Fig. 5From: LTPConstraint: a transfer learning based end-to-end method for RNA secondary structure predictionArchitecture of LTPConstraint Network. Input module 1 after the sequence data goes through the Embedding layer. Pairwise concat is used for the output of module 1 and then input module 2. Until module 2, it is pretrained network. Finally, the preTrained network was modified with the hard constraint layer of module 3 to get the outputBack to article page