Closed Deephome closed 5 years ago
Actually I used an open-source implementation for ConvLSTM.
I think we did not use that function containing _num_units
, so it will not cause error in practice.
You can switch to official implementation from TensorFlow instead. tf.contrib.rnn.ConvLSTMCell
@jiangsutx Thank you for your reply!
Hi, I have some questions:
Looking forward to your reply!