AI BriefingHuggingFace Blog news
The latest AI and developer news from HuggingFace Blog, with short summaries.
Visit website →Self-Speculative Decoding Accelerates Inference
Introducing Self-Speculative Decoding, a technique that leverages the early layers of the same model to improve text generation speed and memory efficiency.
Hugging Face Introduces Chunk-Based Storage
Hugging Face is introducing Content-Defined Chunking (CDC) technology to overcome the limitations of Git LFS, improving storage and transfer efficiency.
BAAI Unveils 'FlagEval Debate,' a Multilingual LLM Debate Evaluation Platform
BAAI has unveiled 'FlagEval Debate,' a multilingual debate-based platform, to address the limitations of existing LLM evaluation.
Judge Arena: Benchmarking LLM Evaluator Models
Judge Arena, a benchmarking platform that compares and ranks LLMs' evaluation capabilities, has launched.
Hugging Face Strengthens Large-Scale Dataset Hosting and Exploration Features
The Hugging Face Hub provides various features for data sharing, including large-scale dataset hosting, a data viewer, and third-party library support.
Hugging Face Integration Revealed in PyCharm
An integration feature has been released that allows users to search for Hugging Face models and insert code directly within the PyCharm IDE.
Argilla 2.4: No-Code Dataset Building Support
Argilla 2.4 has been released, enabling data labeling and feedback collection by importing datasets from the Hugging Face Hub without any code.
UAG Technology Unveiled for All Models
A technology has been developed that can speed up LLM inference using any small model, even if the tokenizer is different.
A Case Study of LLM-as-a-Judge-based RAG Evaluation
This covers a case study of building an LLM-as-a-Judge evaluation system to ensure the reliability of a RAG chatbot for agricultural support.
Multilingual-focused Aya Expanse models released
Cohere For AI unveiled the Aya Expanse 8B and 32B models, which maximize multilingual performance.
HuggingFace Releases CinePile 2.0
CinePile 2.0, a video QA dataset with improved performance through adversarial refinement techniques, has been released.
Hugging Face Launches Inference Optimization Service 'HUGS'
Hugging Face has launched HUGS, an inference service that makes it easy to deploy open models on your own infrastructure.
Google Unveils SynthID Text
Google DeepMind and Hugging Face have released SynthID Text, which embeds and detects watermarks in AI-generated text.
Hugging Face Releases Speech-to-Speech Deployment Guide
Hugging Face has released a guide on how to deploy Speech-to-Speech (S2S), a complex pipeline, to Inference Endpoints.
Outlines-core 0.1.0 Released
Outlines-core 0.1.0, a Rust port of Outlines' core algorithm for supporting structured output in LLMs, has been released.
Diffusers Supports SD 3.5
Hugging Face's Diffusers library officially supports the Stable Diffusion 3.5 Large model.
Transformers.js v3: WebGPU Support
Transformers.js v3 has been released, maximizing browser-based AI inference performance through WebGPU support.
Hugging Face Strengthens Model Security Through Partnership
Hugging Face is enhancing its model security scanning capabilities by integrating Protect AI's Guardian.
Keras Adds Llama 3.2 Support and Usage Guide
Through Keras Hub, Llama 3.2 models can be easily used across JAX, PyTorch, and TensorFlow backends.
Transformers Gradient Accumulation Fix
HuggingFace fixed an issue in the Transformers library where loss function calculation during Gradient Accumulation was mathematically inaccurate.