This is a simple classifier for detecting "Code", "Tags", "Text", and "Title" of StackeOverflow data. Scraped 2M+ questions and answers from StackOverflow and trained different model to classfy the above mentioned classes.
Achieved 89% accuracy for DNN model. see details- https://github.com/pritom007/CodeTextClassifier/blob/master/DNN/report.txt
After the training we use the classifier in our search engine. Read details here- https://github.com/KevinCT/SearchFlow