Unable to understand the generated documentation wrt input code cells

[Tamal-Mondal]

Hello,

I have run the pre-trained model with the dataset publicly available(https://www.dropbox.com/s/vpsst1el7f0jqo6/data_notebookcdg.pkl?dl=0) and have got the predicted documentations as output. But we can't understand relation between output documentation and input codecells as the available input data is preprocessed.

So I request you to please share the kaggle notebooks that you have used as training and test data(only test data is also fine). This would help us greatly to understand the output of HAConvGNN.