anantzoid opened this issue 9 years ago · 1 comments
A typo in seq2seq_f where param name embedding_size is provided while calling tf.nn.seq2seq.embedding_attention_seq2seq.
seq2seq_f
embedding_size
tf.nn.seq2seq.embedding_attention_seq2seq
As per the definition of embedding_attention_seq2seq, the first 6 arguments don't have a name.
embedding_attention_seq2seq
Issue resolved on updating Tensorflow.