Knowledge File / AI小生意项目库
Cohere Transcribe Arabic：专为阿拉伯语最困难转录问题打造的开源模型
Cohere发布了Cohere Transcribe Arabic，一个20亿参数的开源阿拉伯语语音识别模型，在基准测试中优于Whisper Large V3等系统。模型针对方言多样性、双语对话、代码切换和专业词汇等挑战，以Apache 2.0许可发布，可通过Hugging Face和Cohere API获取。
SOURCE / AI小生意项目库
MIN / 4
ACCESS / 会员
POST / 2026-07-08 01:54:08
原贴
查看原文原文
Cohere has released Cohere Transcribe Arabic, an open-source model built for Arabic speech recognition. The 2-billion-parameter ASR model is, according to Cohere, the most accurate open-source Arabic speech-to-text system available. It targets the specific challenges of Arabic speech, including dialect variety, bilingual Arabic-English conversations, code-switching, and specialized vocabulary. Cohere says it outscores Whisper Large V3, the standard Cohere Transcribe model, and other systems in benchmarks. The model ships under the Apache 2.0 license and is available on Hugging Face and through the Cohere API . More benchmarks and examples are on the Cohere blog . Subscribe to THE DECODER for ad-free reading, a weekly AI newsletter, our exclusive "AI Radar" frontier report six times a year, full archive access, and access to our comment section.
中文翻译
Cohere发布了Cohere Transcribe Arabic，这是一个专为阿拉伯语语音识别构建的开源模型。这个20亿参数的ASR模型据Cohere称是现有最准确的开源阿拉伯语语音转文本系统。它针对阿拉伯语语音的具体挑战，包括方言多样性、双语阿拉伯语-英语对话、代码切换以及专业词汇。Cohere表示，它在基准测试中优于Whisper Large V3、标准Cohere Transcribe模型以及其他系统。该模型以Apache 2.0许可证发布，可在Hugging Face和Cohere API上获取。更多基准测试和示例可在Cohere博客上查看。订阅THE DECODER可获取无广告阅读、每周AI通讯、每年六次的专属“AI雷达”前沿报告、完整存档访问以及评论区的访问权限。
核心信息
Cohere发布了Cohere Transcribe Arabic，一个20亿参数的开源阿拉伯语语音识别模型，在基准测试中优于Whisper Large V3等系统。模型针对方言多样性、双语对话、代码切换和专业词汇等挑战，以Apache 2.0许可发布，可通过Hugging Face和Cohere API获取。
- Cohere发布了Cohere Transcribe Arabic，一个20亿参数的开源阿拉伯语语音识别模型，在基准测试中优于Whisper Large V3等系统。模型针对方言多样性、双语对话、代码切换和专业词汇等挑战，以Apache 2.0许可发布，可通过Hugging Face和Cohere API获取。
- 原贴提到：Cohere has released Cohere Transcribe Arabic, an open-source model built
- 来源：the-decoder.com
试看内容
成为会员查看完整内容
你已经看到了这篇内容的前置整理，剩余深度部分仅对会员开放。
详细解读
信息差价值
参考来源
成为会员查看完整内容