatible shapes: [1500,18] vs. [22500,18]
     [[Node: mul = Mul[T=DT_FLOAT, _device="/job:localhost/replica:0/task:0/cpu:0"](_recv_Placeholder_1_0, Log)]]

【问题讨论】：

你的问题是什么？

标签： tensorflow

【解决方案1】：

Caused by op 'mul', defined at: File "C:/Users/ljz/PycharmProjects/TFConvLSTM/LSTMnetwork.py", line 113, in cross_entropy=tf.reduce_mean(-tf.reduce_sum(y_*tf.log(y_conv),reduction_indices=[1])) 和

InvalidArgumentError (see above for traceback): Incompatible shapes: [1500,18] vs. [22500,18] [[Node: mul = Mul[T=DT_FLOAT, _device="/job:localhost/replica:0/task:0/cpu:0"](_recv_Placeholder_1_0, Log)]]

=> 确保 y 和 y_conv 具有相同或可广播的形状。

【讨论】：

【解决方案2】：

有问题的代码： cross_entropy=tf.reduce_mean(-tf.reduce_sum(y_*tf.log(y_conv),reduction_indices=[1]))

当您尝试将两个张量 y_ 和 y_conv 相乘时，它们必须具有相同的形状，即要么都是 [1500,18] 要么都是 [22500,18]

在您的情况下，y_ 应该是输入标签，y_conv 应该是计算的 logits，18 应该是输出神经元的数量。因此我猜你的输入数据（x）和标签（y_）的批量大小不一样。检查它并使它们相同。

【讨论】：