0
From the Hugging Face Hub to robot hardware with Strands Agents and LeRobot
Hugging Face detailed how to deploy models from its hub onto physical robotic hardware using Strands Agents and the LeRobot library.
Hugging Face Blog
Tier 2 source · latest first
0
GLM-5.2: Built for Long-Horizon Tasks
The GLM-5.2 model has been introduced with optimizations specifically designed for handling long-horizon tasks.
0
Agentic Resource Discovery: Let agents search
Hugging Face explored the concept of agentic resource discovery, focusing on enabling AI agents to autonomously search for and locate digital resources.
0
Profiling in PyTorch (Part 2): From nn.Linear to a Fused MLP
This technical guide from Hugging Face explores performance profiling in PyTorch, focusing on optimizing neural network layers from standard linear operations to fused multi-layer perceptrons.
0
How an Agent Built a 3D Paris Gallery by Chaining Two Hugging Face Spaces
A new demonstration shows how an AI agent successfully constructed a 3D virtual gallery of Paris by linking two separate Hugging Face Spaces.
0
Migrating Your GitHub CI to Hugging Face Jobs
Hugging Face has published a guide detailing how developers can transition their continuous integration workflows from GitHub to Hugging Face Jobs.
0
The Open Source Community is backing OpenEnv for Agentic RL
The open-source community has rallied behind OpenEnv, a collaborative environment designed to advance agentic reinforcement learning research.
0
Nemotron 3.5 Content Safety: Customizable Multimodal Safety for Global Enterprise AI
NVIDIA has released Nemotron 3.5 Content Safety, a customizable multimodal safety model designed to help enterprises filter and secure AI-generated content.
0
Designing the hf CLI as an agent-optimized way to work with the Hub
Hugging Face has redesigned its command-line interface to optimize interactions with the Hub for autonomous AI agents.
0
Direct Preference Optimization Beyond Chatbots
Hugging Face explores applications of Direct Preference Optimization beyond conversational agents to other domains.
0
Adding MCP Tools to Reachy Mini
Hugging Face has integrated Model Context Protocol tools into the Reachy Mini robot platform.
0
Holo3.1: Fast & Local Computer Use Agents
Hugging Face has introduced Holo3.1, a system designed for running fast, local computer-use agents.
0
Introducing Mellum2: A 12B Mixture-of-Experts Model by JetBrains
JetBrains has released Mellum2, a 12-billion parameter Mixture-of-Experts model designed for developer workflows.
0
Beyond LLMs: Why Scalable Enterprise AI Adoption Depends on Agent Logic
An article on Hugging Face argues that successful enterprise AI adoption requires shifting focus from large language models to scalable agent logic.
0
Profiling in PyTorch (Part 1): A Beginner's Guide to torch.profiler
Hugging Face has published an introductory guide on using the torch.profiler tool to analyze performance in PyTorch.
0
Shipping a Trillion Parameters With a Hub Bucket: Delta Weight Sync in TRL
Hugging Face has introduced a delta weight synchronization method in its TRL library to optimize the transfer of extremely large models.
0
Reachy Mini goes fully local
The Reachy Mini robot has been updated to support fully local processing and execution.
0
Harness, Scaffold, and the AI Agent Terms Worth Getting Right
This article clarifies key terminology used to describe AI agent architectures, focusing on terms like harness and scaffold.
0
OlmoEarth v1.1: A more efficient family of Earth observation models
The OlmoEarth v1.1 family of models has been released, offering improved efficiency for Earth observation and geospatial analysis tasks.
0
Introducing the Ettin Reranker Family
A new family of search and retrieval models called Ettin Rerankers has been introduced on Hugging Face.
0
PaddleOCR 3.5: Running OCR and Document Parsing Tasks with a Transformers Backend
PaddleOCR 3.5 has been released, integrating a Transformers backend to improve optical character recognition and document parsing tasks.
0
Granite Embedding Multilingual R2: Open Apache 2.0 Multilingual Embeddings with 32K Context — Best Sub-100M Retrieval Quality
IBM has released Granite Embedding Multilingual R2, an open-source Apache 2.0 licensed embedding model featuring a 32K context window and high retrieval quality for models under 100 million parameters.
0
Unlocking asynchronicity in continuous batching
A new technical post on Hugging Face explores methods for enabling asynchronous operations within continuous batching systems to optimize LLM inference.
0
Building Blocks for Foundation Model Training and Inference on AWS
Hugging Face has outlined the essential infrastructure components required for training and deploying foundation models on Amazon Web Services.