does anyone know how to get the logits (predictions) from the embedding_rnn_seq2seq
i need it so i could eventually calculate the accuracy
i've tried doing
correct_prediction = tf.equal(tf.argmax(decoderOutputs, 1), tf.argmax(self.decoderTargets, 1))
accuracy = tf.reduce_mean(tf.cast(correct_prediction, tf.float32))
but the decoderOutputs are not of the same shape of decoderTargets
have anyone tried calculating accuracy from the output of a seq2seq function before?