The archive · 11 labs · 871 dispatches

The labs

No more opening a dozen official blogs every morning. First-hand releases from OpenAI, Anthropic, DeepMind and the rest, each with its substance pulled out.

451

Banque des Territoires, Polyconseil, and Hugging Face Deploy Sovereign RAG Solution for EduRénov Program

Hugging Face, Polyconseil, and Banque des Territoires launched a sovereign, open‑source RAG system to automate email support for France's EduRénov school‑renovation program, ensuring data residency while scaling generative AI for public policy.

452

Hugging Face Dataset Hub Search Features Update

Hugging Face has introduced four new search filters—modality, size, format, and library compatibility—to improve the discoverability of over 180,000 public datasets on the Dataset Hub.

453

Accelerating Protein Language Model ProtST on Intel Gaudi 2

Intel and MILA have optimized the ProtST multi-modal protein language model for Intel Gaudi 2 accelerators, achieving up to 2.92x faster fine-tuning compared to NVIDIA A100 GPUs.

454

Hugging Face Transformers Code Agent GAIA Benchmark Results

Hugging Face achieved a top ranking on the GAIA benchmark using a Code Agent built with the transformers.agents library, demonstrating that code-based actions are more efficient and effective than JSON-based tool calling.

455

Google Gemma 2 Release Notes

Google has released Gemma 2, a family of open-weight LLMs available in 9B and 27B parameter sizes, featuring technical advances in distillation and attention mechanisms to improve performance.

456

XLSCOUT ParaEmbed 2.0 Release

XLSCOUT has released ParaEmbed 2.0, a proprietary embedding model fine-tuned on expert-curated patent data that achieves a 23% increase in accuracy over ParaEmbed 1.0.

457

Hugging Face Ethics and Society Newsletter #6: The Importance of Data Quality

Hugging Face outlines a holistic, responsible approach to data quality, emphasizing that high-quality data must be fit for its intended purpose to ensure AI model performance, fairness, and scientific reproducibility.

458

Fine-tuning Microsoft Florence-2 for DocVQA

Hugging Face demonstrates how to fine-tune Microsoft's Florence-2 vision-language model on the DocVQA dataset, improving validation similarity from 0 to 57.0 after seven epochs.

459

Hugging Face Data Is Better Together Initiative

Hugging Face and Argilla launched the Data Is Better Together (DIBT) initiative to empower the open-source community to collectively create high-quality, diverse, and inclusive datasets for machine learning.

460

Prezi Case Study: Accelerating ML Roadmap with Hugging Face Expert Support

Prezi is leveraging the Hugging Face Expert Support Program and Inference Endpoints to integrate efficient open-source multimodal models into its Prezi AI presentation generation tool.

461

BigCodeBench: A New Benchmark for Complex Python Code Generation

Hugging Face has released BigCodeBench, a benchmark of 1,140 function-level tasks designed to evaluate LLMs on practical programming and diverse library usage, addressing the simplicity and contamination issues of HumanEval.

462

Hugging Face Accelerate: Harmonizing DeepSpeed and FSDP Precision

Hugging Face Accelerate 0.30.0 introduces automatic upcasting for PyTorch FSDP to align its precision handling with DeepSpeed, enabling seamless switching between the two backends without loss of convergence.

463

Stable Diffusion 3 Medium Integration with Diffusers

Hugging Face has integrated Stable Diffusion 3 Medium (2B parameters) into the Diffusers library, introducing a Multimodal Diffusion Transformer (MMDiT) and rectified flow-matching for improved text-to-image synthesis.

464

Hugging Face TRL RLOO Trainer Release

Hugging Face has introduced the RLOO (REINFORCE Leave One-Out) Trainer in TRL, an online RLHF algorithm that uses 50-70% less vRAM and converges up to 3x faster than PPO while remaining competitive in performance.

465

Hugging Face Embedding Container for Amazon SageMaker Release

Hugging Face has released a general availability (GA) Embedding Container for Amazon SageMaker, powered by Text Embedding Inference (TEI) for high-performance deployment of open embedding models.

466

Hugging Face Transformers Documentation Redesign

Hugging Face is redesigning the Transformers documentation to move from a rigid, incremental structure to a code-first, integrated experience tailored for product developers.

467

Artificial Analysis Text to Image Leaderboard and Arena Launch

Hugging Face has launched the Artificial Analysis Text to Image Leaderboard and Arena, using human preference data from over 45,000 votes to rank open-source and proprietary image generation models.

468

Hugging Face NPC-Playground: Integrating LLM-Powered NPCs in 3D Environments

Hugging Face introduces NPC-Playground, a 3D demo utilizing Cubzh and Gigax to enable realistic, LLM-driven NPC interactions through function calling and Lua scripting.

469

Intel Gaudi Assisted Generation Support

Hugging Face has integrated assisted generation (speculative sampling) into Optimum Habana, enabling up to 2x speedups for large transformer-based models on Intel Gaudi processors.

470

Hugging Face Spaces Secrets Security Update

Hugging Face disclosed a security incident involving unauthorized access to a subset of Spaces secrets, leading to the revocation of affected tokens and a transition to fine-grained access tokens.

471

Benchmarking Text Generation Inference

Hugging Face introduces a benchmarking tool for Text Generation Inference (TGI) to help developers profile throughput and latency trade-offs to optimize LLM deployment costs and performance.

472

Training and Finetuning Embedding Models with Sentence Transformers

Hugging Face provides a comprehensive guide on using the Sentence Transformers library to train and finetune embedding models for tasks like semantic search and RAG, introducing the new SentenceTransformerTrainer.

473

Falcon 2 11B Release Notes

TII has released Falcon 2 11B, a series of smaller, high-performance models including a pretrained LLM and a Vision-Language Model (VLM) trained on over 5,000 billion tokens.

474

CyberSecEval 2: Evaluating Cybersecurity Risks in Large Language Models

Hugging Face and Meta have introduced CyberSecEval 2, a comprehensive framework to evaluate LLM susceptibility to code interpreter abuse, offensive cybersecurity capabilities, and prompt injection attacks.

475

Hugging Face AWS Inferentia2 Integration

Hugging Face has expanded support for AWS Inferentia2, enabling the deployment of over 100,000 models via Amazon SageMaker and introducing new Inf2 instance options for Hugging Face Inference Endpoints.

476

Hugging Face and Microsoft Strategic Collaboration Update May 2024

Hugging Face and Microsoft have expanded their strategic partnership to integrate popular open models into the Azure Model Catalog, enable AMD MI300X GPU support on Azure, and introduce a VS Code integration for Hugging Face Spaces.

477

Hugging Face Integration for AMD Instinct MI300 GPU

Hugging Face has enabled first-class integration for the AMD Instinct MI300 GPU, delivering up to 3x faster inference and 2x faster fine-tuning for Llama 3 70B compared to the MI250.

478

Dell Enterprise Hub: On-Premise AI Deployment and Training

Hugging Face and Dell have launched the Dell Enterprise Hub, a platform that simplifies the on-premise training and deployment of open-source large language models on Dell hardware.

479

Hugging Face Spaces Dev Mode Release

Hugging Face has introduced Spaces Dev Mode, allowing PRO subscribers to connect via VS Code or SSH for real-time code editing without requiring git pushes for every change.

480

Hugging Face KV Cache Quantization

Hugging Face has introduced KV Cache Quantization in the Transformers library to reduce memory usage for long-context text generation with minimal impact on model quality.

481

Introducing the Open Arabic LLM Leaderboard

Hugging Face and the Technology Innovation Institute have launched the Open Arabic LLM Leaderboard (OALL) to provide a specialized, open evaluation platform for Arabic Large Language Models.

482

PaliGemma Vision Language Model Release

Google has released PaliGemma, a family of open vision-language models combining SigLIP-So400m and Gemma-2B to enable tasks like image captioning, visual question answering, and object detection.

483

Hugging Face and LangChain Launch langchain_huggingface Partner Package

Hugging Face and LangChain have released langchain_huggingface, a jointly maintained partner package to streamline the integration of Hugging Face models into the LangChain ecosystem.

484

Transformers Agents 2.0 release notes / what's new

Hugging Face introduces Transformers Agents 2.0, a modular agent framework featuring iterative problem-solving agents and high performance on the GAIA benchmark using Llama-3-70B-Instruct.

485

Building Cost-Efficient Enterprise RAG Applications with Intel Gaudi 2 and Intel Xeon

Hugging Face demonstrates how combining Intel Gaudi 2 accelerators and Intel Xeon CPUs via the OPEA framework allows enterprises to build RAG applications with superior performance-per-dollar compared to H100-based systems.

486

Hugging Face Enterprise Hub now available on AWS Marketplace

Hugging Face has integrated Enterprise Hub with the AWS Marketplace, allowing organizations to upgrade their accounts and manage billing directly through their AWS accounts.

487

Hugging Face Open Leaderboard for Hebrew LLMs

Hugging Face has launched an open leaderboard specifically designed to evaluate and improve Large Language Models (LLMs) in Hebrew, addressing the challenges of low-resource and morphologically complex languages.

488

Artificial Analysis LLM Performance Leaderboard on Hugging Face

Hugging Face has integrated the Artificial Analysis LLM Performance Leaderboard, providing AI engineers with a unified metric system for comparing the quality, price, and speed of over 100 serverless LLM API endpoints.

489

Hugging Face Inference Endpoints ASR and Diarization Pipeline

Hugging Face introduces a custom inference handler for deploying a modular pipeline combining Whisper ASR, Pyannote diarization, and speculative decoding on Inference Endpoints.

490

Improving Prompt Consistency with Structured Generations

Hugging Face and Dottxt research demonstrates that using structured generation to constrain LLM outputs reduces performance variance and improves ranking consistency across different prompt formats and shot orders.

491

StarCoder2-15B-Instruct-v0.1 release notes / what's new

Hugging Face introduces StarCoder2-15B-Instruct-v0.1, the first entirely self-aligned code LLM trained with a fully transparent and permissive pipeline that outperforms CodeLlama-70B-Instruct on HumanEval.

492

Hugging Face Open Chain of Thought Leaderboard

Hugging Face has introduced the Open Chain of Thought Leaderboard to measure the specific accuracy gain provided by chain-of-thought prompting across various LLMs on challenging reasoning tasks.

493

Jack of All Trades (JAT) Multi-Purpose Transformer Agent

Hugging Face introduces Jack of All Trades (JAT), a single transformer-based agent capable of performing diverse sequential decision-making tasks across Atari, BabyAI, Meta-World, and MuJoCo environments.

494

The Open Medical-LLM Leaderboard: Benchmarking Large Language Models in Healthcare

Hugging Face has introduced the Open Medical-LLM Leaderboard, a standardized platform to evaluate and compare the performance of LLMs across diverse medical datasets to improve reliability and patient safety.

495

Meta Llama 3 Release Notes

Meta has released Llama 3, an open-access LLM family featuring 8B and 70B parameter models with improved tokenization and training on 15 trillion tokens.

496

Ryght Case Study: Building a Life Sciences Generative AI Platform with Hugging Face

Ryght has launched Ryght Preview, an enterprise-grade generative AI platform for healthcare and life sciences that leverages Hugging Face's Expert Support, TGI, and TEI to provide secure, flexible, and high-performance AI copilots.

497

Running Privacy-Preserving Inferences on Hugging Face Endpoints

Hugging Face and Zama have enabled the deployment of Fully Homomorphic Encryption (FHE) models via Hugging Face Endpoints, allowing users to perform machine learning inferences on encrypted data without decrypting it.

498

LiveCodeBench Leaderboard: Contamination-Free Evaluation for Code LLMs

Hugging Face has introduced the LiveCodeBench leaderboard, a new benchmark developed by researchers from UC Berkeley, MIT, and Cornell to evaluate LLM code generation and reasoning capabilities while preventing benchmark contamination using time-windowed problem sets.

499

Gradio Reload Mode for Faster AI App Development

Gradio's reload mode enables developers to apply source code changes to AI applications instantly without restarting the server, significantly reducing development latency.

500

Idefics2 8B Vision-Language Model Release – Architecture, Data, and Performance

Hugging Face released Idefics2, an 8B open‑source vision‑language model that outperforms other 8‑B models on VQA and OCR benchmarks and is ready for fine‑tuning via 🤗 Transformers.