In: Computer Science
Answer)
The Autoencoder is not considered to be a classifier which is the nonlinear feature with extraction mechanism and which helps in dimensional reduction mechanism that is mostly being used earlier than the classification of the high dimensional dataset for removing the redundant information from that of the data.
For training the autoencoder one would not need anything extra while one just needs to throw the input data at this and the goal is for getting the output with that of the identical and the input while the architecture would not change at all except for the fit function.
Some of the steps to be taken are as follows:
1) Using the libraries for loading, exploring, analyzing the data
2) Preprocessing the data and understanding means to rescale, resizing the data, verifying the data or even splitting the data in sets or training
3) Constructing the model and know more about the data and network
4) Converting the labels, encoding vectors, spliting training and validation images
5) Verifying values, matrices and models
6) Visualizing classification reports