Fig. 1
From: Prediction of plant secondary metabolic pathways using deep transfer learning
Flowchart of the proposed GTC. The whole pipeline contains two processes: pre-training and fine-tuning. The model is initially constructed on the KEGG dataset and then leverages the learned knowledge to predict plant secondary metabolic pathways. Conv1D stands for one-dimensional convolutional layer; Pool1D stands for one-dimensional max-pooling layer; FC stands for fully connected layer