FullConnectedLayer is an intermediate layer used for updating all parameters between every units.
Convert multiple vector Blob into uni Blob.
Layer is an abstraction of neural network layer.
Rectified linear unit layer
Sigmoid function layer
Sigmoid function layer
Supports only uni blob (which is a Blob includes a Vector in the channel)
Layer is an abstraction of neural network layer. This class only retains the size of input and output not coefficient and intercept. Actual parameters are kept in Model class. The parameters can be accessed with id.