dssm model gragh.py forward() logits = tf.concat([pos_result, neg_result], axis=1) there why pos_result put forward? Is wrong?
yangyan1227 opened this issue · 1 comments
yangyan1227 commented
dssm model gragh.py forward() logits = tf.concat([pos_result, neg_result], axis=1) there why pos_result put forward? Is wrong?
yangyan1227 commented
dssm model gragh.py forward() logits = tf.concat([pos_result, neg_result], axis=1)
there why pos_result put forward? Is wrong?