I have a very basic autoencoder model. I am trying to train it on one hot encoded vector.
#make the AutoEncoder Model # this is the size of our encoded representations encoding_dim = 64 # this is our input placeholder input_vector = Input(shape=(227,)) # "encoded" is the encoded representation of the input encoded = Dense(encoding_dim, activation='sigmoid')(input_vector) # "decoded" is the lossy reconstruction of the input decoded = Dense(227, activation='sigmoid')(encoded) # this model maps an input to its reconstruction autoencoder = Model(input_vector, decoded)
Here the encoding dimension is 64. How can I calculate it according to my data ? Does it affect if I change it. I already tried but no obvious influence for now.