0
vLLM V0 to V1: Correctness Before Corrections in RL
The Hugging Face blog details the transition of the vLLM engine from version 0 to 1, emphasizing reinforcement learning correctness.
Hugging Face Blog
Tier 2 source · latest first
0
Adding Benchmaxxer Repellant to the Open ASR Leaderboard
Hugging Face has introduced a mechanism called Benchmaxxer Repellant to its Open ASR Leaderboard to improve evaluation integrity.
0
Granite 4.1 LLMs: How They’re Built
A technical overview has been published detailing the architecture and training methodology behind the Granite 4.1 large language models.
0
DeepInfra on Hugging Face Inference Providers 🔥
DeepInfra has been integrated as an inference provider on the Hugging Face platform to offer scalable model hosting.
0
Introducing NVIDIA Nemotron 3 Nano Omni: Long-Context Multimodal Intelligence for Documents, Audio and Video Agents
NVIDIA has introduced Nemotron 3 Nano Omni, a lightweight multimodal model designed for long-context processing of documents, audio, and video.
0
How to build scalable web apps with OpenAI's Privacy Filter
This guide explains how developers can utilize OpenAI's Privacy Filter to construct scalable web applications.
0
DeepSeek-V4: a million-token context that agents can actually use
DeepSeek has introduced DeepSeek-V4, a new model featuring a highly usable one-million-token context window optimized for AI agents.
0
How to Use Transformers.js in a Chrome Extension
This guide explains how developers can integrate Transformers.js into Google Chrome extensions to run machine learning models directly within the browser.
0
QIMMA قِمّة ⛰: A Quality-First Arabic LLM Leaderboard
Hugging Face introduced QIMMA, a new evaluation leaderboard focused on assessing the quality of Arabic large language models.
0
AI and the Future of Cybersecurity: Why Openness Matters
Hugging Face discussed the role of open-source artificial intelligence in strengthening cybersecurity defenses and fostering collaborative security research.
0
Training and Finetuning Multimodal Embedding & Reranker Models with Sentence Transformers
Hugging Face published a technical guide on training and fine-tuning multimodal embedding and reranker models using the Sentence Transformers library.
0
The PR you would have opened yourself
Hugging Face detailed a new automated pull request feature designed to streamline code contributions and repository maintenance.
0
Ecom-RLVE: Adaptive Verifiable Environments for E-Commerce Conversational Agents
Researchers introduced Ecom-RLVE, a framework designed to provide adaptive and verifiable environments for training e-commerce conversational agents.
0
Inside VAKRA: Reasoning, Tool Use, and Failure Modes of Agents
Hugging Face has published an analysis of VAKRA, examining how AI agents handle reasoning, tool integration, and common failure modes.
0
Meet HoloTab by HCompany. Your AI browser companion.
HCompany has introduced HoloTab, an AI-powered browser companion designed to assist users with web navigation and tasks.
0
Waypoint-1.5: Higher-Fidelity Interactive Worlds for Everyday GPUs
Hugging Face introduced Waypoint-1.5, a framework designed to generate high-fidelity interactive virtual environments on consumer-grade GPUs.
0
Multimodal Embedding & Reranker Models with Sentence Transformers
Hugging Face detailed the integration of multimodal embedding and reranker models within the Sentence Transformers library.
0
Safetensors is Joining the PyTorch Foundation
The Safetensors format is officially joining the PyTorch Foundation to foster open-source collaboration and standardization in model security.
0
Welcome Gemma 4: Frontier multimodal intelligence on device
Google has introduced Gemma 4, a frontier multimodal model designed to bring advanced intelligence directly to on-device applications.
0
Falcon Perception
The Technology Innovation Institute has introduced Falcon Perception, a new model or framework focused on perception capabilities within the Falcon open-source AI family.
← NewerPage 3