Thrilled to celebrate 5 years of AlphaFold 2! It’s now been used by over 3 million researchers around the world to accelerate their vital research - and it was an honour of a lifetime for our work to be recognised last year with the Nobel Prize! Proof of AI’s potential to enable science at digital speed 🚀
To honour the anniversary, we’ve made The Thinking Game film available for free on our YouTube channel - it’s a great look behind the scenes of AlphaFold & our journey to AGI.
open-source OCR models are super cheap to run and privacy first 🤝
BUT there's a ton of new models out there: DeepSeek-OCR, Nanonets, PaddleOCR, how do you pick them? 🤯
don't worry though, @huggingface got you covered! 🫡🧶
🚀 DeepSeek-OCR — the new frontier of OCR from @deepseek_ai , exploring optical context compression for LLMs, is running blazingly fast on vLLM ⚡ (~2500 tokens/s on A100-40G) — powered by vllm==0.8.5 for day-0 model support.
🧠 Compresses visual contexts up to 20× while keeping 97% OCR accuracy at <10×.
📄 Outperforms GOT-OCR2.0 & MinerU2.0 on OmniDocBench using fewer vision tokens.
🤝 The vLLM team is working with DeepSeek to bring official DeepSeek-OCR support into the next vLLM release — making multimodal inference even faster and easier to scale.
🔗 https://t.co/rnBG9VUuMy
#vLLM #DeepSeek #OCR #LLM #VisionAI #DeepLearning
@josasis@joanvilallanars@tradiactivitat No contradiu la teva afirmació. Tenies raó. Les inversions en CapEx son tan besties que soc bullish tant a AMD i NVIDIA. AMD la seva part de CPU🚀🚀
La plataforma Rubin té algun problema? 🤔 Sobre el paper quan arribi Helios de AMD, la Rubin hauria de tornar a ser molt superior.
@josasis@joanvilallanars@tradiactivitat Opino que NVIDIA sempre va un parell de passos per endevant dels fabricants ASICs i AMD, pero esta clar que ningu vol tenir tanta dependencia de una empresa i diversifiquen.
La métrica clau sera tokens/GWatts disponibles, i Jensen Huang porta anys preparant aixo.
Algun consell?
@josasis@joanvilallanars@tradiactivitat Si et refereixes a AMD no crec.
En el tema conectivitat GPUs en racks no tenen rival (a part de que CUDA superior a ROC, encara que ROC estigui millorant).
@tradiactivitat Se sap qui son aquests 2 clients?
MSFT? Open AI, Oracle, Meta. xAI?
Crec que els nous GPU no tenen prou capacitat de produccio, i prioritzen els clients vip estrategics. Si tinguessin més oferta de producte estaria més diversificat.
@siemprepositifo La herencia deixada de Johan Cruyff es la més important de la historia del club. Period.
Pero no entenc perque aquesta insistencia en denigrar el Nuñez. Tan mal president va ser? En lo económic ho va fer bé.
Fer fora al Cruyff, fins a cert punt llogic per ficar a Jordi Jr de 9.
@marquillO_os@botealberto Claro. La recupera desde fuera y la pone por la puerta dentro del campo del rival.
Es habitual en padel, pero aqui Tapia tenia poquisimo angulo, y fue a recuperar el smash x3 de Galan casi hasta Cuenca.
Check out the 999 open models that Google has released on @huggingface:
https://t.co/Fo4Ycn9ARi
(Comparative numbers: 387 for Microsoft, 33 for OpenAI, 0 for Anthropic).