arXiv cs.CL· Sahil Mahendrakar·· 10 小时前AI 评分42
Earlybird：29M 参数的预测式流式语音识别模型
Earlybird: Predictive Streaming Speech Recognition in 29M Parameters
AI 导读
Earlybird 是一个 29M 参数（4 bit 量化后 16.8 MB）的流式语音识别模型，能在说话人说完单词前就预测出该词。在 LibriSpeech test-clean 上，中位单词比说话人说完提前 90 ms 被正确预测，而已发布的流式模型要滞后 90 至 400 ms。在全部六个测试集上它都是最早输出的一方，同套有声书上 69% 的单词在结束前被正确预测，其他模型仅 3-25%。
正文 · 原文
Abstract:Streaming speech recognizers output a word only after the speaker has finished it, typically 90-400 ms later. For a voice agent every millisecond counts: the sooner the words are outputted, the sooner the agent can start a search or plan its reply. We present Earlybird, a 29M-parameter streaming recognizer (16.8 MB at 4 bits). Earlybird learns from audio cut off at random points, including partway through a word, and is asked to predict every word that is due by then in full, plus a guess of the next two words. Its raw token probabilities give a confidence for every word including the one still being spoken. On clean audiobook speech (LibriSpeech test-clean), the median word is predicted correctly 90 ms before the speaker finishes it, while released streaming models get it 90 to 400 ms after it ends. Earlybird is the earliest on all six test sets we time, covering audiobooks, TED talks, public speech, earnings calls, and synthetic speech. On the same audiobooks, 69% of words are predicted correctly before they end, against 3-25% for the other models. For words still being spoken, confidence is calibrated, with an expected calibration error (ECE) of 3.7 to 5.5. On the eight English Open ASR Leaderboard sets, it is more accurate than Moonshine Tiny on every set and roughly ties Whistle, a model of the same size.
| Comments: | 9 pages, 3 figures, 6 tables. Model: this https URL |
| Subjects: | Sound (cs.SD); Computation and Language (cs.CL); Audio and Speech Processing (eess.AS) |
| Cite as: | arXiv:2611.14421 [cs.SD] |
| (or arXiv:2611.14421v1 [cs.SD] for this version) | |
| https://doi.org/10.48550/arXiv.2611.14421 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Sahil Mahendrakar [view email]
[v1]
Tue, 10 Nov 2026 17:59:28 UTC (15 KB)
来源：arXiv cs.CL · arxiv.org