AI 음성-텍스트 변환: NVIDIA 카나리아-1b

Canary is a new multilingual speech-to-text recognition model from NVIDIA. According to Hugging Face, it's the best model currently available on the market, boasting the lowest error rate (about 6.67%).

And now it is available to every Latenode user!

👇 It's part of a series of applications designed to give you access to unique and otherwise hard-to-reach AI tools, thus broadening your toolkit with the most cutting-edge tools available.

The innovative Canary model by NVIDIA's NeMo team is a leap forward in multilingual speech-to-text recognition and translation, supporting English, Spanish, German, and French. Trained on a unique mix of 85,000 hours of annotated speech and refined through advanced machine translation techniques, Canary surpasses models like Whisper-large-v3 and SeamlessM4T-Medium-v1 in accuracy, despite using less data. It is now accessible on latenode.com, offering unparalleled efficiency and performance in understanding and translating spoken language.