Structured Self-attentive sentence embeddings

Implementation for the paper A Structured Self-Attentive Sentence Embedding, which was published in ICLR 2017: .


For binary sentiment classification on imdb dataset run : python "binary"

For multiclass classification on reuters dataset run : python "multiclass"

You can change the model parameters in the model_params.json file Other tranining parameters like number of attention hops etc can be configured in the config.json file.

If you want to use pretrained glove embeddings , set the use_embeddings parameter to "True" ,default is set to False. Do not forget to download the glove.6B.50d.txt and place it in the glove folder.


  • Classification using self attention
  • Regularization using Frobenius norm
  • Gradient clipping
  • Visualizing the attention weights

Instead of pruning ,used averaging over the sentence embeddings.


After training, the model is tested on 100 test points. Attention weights for the 100 test data are retrieved and used to visualize over the text using heatmaps. A file visualization.html gets saved in the visualization/ folder after successful training. The visualization code was provided by Zhouhan Lin (@hantek). Many thanks.

Below is a shot of the visualization on few datapoints. alt text

Training accuracy 93.4% Tested on 1000 points with 90.2% accuracy


GitHub - kaushalshetty/Structured-Self-Attention: A Structured Self-attentive Sentence Embedding
A Structured Self-attentive Sentence Embedding. Contribute to kaushalshetty/Structured-Self-Attention development by creating an account on GitHub.