OpenAI’s Whisper models fail up to 300% more often on African accents vs. standard English. We tested 20+ African and 40+ global accents Here’s why this matters 👇

@julien_c @joinstationf Can you make us cut the line plsss ?

@aiandcivil75700 @ZyphraAI Hey you can try it here :)

The space is available here:

@reach_vb @Gautam_A_k wait wat, im the intern here

Xavier Niel, one of the major investor of Mistral said that Mistral is going to have a model that will kill DeepSeek R1 in two-three months. Mmmmhhhh

Having a climbing wall and a sauna at the office is really what keeps me coming

Model weights here:

RT @Tu7uruu: WE ARE SO BACK! Some really exciting news for France 🇫🇷 > French president Macron, just announced a 109 BILLIONS Investment in…

@qtnx_ Make it part of your jobs and you can post as much as you want 🤗

@AnthropicAI ok cool, but where is sonnet 4 ?

RT @JiaLi52524397: 🚀 NuminaMath 1.5 is here! 🚀 900k+ high-quality competition math problems with CoT solutions, new problem metadata, manua…

🚨A new Math dataset released! NuminaMath 1.5 is now available and features over 900,000 high-quality competition math problems complete with chain-of-thought solutions. The update also includes enhanced problem metadata and a selection of manually verified Olympiad problems. The dataset is accessible on HuggingFace, with special thanks given to @JiaLi52524397 @Will424408 and @dsleo for their contributions! Check it out here!

@miniapeur Agree, I get more likes when posting funny meme. Kinda boost my ego regarding my skills in meme but it also makes me wonder if people appreciate my actual content as much as the humor :')

I think we dont talk enough about the dopamine shot when you get a lower eval loss

The model is available here: