Multi modal

• COATI: A pre-trained, multi-modal encoder-decoder model designed for navigating and representing chemical space.
  

  

• DRAGONFLY: Prospective de novo drug design with deep interactome learning
  

  

• DrugEx: Library for de novo drug design using RNNs, Transformers within a multi-objective reinforcement learning framework
  

  

• Sc2Mol: A scaffold-based two-step molecule generator that combines variational autoencoders with transformers to generate molecules, supporting batch random generation for efficiency.