WAND: Windowed Attention and Knowledge Distillation for Efficient Autoregressive Text-to-Speech Models
This addresses efficiency bottlenecks for text-to-speech applications, offering incremental improvements in model deployment.
The paper tackled the high memory and compute costs of autoregressive text-to-speech models by proposing WAND, a framework that uses windowed attention and knowledge distillation to achieve constant complexity, resulting in up to 66.2% KV cache memory reduction and near-constant latency while preserving quality.
Recent decoder-only autoregressive text-to-speech (AR-TTS) models produce high-fidelity speech, but their memory and compute costs scale quadratically with sequence length due to full self-attention. In this paper, we propose WAND, Windowed Attention and Knowledge Distillation, a framework that adapts pretrained AR-TTS models to operate with constant computational and memory complexity. WAND separates the attention mechanism into two: persistent global attention over conditioning tokens and local sliding-window attention over generated tokens. To stabilize fine-tuning, we employ a curriculum learning strategy that progressively tightens the attention window. We further utilize knowledge distillation from a full-attention teacher to recover high-fidelity synthesis quality with high data efficiency. Evaluated on three modern AR-TTS models, WAND preserves the original quality while achieving up to 66.2% KV cache memory reduction and length-invariant, near-constant per-step latency.