Cohere 发布了 Cohere Transcribe Arabic，这是一个用于阿拉伯语语音识别的开源模型。据报道，该模型在转录准确性方面优于 Whisper 和 OmniASR 等现有基准，在人类评审测试中，其准确率达到 96%。Cohere 旨在将这项前沿技术提供给全球阿拉伯语使用者，模型权重可在 Hugging Face 上获取，以便在各种硬件上部署。 AI
影响 此次发布为阿拉伯语语音识别提供了一个高度准确的开源解决方案，有可能加速其在商业和开发者社区中的应用。
排序理由 前沿实验室模型发布，附带系统卡和基准测试声明。
AI 生成摘要 · Google Gemini · 来自 4 个来源。 我们如何撰写摘要 →
Sovereign AI belongs in all regions of the world. We’re working hard to make it accessible to everyone. Model weights are available on @huggingface. Run it anywhere from corporate hardware to your personal laptop. Try it out: https://t.co/gr3ZIyX4Pj
Transcribe Arabic is built to bring frontier capabilities to the millions of Arabic speakers in business and developer communities. Built to handle code-switching, multiple dialects, and Arabic-accented English, we’re taking real-life transcription to the next level.
Our model tops the Open Universal Arabic ASR Leaderboard, outperforming Whisper and OmniASR in transcription accuracy. Human reviewers also preferred us to Whisper in 96% of tests. https://t.co/12gmOS76NS
We’ve built Cohere Transcribe Arabic, the world’s most accurate open-source model for Arabic speech recognition. Available under Apache 2.0 https://t.co/KKYAjdoTX3