The open-source data curation and annotation platform Argilla has officially released version 2.4, with the core of this update being deep integration with…
In the deployment and inference of large language models (LLMs), reducing generation latency has always been a critical challenge. The traditional approach of…
This case study provides a detailed account of how non-profit organization Digital Green, with support from Hugging Face's Expert Support team, optimized its…
Cohere For AI (C4AI) has officially launched Aya Expanse, a family of open-weight models designed specifically for multilingual tasks. The family includes two…
CinePile is a multimodal question-answering dataset focused on movie and long-video understanding. In traditional dataset construction, researchers commonly…
On October 23, 2024, Google and Hugging Face jointly announced the open-sourcing of Google's "SynthID Text" technology and its integration into Hugging Face's…
Hugging Face has officially launched HUGS (Hugging Face Microservices), a brand-new microservices solution designed to address the pain points enterprises face…
In LLM application development, ensuring that a model outputs content that 100% conforms to a specific format — such as a JSON Schema, a regular expression, or…
As real-time voice interaction technologies like GPT-4o become more widespread, the open-source community is also actively developing speech-to-speech (S2S)…
As open-source AI models have flourished, Hugging Face has become the central platform through which developers worldwide access and share machine learning…
Stability AI officially launched the Stable Diffusion 3.5 (SD3.5) model series in late October 2024, and Hugging Face's Diffusers team simultaneously announced…
Meta's Llama 3.2 release includes lightweight 1B and 3B text models designed specifically for edge computing and mobile devices. These models have now been…
As generative AI applications become more widespread, one of the biggest challenges developers face is the "non-deterministic" output of large language models…
### The Mathematical Flaw in Traditional Gradient Accumulation Gradient accumulation is an extremely common technique in deep learning. When VRAM is limited…
The well-known AI model hosting platform Replicate has announced a significant speed improvement for FLUX image generation models running on its platform. FLUX…
As AI applications become increasingly widespread, Gradio has become one of the most popular tools for developers to showcase and deploy machine learning…
AMD has officially launched its 5th-generation EPYC processor, codenamed "Turin," and Hugging Face has promptly published a blog post detailing the deep…
Gradio, one of the most popular rapid prototyping tools for AI in the Python ecosystem, has officially reached the major milestone of Gradio 5. This update is…
As the scale of AI models and the volume of training data grow dramatically, the computational capacity and memory (RAM) of a single machine often become…
Hugging Face has published a technical blog post on "Dynamic Speculation," aimed at optimizing the inference speed of large language models (LLMs)…
The Hugging Face Hub, as the world's largest open-source AI community and dataset hosting platform, automatically converts datasets uploaded in various formats…
Hugging Face has officially launched the "Open FinLLM Leaderboard" — a new platform dedicated to evaluating and tracking the performance of large language…
This article from the Hugging Face blog takes an in-depth look at how China's artificial intelligence forces have successfully gone global in recent years…
The Hugging Face team and its collaborators have jointly launched a new benchmark called "BenCzechMark," designed to evaluate the understanding and generation…
In the field of 3D generative AI (encompassing models such as InstantMesh and Tripo3D), generated 3D models typically represent color using "vertex coloring."…
Meta has officially introduced the Llama 3.2 family of open-source models, marking a significant architectural upgrade with two major breakthroughs: multimodal…
With the explosion of video generation and understanding models such as Sora and Gen-3, high-quality video training data has become a key battleground for…
In today's rapidly evolving AI landscape, keeping up with the latest academic papers is a major challenge. Hugging Face's "Daily Papers" page was created…
This article provides a detailed look at how to use Hugging Face's `optimum-intel` library and Intel's OpenVINO GenAI toolkit to optimize and deploy generative…
The deployment of large language models (LLMs) has long faced a dual bottleneck of VRAM capacity and memory bandwidth. Microsoft previously introduced the…