模型多源确认
网易有道 Confucius4-R2T2 发布 4 天登顶 Hugging Face ASR 热榜
精选理由
网易有道的语音识别模型，上线4天就拿了Hugging Face ASR榜第一，做实时转写的可以看看他们论文里的流式方案。
网易有道发布流式语音识别模型 Confucius4-R2T2，上线 4 天登上 Hugging Face ASR Trending 榜第一。目前含量化版本在内全球累计下载量超过 20 万次。作者还读了其配套论文，梳理了该模型流式语音识别的实现思路。
原文 · Ate-a-Pi
The way these guys are doing streaming speech recognition is really cool.
#1 on Hugging Face's ASR Trending list: Confucius4-R2T2 from @NetEaseYouDaoAI. It hit #1 within 4 days of release.
It now has 200K+ cumulative downloads worldwide, including quantized versions.
I read their paper, and here is what you should know:
https://t.co/jbJblFt2AO
