Enterprise AI company Cohere has introduced Transcribe, an open-source automatic speech recognition model designed for tasks like note-taking and speech analysis. With just 2 billion parameters, Transcribe is optimized for consumer-grade GPUs, supporting 14 languages including English, French, German, and more. Cohere’s Transcribe model outperforms competitors on the Hugging Face Open ASR leaderboard with an average word error rate (WER) of 5.42.
While Transcribe excels in most languages, it lags behind in transcribing Portuguese, German, and Spanish compared to other models. However, Cohere claims that Transcribe can process 525 minutes of audio in just one minute, positioning it as a high-performance model in its class. The company plans to integrate Transcribe into its North platform and offers it for free via its API and Model Vault.
Speech recognition models like Transcribe are gaining popularity due to the rising demand for note-taking and dictation apps. Cohere’s introduction of Transcribe showcases advancements in speech recognition technology.
Source: TechCrunch