For the last 5-6 years I have been kind of obsessed (in a good way) with how perhaps we could try to get computational machine learning models for drug discovery to a point where they could be shared. The reasoning behind this being that we publish papers, but the models described in them never really …