Skip to main content

Table 5 Formulas used to calculate the complexity of the networks

From: Performance vs. hardware requirements in state-of-the-art automatic speech recognition

Network complexity Formula
Parameters (model size) in fully connected layers input sizeoutput size+bias
Parameters (model size) in time-delay layers (input sizeoutput size+bias)context size
Parameters (model size) in convolutional layers (filter sizenumber of input filters+bias)number of output filters
Parameters (model size) in recurrent layers (input size+recurrent layer size)output sizenumber of gatesdirectionality factor
Multiply-accumulate operations (MACs) parametersfeatures vector lengthsequence length in time
Operations (Ops) MACs2
Activations in fully connected layers and time-delay layers output sizeoutput sequence length
Activations in convolutional layers output sizenumber of output filtersoutput sequence length
Activations in recurrent layers recurrent layer sizeoutput sequence lengthdirectionality factor