ValueError: when making a template about convolutional-lstm-in-tensorflow HOT 3 OPEN

Could you send a snippet of the network creation + input code?

from convolutional-lstm-in-tensorflow.

hmm, Im not really sure whats happening with just this info. I think it might be that you need to specify the batch size or you are giving it the wrong input_shape. Note that the convlstm stuff is merged into tensorflow now and a little example can be seen here https://github.com/loliverhennigh/dynamic_rnn_conv_lstm/blob/master/mnist_deep.py#L57.

from convolutional-lstm-in-tensorflow.

@loliverhennigh Thanks for your reply.
I have a few questions want to ask you.

1. Is the ''number of hidden states'' in the convlstm layer (mentioned in the paper: Convolution LSTM network: ...for precipitation nowcasting) equal to the ''number of feature maps'' (output_channels) in this example ?
2. Taking this mnist_deep as example, how can I stack multiple convlstm ?
3. Have you tried the 3d convlstm for forecasting/classifying the images ? How is the performance ?.

I am quite new in tensorflow, wish to know some implementation experience in tf, especially for convlstm.

Thanks in advance.

from convolutional-lstm-in-tensorflow.