I have a Siamese neural network and I want to apply a linear transformation on the extracted image feature to reduce the dimensionality using PCA or autoencoders. how can I
