Google DeepMind officially launched Gemma 3n along with its developer guide. The Gemma series, as Google's open-weights model family, has long been a favorite…
LeRobot, the open-source robotics project under Hugging Face, has officially released version v0.4.0 — a major update aimed at comprehensively accelerating…
Google DeepMind has officially announced the addition of a highly distinctive and specialized new member to its open-source model family — Gemma 3 270M. This…
Google DeepMind has officially launched VaultGemma, currently the world's most capable large language model (LLM) trained entirely from scratch using…
Hugging Face has announced the launch of a new open-source project called "OpenEnv," aimed at collaboratively building an open and standardized execution…
As open-source AI models have grown explosively, model security has become an issue that can no longer be ignored. Traditional machine learning model formats…
In the fields of natural language processing (NLP) and vector retrieval, Sentence Transformers — founded by Nils Reimers — has long been the industry-standard…
Hugging Face recently published a feature article on "AI for Food Allergies" in its "Hugging Science" column. Food allergies are a global health concern…
Google Cloud has announced a deep collaboration with Intel and Hugging Face on its new C4 instances to comprehensively optimize open-source GPT (Generative…
Visual Language Models (VLMs) combine computer vision with natural language processing, enabling complex tasks such as image captioning and visual question…
As "Sovereign AI" becomes a global trend, countries around the world are actively seeking to build AI models that reflect their own culture, values, and…
Arm has officially announced on the Hugging Face blog that it will actively participate in the upcoming PyTorch Conference. As the Arm architecture gains…
Hugging Face and the BigCode community have jointly launched a new code model evaluation platform called "BigCodeArena." As AI-assisted coding (such as Copilot…
Cloud AI model hosting platform Replicate has announced official support for IBM's latest Granite 4.0 model family. This means developers and enterprise users…
As AI Agent applications become increasingly widespread, running large language models (LLMs) efficiently on personal computers (such as AI PCs powered by…
### Background and Challenge: The Rise of Local "Computer Use" With Anthropic's introduction of Computer Use and the development of various OS-level agents…
ServiceNow AI recently published a post on the Hugging Face blog introducing a brand-new open-source framework called "SyGra" — a one-stop synthetic data…
AI agents are currently the hottest research direction in the AI field, but how to objectively, safely, and reproducibly evaluate agent capabilities has long…
Hugging Face has announced a deep partnership with Scaleway, a leading European cloud infrastructure provider, with Scaleway officially joining the Hugging…
Hugging Face continues to expand its "Inference Providers" program, aimed at enabling developers to run open-source models from Hugging Face Hub in the…
Hugging Face has officially released `LeRobotDataset:v3.0`, a critical technical upgrade to its open-source robot learning library `lerobot`, with the core…
With the explosive growth of generative AI — particularly in image and multimedia generation — how to identify AI-generated content and protect copyright has…
Vercel officially announced the launch of "402-mcp" in its Changelog — an important update that injects commercial capabilities into the Model Context Protocol…
Vercel has officially launched a new open protocol called "x402-mcp," designed to establish a standardized payment and billing mechanism for Model Context…
Writer, a leading provider of enterprise AI solutions, has officially announced the launch of its new "Palmyra-mini" model series on the Hugging Face platform…
### Background and the LLM Inference Bottleneck When running large language models (LLMs), autoregressive generation is inherently "memory-bandwidth-bound"…
Hugging Face and Together AI have announced a deep partnership, launching a new integration designed to streamline the fine-tuning workflow for open-source…
### Background and Core Concepts Traditional large language models (LLMs), when faced with complex mathematics, data analysis, or programming tasks, can…
In today's era dominated by generative AI and large language models (LLMs), bidirectional encoder models (such as BERT and RoBERTa) still play an indispensable…
When deploying modern AI models (such as LLaMA, Flux, or Stable Diffusion), `torch.compile` — introduced in PyTorch 2.0 — is a powerful performance…