MOGNET: A mux-residual quantized network leveraging online-generated weights

authors

• Nguyen Van Thien
• Guicquero William
• Sicard Gilles

keywords

• CNN
• Quantized neural networks
• Skip connections
• Channel attention
• Logic-gated CNN
• Cellular Automaton

abstract

This paper presents a compact model architecture called MOGNET, compatible with a resource-limited hardware. MOGNET uses a streamlined Convolutional factorization block based on a combination of 2 point-wise (1×1) convolutions with a group-wise convolution in-between. To further limit the overall model size and reduce the on-chip required memory, the second point-wise convolution's parameters are on-line generated by a Cellular Automaton structure. In addition, MOGNET enables the use of low-precision weights and activations, by taking advantage of a Multiplexer mechanism with a proper Bitshift rescaling for integrating residual paths without increasing the hardware- related complexity. To efficiently train this model we also introduce a novel weight ternarization method favoring the balance between quantized levels. Experimental results show that given tiny memory budget (sub-2Mb), MOGNET can achieve higher accuracy with a clear gap up to 1% at a similar or even lower model size compared to recent state-of-the-art methods.

more information