ValueError: Dimensions must be equal

Error code:
File "/usr/local/lib/python2.7/dist-packages/tensorflow/python/framework/common_shapes.py", line 691, in _call_cpp_shape_fn_impl
    raise ValueError(err.message)

ValueError: Dimensions must be equal, but are 1024 and 596 for 'rnn/while/rnn/multi_rnn_cell/cell_0/basic_lstm_cell/MatMul_1' (op: 'MatMul') with input shapes: [100,1024], [596,2048].

 

Source code:

    # Build the RNN layers
    
    lstm = tf.contrib.rnn.BasicLSTMCell(lstm_size)
    drop = tf.contrib.rnn.DropoutWrapper(lstm, output_keep_prob=keep_prob)
    cell = tf.contrib.rnn.MultiRNNCell([drop] * num_layers)

    initial_state = cell.zero_state(batch_size, tf.float32)

    # Run the data through the RNN layers
    outputs, state = tf.nn.dynamic_rnn(cell, x_one_hot, initial_state=initial_state)
    final_state = state

修改成：

    def get_a_cell(lstm_size, keep_prob):
        lstm = tf.contrib.rnn.BasicLSTMCell(lstm_size)
        drop = tf.contrib.rnn.DropoutWrapper(lstm, output_keep_prob=keep_prob)
        return drop
    
    cell = tf.contrib.rnn.MultiRNNCell([get_a_cell(lstm_size, keep_prob) for _ in range(num_layers)])

    initial_state = cell.zero_state(batch_size, tf.float32)

    # Run the data through the RNN layers
    outputs, state = tf.nn.dynamic_rnn(cell, x_one_hot, initial_state=initial_state)
    final_state = state
   

 运行成功！

 

cell = tf.contrib.rnn.MultiRNNCell([drop] * num_layers) 不能使用？ 具体原因不清楚，理论上和cell = tf.contrib.rnn.MultiRNNCell([get_a_cell(lstm_size, keep_prob) for _ in range(num_layers)])应该一样的效果，建议不要使用[drop] * num_layers这种用法。