ROSE-CD distills a robust one-step consistency model for speech enhancement that outperforms its 30-step diffusion teacher on quality metrics while running 54 times faster.
End-to-End Multi-Task Denoising for joint SDR and PESQ Optimization
2 Pith papers cite this work. Polarity classification is still indexing.
2
Pith papers citing it
verdicts
UNVERDICTED 2representative citing papers
Shared-score quaternion self-attention reduces score multiplications by 75% and softmax operations from four to one while proving equivalence to component-wise attention under quaternion linear projections.
citing papers explorer
-
Robust One-step Speech Enhancement via Consistency Distillation
ROSE-CD distills a robust one-step consistency model for speech enhancement that outperforms its 30-step diffusion teacher on quality metrics while running 54 times faster.
-
Quaternion Self-Attention with Shared Scores
Shared-score quaternion self-attention reduces score multiplications by 75% and softmax operations from four to one while proving equivalence to component-wise attention under quaternion linear projections.