We’re open-sourcing our entire STT benchmarking methodology.
Normalization rules, scoring scripts, reproducibility, all of it.
Because no model can be the best at everything, and everyone deserves transparency about this.
Watch MadMax, our CTO at @gladia_io, explain it.
A huge step for me, and the team that, I hope, will benefit everyone - founders, customers, and the community.
Test and challenge it. Repo + results here: https://t.co/9Ce3m9L3Ce
Here is a new automatic speech recognition system with no hallucinations and much more accurate than Whisper:
Its name is Whisper-Zero, and it's a big improvement over OpenAI's model.
• It removes 99.9% of hallucinations
• It dramatically increases accuracy
• It's much faster
• It's much simpler to use
• It comes with no limitations
I recorded a live demo using their playground. It's fast and very accurate! More impressive: it switches between different languages without skipping a bit.
This is not a toy. Whisper-Zero's goal is to work in real-life environments.
Some additional notes:
• It's hallucination free
• Supports custom vocabulary
• It does automatic noise reduction
• Supports word-level timestamps
• Supports speaker diarization
• It works with accents, even thick ones!
You can use Whisper-Zero with any audio or video file type, including large files of up to 500MB and 135 mins. It also supports processing any URL to a video or audio file.
You can check Whisper-Zero here: https://t.co/UCmGxfLMIr.
Thanks to the @gladia_io team for partnering with me on this post.
"For me, success is the ability to realize your dreams ... to turn your ideas into something that exists and works," says Anna Valeri Jelezovskaia, an alum of the Bachelor of Laws (LL.B.) and a content strategist at @gladia_io . Listen to her full story: https://t.co/Qu6npVnEnY
Our seed round and product launch are featured in @TechCrunch! And according to @romaindillet, @gladia_io is the best transcription API he's ever used...We're humbled and thrilled. Off we go team!! https://t.co/9Ve3zhAQle
It's been hell of a ride, but here we are!! Audio Intelligence API by @gladia_io is out now! ⚡️ More accurate than ever, fast, and priced to scale, with speaker diarization, word-level timestamp, and code-switching in 99 languages in the core offer.
It's official: Gladia's production-ready API is out now, fully revamped for stability! 🚀 More accurate than ever, fast, and priced to scale, the API includes transcription, diarization, word-level timestamp, and code-switching, all supported in 99 languages!
🎙Exciting news! Introducing the CS Café podcast ☕️ hosted by @CobolStone 🌐 Join us as we dive into the amazing world of #Security, #AI and more! 🚀
📺 Watch on YouTube: https://t.co/uBhAERFnNk
🎧 Listen on Spotify: https://t.co/f352Z2LAsL
Working in AI makes me question many things. For instance, how can we possibly create responsible AI without properly understanding / agreeing on what actually makes us human (which we don’t). If AI is to perpetuate this ignorance, we’re def in for a crazy ride.
MAJOR ANNOUNCEMENT
Our 🎙 Audio Transcription API 🎙 is now live! Built on @OpenAI Whisper-Large-v2, 10s to transcribe a 1h file, and more accurate than everything else on the market 🔥
I’m so so proud of the team, and of the tech we created!!
Why?
🌐1 Minute in the Metaverse🌐
Ep 8: A Space Odyssey of the Metaverse
w/@pabloiquiroga, Co-Founder & CRO at @staratlasgame
What if we could build a perfect world in space?👽🚀
Star Atlas is turning to web3 gaming to give it a try. Here's how.
#staratlas#powder#gaming#web3