The inspiration came from the tough days of organic chemistry study of colleges where it was difficult to predict the reactions and we thought of a concept that using deep learning and seq2seq model can we predict the reaction and came the success.
What it does
If we provide canonical SMILES for organic compounds we can predict the reaction using Deep learning and Seq2seq model We have to pass two or more canonical smiles separated by a dot such as *.@@@@ it will give us the result(replace ** and @@@@ with appropriate SMILES )and click on the box for create reaction it will give us result.
How we built it
At back-end we have used Tensorflow and Seq2seq model for doing predictions.Its completely html5 based.
Challenges we ran into
Creating the model was tough we had to try different ways to get it right.
Accomplishments that we're proud of
The complete project was a rewarding journey for us we learnt a lot and this will help so many people.
What we learned
Implementing Deep learning in web.
What's next for Organic Compounds predictor using Open VINO
Complete end to end game based system where people can share the reaction in Facebook too.