SSVAE is a Generative Model for molecular design that enables the generation of new molecules according to desired physicochemical properties.

This is the updated code for Tensorflow 2.0+ for the model described in the paper "Conditional molecular design with deep generative models". See the paper here. As well, the Tensorflow 1.0 code can be found here.

• SSVAE2.0 molecule conditional prediction now supports an arbitrary number of molecule properties.
• The way batches are generated is up to date and implemented in the full_preprocessing.py.
• The stop condition for the training is now defined as its own class.

• Tensorflow 2.0+
• Scikit-learn
• RDKit
• Pandas/NumPy