huxian0402/convlstm-moving-mnist

Using convlstm to prediction moving mnist dataset.

convlstm-moving-mnist's Issues

RNN_convLSTM structure Error

# RNN_convLSTM  structure
def RNN_convLSTM(x ):

    #(1,19,64,64,1)     #num_filter, kernei_size
    x = tf.contrib.slim.conv2d(x, 1, 1, padding='SAME',activation_fn=tf.nn.relu)
    x = tf.contrib.slim.conv2d(x, 8, 1, padding='SAME', activation_fn=tf.nn.relu)
    x = tf.contrib.slim.conv2d(x, 16, 1, padding='SAME', activation_fn=tf.nn.relu)
    #(1,10,64,64,16)
   
                                      # [3 ,                       3]      ,                1024
    rnn_cell = BasicConvLSTMCell([64, 64], [config.conv_filter_size, config.conv_filter_size], config.hidden_size,
                                 is_train,
                                 forget_bias=1.0, activation=tf.nn.tanh)

    state_size = rnn_cell.state_size  #c(64,64,1024) ,h(64,64,1024)

我应该把tf.contrib.slim.conv2d 改成tf.contrib.slim.conv3d吗？，因为前者的输入只能是一个4维tensor,会报错的！ 求解？

Convolution expects input with rank 4, got 5