Pith. sign in

Com- pared to the static baseline in Table 2, our dynamic models can save up to 29.6 % of MACs while only incurring a 0.75 % drop in PESQ

CONCLUSION We presented ConvFSE-Net, a neural network architecture for SE, extended it with a dynamic pruning system that learns to skip unnecessary convolutional channels base

1 Pith paper cite this work. Polarity classification is still indexing.

1 Pith paper citing it
browse 1 citing papers

fields

eess.AS 1

years

2024 1

verdicts

CONDITIONAL 1

representative citing papers

Scalable Speech Enhancement with Dynamic Channel Pruning

eess.AS · 2024-12-22 · conditional · novelty 5.0

A custom convolutional speech enhancement network with a learned gating module skips individual channels at runtime, saving up to 29.6% of MACs on VoiceBank+DEMAND with a negligible PESQ drop.

citing papers explorer

Showing 1 of 1 citing paper.

  • Scalable Speech Enhancement with Dynamic Channel Pruning eess.AS · 2024-12-22 · conditional · none · ref 7

    A custom convolutional speech enhancement network with a learned gating module skips individual channels at runtime, saving up to 29.6% of MACs on VoiceBank+DEMAND with a negligible PESQ drop.