Self-Attentive Implementation of "A Structured Self-Attentive Sentence Embedding" Result Accuracy: 0.85 Visualization Multi annotation vector - multi r hops Single annotation vector - single r hop Reference https://arxiv.org/abs/1703.03130 https://github.com/flrngel/Self-Attentive-tensorflow