Moonshine v2: Ergodic Streaming Encoder ASR for Latency-Critical Speech Applications

TL;DR

Moonshine v2 introduces an ergodic streaming encoder with sliding-window self-attention for low-latency, high-accuracy speech recognition on resource-constrained devices, outperforming larger models in speed and efficiency.

Contribution

It presents a novel ergodic streaming-encoder architecture using sliding-window self-attention to reduce latency while maintaining state-of-the-art accuracy.

Findings

Achieves state-of-the-art WER on standard benchmarks.

Attains accuracy comparable to models 6x larger.

Runs significantly faster with low latency.

Abstract

Latency-critical speech applications (e.g., live transcription, voice commands, and real-time translation) demand low time-to-first-token (TTFT) and high transcription accuracy, particularly on resource-constrained edge devices. Full-attention Transformer encoders remain a strong accuracy baseline for automatic speech recognition (ASR) because every frame can directly attend to every other frame, which resolves otherwise locally ambiguous acoustics using distant lexical context. However, this global dependency incurs quadratic complexity in sequence length, inducing an inherent "encode-the-whole-utterance" latency profile. For streaming use cases, this causes TTFT to grow linearly with utterance length as the encoder must process the entire prefix before any decoder token can be emitted. To better meet the needs of on-device, streaming ASR use cases we introduce Moonshine v2, an ergodic streaming-encoder ASR model that employs sliding-window self-attention to achieve bounded, low-latency inference while preserving strong local context. Our models achieve state of the art word error rates across standard benchmarks, attaining accuracy on-par with models 6x their size while running significantly faster. These results demonstrate that carefully designed local attention is competitive with the accuracy of full attention at a fraction of the size and latency cost, opening new possibilities for interactive speech interfaces on edge devices.