This is a tool to visualize the distribution of importance scores over timesteps (example words) in a sequence classification task. As you hover your mouse over the decoded words, the tool shows a heatmap of importance scores over the source words.
To run the visualizer, run
python -m SimpleHTTPServer
from this directory then navigate to http://localhost:8000/ in browser. The visualizer will show some example data.
To visualize your own data, you need to replace imp_scores_test.json with a similar file which provides the importance scores over the sequences, the sequences themselves, and the gold and the predicted labels. In particular imp_scores_test.json should contain the following fields:
seq_lst: multiple sequences as a list of list of words.imp_scores: the importance scores of every word for every sequence. List of list.pred: List of predicted labels for every sequencegold: List of gold labels for every sequence
WARNING: Make sure that none of the strings in seq_lst, imp_scores, pred or gold contain <angled brackets>. These will interfere with the HTML and can result in text not being displayed.