encoder output shape (batch size, sequence length, units) (64,88,1024)
encoder hidden state shape (batch size, units) (64,1024)
attention result shape (batch size, units) (64,88,1)
attention weights shape (batch_size, sequence_length,1) (64,1024)
decoder output shape (batch_size, vocab size) (64,13053)
相关文章：