In Miami you do AI demos on the beach with will.i.am FYI Hugging Face

How it started / How it's going

The quantized version of Idefics 2 gets a 2.27x speed up M3 Max.

9 tokens/s —> 51.5 tokens/s 🔥🚀

Llama 3 70B takes the king’s crown 👑 from GPT-4 Turbo

- 100% in coverage and 70% in quality of code
- 💵 most cost-effective inference
- open-weight

We tested OpenAI's GPT 3.5 & 4, Meta's Llama 3, Google's Gemini 1.5 Pro, cohere's Command R+ and 130+ other LLMs

The main goal of my trip to San Francisco next week is to hire our first Hugging Face Community Lead there. To make it 🌶️🌶️🌶️, I’ll do it in public & won’t leave the city before we find them. Time to make it happen folks!

More Options for open LLMs on Hugging Face! 🤗 Learn how to deploy AI at Meta Llama 3 using vLLM on Hugging Face Inference Endpoints. 🚀
We created a detailed blog post showing you how to use vLLM as a custom container within Inference Endpoints to deploy any vLLM support…

TGI v2.0.2 is out!
- New models (idefics2, phi3)
- Cleaner VLM support in the openai layer
- Upgraded to pytorch 2.3.0

github.com/huggingface/te…

👏 Nicolas Patry and contributors!

Now Qwen1.5-110B is on the top among pretrained models in Open LLM Leaderboard😉

Getting in the mood for the weekend! Anyone has tickets for the HF team?

Released Hermes 2 Pro on Llama-3 8B today!

Get it here: huggingface.co/NousResearch/H…

or the GGUF here: huggingface.co/NousResearch/H…

Announcing Hermes 2 Pro on Llama-3 8B! Nous Research's first Llama-3 based model is now available on HuggingFace.

Hermes Pro comes with Function Calling and Structured Output capabilities, and the Llama-3 version now uses dedicated tokens for tool call parsing tags, to make…

🎉🤗 #SnowflakeArctic is a top trending model on Hugging Face!

Get started with Arctic today by going to Hugging Face to directly download Arctic and use our Github repo for inference and fine-tuning recipes: huggingface.co/models

Wow the 15T tokens dataset was just released and so many downloads already 🤯

PSA:

ASKING AN LLM ABOUT ITSELF IS NOT A RELIABLE WAY OF INVESTIGATING THE LLM

'Oh the model told me it's made by OpenAI' so what?!

I can't believe smart people keep making this rookie mistake!

It's not just another Monday. Today, you check out the Hugging Face 🤗 demo from the Developer Keynote at #GoogleCloudNext → goo.gle/49ZsS2g

I'll be in San Francisco/the Bay Area next week. What are some cool startups using HF or AI in general there?

MAU 📈 from Gradio! 1M when?

The April leaderboards are here! 🤗 Meet @HuggingFace’s most impactful users (models, datasets, spaces, followers):

🏆 Top Contributors To Follow - huggingface.co/spaces/mvaloat…

Create domain-specific synthetic datasets in under 20 minutes?

Check out Argilla's Domain Specific Dataset Project to generate chat datasets tailored to a domain. I used it to craft a dataset focused on data-centric machine learning (how meta 😅): huggingface.co/datasets/davan….