Dynamic nsNet2: Efficient Deep Noise Suppression with Early Exiting
This is an incremental improvement for efficient audio processing on devices with limited resources.
The paper tackles the challenge of deploying deep noise suppression models on resource-constrained devices by introducing an early-exiting version of nsNet2 that allows halting computations at different stages, showing trade-offs between performance and computational complexity with established metrics.
Although deep learning has made strides in the field of deep noise suppression, leveraging deep architectures on resource-constrained devices still proved challenging. Therefore, we present an early-exiting model based on nsNet2 that provides several levels of accuracy and resource savings by halting computations at different stages. Moreover, we adapt the original architecture by splitting the information flow to take into account the injected dynamism. We show the trade-offs between performance and computational complexity based on established metrics.