Problem in Convolutional autoencoder.

Hey everyone,

As a newbie I am working on audio feature extraction using convolutional autoencoder.After applying convolutional layers when i give this 2D matrix to training network it gives error "Not enough input arguments". How to train this data?