search

carlthome / tensorflow-convlstm-cell

A ConvLSTM cell with layer normalization and peepholes for TensorFlow's RNN API.
MIT License
399 stars 150 forks source link

About padding before each time step #30

Closed PeiYingjun closed 5 years ago

PeiYingjun commented 6 years ago

Thanks for this implementation, it helps save my strength, yet there's something confuses me about ConvLSTMCell: to ensure h and x have the same row and column, don't you think we should add padding op before each time step instead of using static_rnn directly?

PeiYingjun commented 5 years ago

I see now, if we set param padding = 'same' and strides = 1, then we get the same size as inputs, no matter how large is the kernel_size