@osanseviero Profile picture

Omar Sanseviero

@osanseviero

Llama farmer ex-Chief Llama Officer @huggingface 🦙 Founder @AI_Learners. Xoogler (SWE at Google Assistant) 100% Hacker Llama🇵🇪🇲🇽

Similar User
Lilian Weng photo

@lilianweng

clem 🤗 photo

@ClementDelangue

Hugging Face photo

@huggingface

Julien Chaumond photo

@julien_c

Gradio photo

@Gradio

Aran Komatsuzaki photo

@arankomatsuzaki

Soumith Chintala photo

@soumithchintala

Philipp Schmid photo

@_philschmid

Sasha Rush photo

@srush_nlp

Nils Reimers photo

@Nils_Reimers

Sylvain Gugger photo

@GuggerSylvain

Lysandre photo

@LysandreJik

Thomas Wolf photo

@Thom_Wolf

Jay Alammar photo

@JayAlammar

Lewis Tunstall photo

@_lewtun

Working on a timeline of open transformer models. Which ones are missing?

Tweet Image 1

Microsoft released LLM2CLIP, a technique in which a LLM acts as a teacher for CLIP's visual encoder. 🧠Unlocks longer/complex captions 🤗Apache 2 licensed 🚀Larger models are being trained Website: microsoft.github.io/LLM2CLIP/ Models: hf.co/collections/mi…

Tweet Image 1

Omar Sanseviero Reposted

Releasing two trillion tokens in the open. huggingface.co/blog/Pclanglai…

Tweet Image 1

AlphaFold should be on @huggingface 🤗


Omar Sanseviero Reposted

🚀Now it is the time, Nov. 11 10:24! The perfect time for our best coder model ever! Qwen2.5-Coder-32B-Instruct! Wait wait... it's more than a big coder! It is a family of coder models! Besides the 32B coder, we have coders of 0.5B / 1.5B / 3B / 7B / 14B! As usual, we not only…

Tweet Image 1

Omar Sanseviero Reposted

Trust me next week will be totally different


If you felt this week was very slow in open-source ML...it's because the Hugging Face team is in a company-wide offsite and ML just slowed down by 10 months 😭


Omar Sanseviero Reposted

Yesterday, we had @osanseviero in Madrid enjoying delicious croquetas and torreznos from Casa Mortero with @jjmata, Álvaro Correa from Taxdown, Álvaro, our PM at @graphext, and myself while we chatted about LLMs, AI, and other life stuff :)

Tweet Image 1

Omar Sanseviero Reposted

🚀 Today, we are introducing SmolTools! 🚀 Last week, at Hugging Face we made a significant leap forward with the release of SmolLM2, a compact 1.7B language model that sets a new benchmark for performance among models of its size. But beyond the impressive stats, SmolLM2 truly…


Omar Sanseviero Reposted

Gracias a @multimodalart y a @huggingface, tenemos un espacio en HuggingFace donde usar el modelo F5 en castellano de forma gratuita: huggingface.co/spaces/jpgalle…


Omar Sanseviero Reposted

🚀 Excited to introduce a new member of the OS-Copilot family: OS-Atlas - a foundational action model for GUI agents Paper: huggingface.co/papers/2410.23… Website: osatlas.github.io A thread on why this matters for the future of OS automation 🧵 TL;DR: OS-Atlas offers: 1.…

Tweet Image 1

10 big Open ML releases from Oct 28 - Nov 1st 1. 🎨Stable Diffusion 3.5 Medium by StabilityAI. A 2B image generation model with a permissive license 2. 📹LongVU by Meta, a video LM that can handle long videos (1 to 7B params) 3. 🗣️MaskGCT by the Chinese University of Hong Kong.…

Tweet Image 1

I'll be in Madrid from Thursday->Monday. Are any AI events happening? 🦙


Omar Sanseviero Reposted

Introducing SmolLM2: the new, best, and open 1B-parameter language model. We trained smol models on up to 11T tokens of meticulously curated datasets. Fully open-source Apache 2.0 and we will release all the datasets and training scripts!

Tweet Image 1

6 years ago, I had a pneumothorax (🫁collapse) 1 year ago, I had a long-lasting plantar fasciitis 3 months ago, I was able to resume running and training seriously for the first time I'm super excited because I ran my first Half Marathon🏃Next stop: the full marathon in April!

Tweet Image 1

Each time I review a new agent paper 😅 Agents present very interesting engineering, data, and research challenges - long context, prompt formatting for tools, etc., but the way they are being marketed feels like a huge revolution when it's...prompts with extra steps/tools?

Tweet Image 1

Omar Sanseviero Reposted

We just released the weights of Pixtral 12B base model on HuggingFace: Pixtral 12B Base: huggingface.co/mistralai/Pixt… Also link to Pixtral 12B Instruct: huggingface.co/mistralai/Pixt…


This week in open ML 🤯 - IBM Granite - Allegro by Rhymes for video generation - Stability Stable Diffusion 3.5 - Genmo Mochi for video gen - Moonshine for speech recognition on edge - Cohere multilingual Aya Expanse - Microsoft OmniParser for screen parsing - Meta Quant LLamas


Loading...

Something went wrong.


Something went wrong.