How you save the NN inputs from the autoencoder ?

[abhisekbakshi]

After analyzing your code thoroughly I understood that the Autoencoders have extracted the high-level features that act as input to the fully connected layers. Both of the autoencoders have extracted 64 feathers of the drugs and proteins. However, you have read a train and a test file containing the concatenation of the 64 features of each autoencoder. Thus NN input size is 128. Unfortunately, I could not understand how the file is created. In other words, how the features are extracted from the encoder's output and concatenated. Please help me to understand this.