Sources
Loading...
Additional media
Loading...
DeepNewz, mobile.
People-sourced. AI-powered. Unbiased News.
Israeli company aiOla has introduced an ultra-fast, open-source speech recognition model called Whisper-Medusa, which significantly outperforms OpenAI's Whisper. The Whisper-Medusa model incorporates a multi-head attention mechanism, enabling it to adapt instantly to new jargon without requiring retraining. aiOla's model is reported to be 50% faster than OpenAI's Whisper, marking a notable advancement in the field of automatic speech recognition.
Move over, Whisper! Israeli company @_aiOla drops open-source speech recognition model that's 50% faster than OpenAI's Whisper. It uses "multi-head attention". Read details, and everything else going on in AI, including Gemini 1.5 pro win, GitHub models, and Stable Fast 3D...
new sota open source text to image model w a new ai shop out of stealth https://t.co/hOSGTlwzEr https://t.co/o3eJyJr8DA
MMS Zero-shot Released: A New AI Model to Transcribe the Speech of Almost Any Language Using Only a Small Amount of Unlabeled Text in the New Language https://t.co/bQVnHDwwOl #SpeechRecognition #AIModel #InclusiveTechnology #ZeroShotApproach #LanguageDiversity #ai #news #llm … https://t.co/hUSqUVwfsL
