https://hgpu.org/?p=24155
Efficient Deep Neural Network Inference for Embedded Systems: A Mixture of Experts Approach