“…The evolution of DNNs has already piqued interest in hardware acceleration as both DNNs training and inference demand a tremendous amount of computation. As a result, hardware accelerators such as GPUs, FPGAs, and customized ASICs have been employed to accelerate DNNs [4], [5], [6], [7], [8], [9], [10], [11], [12], [13], [14], [15], [16], [17], [18], [19], [20], [21], [22], [23], [24], [25]. However, DNN designers are still hampered by the growing complexity of DNN models.…”