Tag
推文分享了一套实用配方:找到可重复任务,用任意模型完成并记录运行结果,然后微调最小的开源权重模型(如 FastConformer)达到目标精度,从而以极小算力获得接近 SOTA 的速度与性能。作者以自己用 Parakeet 采集数据训练 FastConformer 的实际经验为例。
This paper investigates the impact of data scale versus latency on cross-lingual transfer for streaming ASR, finding that multilingual initialization benefits are data-limited, not latency-limited, and diminish as target-language data increases.