The AI community building the future.
External link for Hugging Face
We’re on a journey to solve and democratize artificial intelligence through natural language.
Paris, FR
Hugging Face reposted this
We just made it much easier to build and share your MCP Servers with Gradio! 🔥 MCP is now selected by default (instead of Python) when you click on the 'Use via API' page. 🧭 The chosen language now appears as a query parameter in the web address, making it accessible programmatically.
Hugging Face reposted this
🔥🆕ObjectClear, an object removal model capable of eliminating target object and its effects (shadow etc.) Object Clear app on Hugging Face: https://lnkd.in/g8NgaSkB
Hugging Face reposted this
🎉 Big news! Google Colab now comes with Gradio pre-installed (v5.38)! No more pip install gradio needed - just import and start building AI apps instantly. 🙌 Thanks to Chris Perry and the whole Google Colab team for making Gradio more accessible to millions of developers worldwide!
Hugging Face reposted this
🚀 Big big news for multimodal devs! The transformers ↔️ vLLM integration just leveled up: Vision-Language Models are now supported out of the box If the model is integrated into Transformers, you can now run it directly with vLLM — no need to rewrite or duplicate code. Just plug it in and go. Zero extra effort Performance might differ model to model (we’re working on that!), but functional support is guaranteed Curious how to serve Transformers models with vLLM? Full docs here 👉 https://lnkd.in/d-KjqbmU #multimodal #transformers #vLLM #VLM #opensource
Hugging Face reposted this
🔥 Vibe coding for data science with Kimi K2 - a step-by-step tutorial you can run for free! I'm excited to share a new approach to working with unstructured datasets, powered by the Hugging Face Hub and open models! https://lnkd.in/dBwFrS_g
Hugging Face reposted this
Mixture of Experts? No no, Roundtable of Experts! Meet CONSILIUM - Where AI Models Debate Around a Table! Amazing results! 👇 Let's learn more 📊 THE VALIDATION Microsoft proved why multi-agent setup matters: - AI panel of "doctors": 85.5% accuracy / Human doctors: 20% accuracy - 4X BETTER at diagnosis - While CUTTING costs More here: AI Diagnostic Orchestrator (MAI-DxO) : https://lnkd.in/dF-fDzcM 🎰 WHAT IS CONSILIUM? Built during Agents-MCP hackathon, Consilium seats LLMs around a visual poker table where they: - Debate complex questions with distinct roles - Research using 5+ external sources - Reach consensus through structured discussion - Show real-time "thinking" bubbles It's like having an expert panel on demand ⚡ 🎭 The breakthrough? Each LLM gets a personality: - 🔥 Expert Advocate: Passionate specialist - 🔍 Critical Analyst: Finds flaws & risks - 🎯 Strategic Advisor: Practical solutions - 📚 Research Specialist: Deep domain knowledge - ⚡ Innovation Catalyst: Challenges conventional thinking Real and longer debate = better decisions 🧠 🌐 RESEARCH AGENT MAGIC : When models need facts, a dedicated research agent joins the table and searches following sources -- - 🔍 Web Search - 📖 Wikipedia - 📄 arXiv papers - 💻 GitHub - 📈 SEC EDGAR Recorded demo: https://lnkd.in/gMzrgqJK 🛠️ TECHNICAL BRILLIANCE -- Built as : - A Gradio UI (that poker table is 🔥) - MCP server for integrating Consilium in your workflows - A Custom Gradio component - Session-based state management - Real-time visual updates Developer experience = 10/10 ✨ MCP demo: https://lnkd.in/gaEAVUMS 💡REAL-WORLD APPLICATIONS -- Imagine using this for: - 🏥 Medical diagnosis (already proven) - 💼 Business strategy decisions (ooh 🤩) - 🔬 Research paper analysis - ⚖️ Legal case evaluation Any complex decision = perfect use case 🎯 Play with the App: https://lnkd.in/gK-w6sYA 🔥 Consilium is built by Andreas Zettl 📖 READ MORE ON CONSILIUM in this Hugging Face Report -- Consilium: When Multiple LLMs Collaborate: https://lnkd.in/gmT65peC
You can now seamlessly switch between providers like Groq, SambaNova, Together AI, Cerebras Systems and more and LLMs like Kimi K2, DeepSeek R1 and 6000+ models using the OpenAI client 🤗 Time to build is now! 🔥
You asked we delivered! Hugging Face Inference Providers is now fully OpenAI client compatible! 🔥 Simply append the provider name to the model ID and.. that's it! Read more here: https://lnkd.in/eNsZsmci OpenAI client is arguably the most used client when it comes to LLMs, so getting this right is a big milestone for the team! 🤗
Hugging Face reposted this
Fighting with prompts for data augmentation is a thing of the past. Here's a workflow where Hugging Face AISheets handles few-shot and context engineering for you! 1. Load your dataset (csv, parquet, Excel, HF hub) 2. Write a VERY simple prompt (e.g., categorize the following text). 3. Generate a few rows and inspect the results. Even SOTA models like Kimi K2 will make mistakes. 4. While inspecting the results, mark good examples with the thumbs-up button. 5. Regenerate and profit! Under the hood, AISheets will manage the context in the prompt (adding few-shot examples and web docs if search is enabled). This will instantly improve the quality of the generated cells! Check the dataset I've built in the comments. If you like this, try it out and leave ❤️ to support us: https://lnkd.in/dVu2f2SB
Excited to bring the collective power of Inference Providers to Pydantic AI 🔥 Go check it out!
Pydantic AI now supports Hugging Face as a provider! Hugging Face (https://huggingface.co/) is an AI platform with all major open source models, datasets, MCPs, and demos. You can use it to run open source models like DeepSeek R1 on scalable serverless infrastructure. They have a free tier allowance so you can test it out. Thanks to the Hugging Face team (Célina Hanouti) for this great contribution. For more details, check out the docs (https://lnkd.in/eTwF-dHi)
Hugging Face reposted this
New Tutorial in the MCP Course! This tutorial walks you through building a tiny agents application using AMD hardware. 🔗 https://lnkd.in/efv9bk8H - You can use Neural Processing Unit (NPU) and integrated GPU (iGPU) from AMD - Sets up local file MCP for dealing with sensitive files It's sick to see local assistants make it on to even more hardware.