How Google Uses Hugging Face

Deploying secure, real-time Edge AI on Raspberry Pi is now simplified using LiteRT and lightweight Gemma open models. LiteRT optimizes CPU and GPU performance, delivering fast token speeds for models like Gemma4, enabling real-time local reasoning for robotics. Developers can quickly convert, quantize, and run these models using the lightweight LiteRT CLI tool. Support for Hailo AI accelerators is also coming very soon.

The 2026-07-28 Model Context Protocol (MCP) specification replaces legacy stateful constraints with a fully stateless core, enabling cloud-native horizontal scaling, serverless deployments, and standard round-robin load balancing. This architectural shift introduces standardized HTTP headers for efficient routing without deep packet inspection, caching controls, and Multi Round-Trip Requests (MRTR) to handle interactive and long-running tasks without blocking connections. Developers can immediately begin migrating their agentic applications to this highly scalable infrastructure using the newly available beta SDKs for Python, TypeScript, Go, and C#.

Tunix is Google’s new JAX-native post-training library designed to eliminate TPU idling bottlenecks when training multi-turn, tool-using LLM reasoning agents. It maximizes hardware throughput by combining highly concurrent, asynchronous rollouts with a decoupled producer-consumer pipeline, ensuring the trainer is constantly fed even while agents wait on network I/O or environment steps. Additionally, Tunix provides plug-and-play abstractions and continuous macro-level profiling, allowing developers to easily integrate custom open-source environments and optimize complex distributed workflows without massive code rewrites.

We're excited to introduce LiteRT.js, the newest member of the LiteRT family! LiteRT.js is our powerful solution for running machine learning models directly in the browser, extending Google's cross-platform edge AI runtime to the web. Built for JavaScript developers, LiteRT.js delivers state-of-the-art ML model inference performance on WebGPU and upcoming WebNN, with a fallback to WebAssembly for CPU. This post provides a quick tour of LiteRT.js and gives web developers everything they need to get started.

DiffusionGemma is an experimental text-generation model built on the Gemma 4 architecture that uses diffusion-based parallel generation instead of token-by-token autoregression, enabling much faster inference, bidirectional context awareness, and real-time self-correction while remaining deployable on consumer GPUs. Its architecture generates and refines 256-token blocks in parallel through iterative denoising, allowing it to handle complex constraint-based tasks such as Sudoku more effectively than traditional language models and demonstrating strong gains from fine-tuning. The model integrates with vLLM and other popular inference frameworks, giving developers access to a new non-autoregressive approach that combines high performance, efficient long-context scaling, and straightforward customization and deployment.

The newly released Gemma 4 12B is a dense, multimodal model designed for high-performance local AI execution on consumer devices. By introducing a novel, encoder-free architecture, it bypasses traditional visual and audio encoders to feed multimodal data directly into the LLM backbone.

The Google Tensor ML SDK is graduating to its Beta phase, allowing developers to build and deploy high-performance machine learning models directly onto the TPU of Google Pixel 10 devices. By integrating with LiteRT, Google's edge deployment framework, the SDK provides a unified workflow for developers to convert, compile, and run PyTorch or TFLite models with robust fallback options. Additionally, a new model garden offers over 100 classic and generative AI models, including Gemma 3, enabling low-latency, private features like speech recognition, computer vision, and text generation.

Google Cloud has introduced a high-performance integration that connects Rapid Storage directly to PyTorch via the fsspec interface to eliminate AI training bottlenecks. By utilizing Google’s Colossus architecture and bidirectional gRPC streaming, the solution offers up to 15 TiB/s aggregate throughput and significant reductions in latency. These improvements allow developers to speed up total training time by 23% with zero code changes required beyond updating the storage bucket type.

LiteRT is a production-ready framework designed to help mobile developers unlock the power of Neural Processing Units (NPUs), overcoming the performance and battery limitations of traditional CPU or GPU processing. By providing a unified API that abstracts away hardware complexities, it allows industry leaders like Google Meet and Epic Games to deploy sophisticated AI models for real-time video, animation, and speech recognition with significantly higher efficiency. The platform further supports developers through benchmarking tools and cross-platform compatibility, enabling seamless AI deployment across mobile devices, AI PCs, and industrial IoT hardware.

MaxText has introduced new support for Supervised Fine-Tuning (SFT) and Reinforcement Learning (RL) on single-host TPU configurations, leveraging JAX and the Tunix library for high-performance model refinement. These features enable developers to easily adapt pre-trained models for specialized tasks and complex reasoning using efficient algorithms like GRPO and GSPO. This update streamlines the post-training workflow, offering a scalable path from single-host setups to larger multi-host configurations.

Agent Development Kit (ADK) now supports a robust ecosystem of third-party tools and integrations. Connect your agents to GitHub, Notion, Hugging Face, and more to build capable, real-world applications.

This tutorial demonstrates how to fine-tune FunctionGemma, a small language model for translating natural language into API calls, using Google's Tunix library on TPUs.

LiteRT has evolved from its TensorFlow Lite foundation into a universal on-device AI inference framework, now offering production-ready GPU acceleration across six platforms and streamlined NPU int...

This article demonstrates how to fine-tune FunctionGemma, a specialized 270M parameter Gemma 3 model designed for function calling in agentic AI systems.

The article discusses how to fine-tune the Gemma 3 270M model for on-device applications, enabling developers to create custom AI models without the need for expensive hardware.

The article provides an in-depth exploration of the EmbeddingGemma architecture, detailing its origins, embedding generation process, and the comprehensive training methodology.

The article discusses the launch of the Google AI Edge Gallery app, which now includes audio capabilities and is available on Google Play.

This article discusses the integration of Google's EmbeddingGemma model with Google Cloud's Dataflow to create a scalable embedding pipeline for AI applications.

EmbeddingGemma is an innovative open embedding model designed for on-device AI applications, featuring 308 million parameters for efficient performance.

The article introduces Gemma 3 270M, a compact AI model designed for hyper-efficient task-specific fine-tuning.

LangExtract is an open-source Python library powered by Gemini, designed to facilitate the extraction of structured information from unstructured text.

Stanford's Marin project introduces the first fully open foundation model developed using JAX, emphasizing transparency in the scientific process behind AI models.

The article introduces T5Gemma, a new collection of encoder-decoder models derived from pretrained decoder-only models.

The article introduces Gemma 3n, a mobile-first architecture designed for on-device AI, highlighting its multimodal capabilities and architectural innovations.

The article discusses how to use KerasHub for loading model weights from SafeTensors into Keras, enabling flexible end-to-end machine learning workflows across different frameworks like JAX, PyTorc...

The article discusses the expansion of Google's AI Edge platform to support on-device small language models (SLMs) with multimodal capabilities, including the introduction of the Gemma 3 and Gemma ...

The article discusses the launch of Gemma 3, a state-of-the-art AI model optimized for consumer GPUs through Quantization-Aware Training (QAT).

TxGemma is a collection of open models designed to enhance the efficiency of therapeutic development by utilizing large language models.

Gemma 3 is the latest version of the Gemma open-model family, boasting enhanced capabilities such as multimodality, longer context windows, and improved reasoning.

The article discusses the Gemma 3 1B model, a lightweight language model designed for mobile and web applications using Google AI Edge.

The article discusses the launch of ShieldGemma 2, a safety content classifier model built on Gemma 3, aimed at detecting harmful content in both synthetic and natural images.

PaliGemma 2 mix is an advanced vision-language model designed for multiple tasks, allowing developers to utilize a single model for various applications such as image captioning, object detection, ...

The article reflects on the first year of the Gemini era, highlighting the advancements made in AI technology and the community's contributions.

PaliGemma 2 is the latest vision-language model from Google, designed to simplify the process of building advanced AI that can interpret visual inputs.

The Web AI Summit 2024, hosted by Google on October 18, 2024, focused on client-side AI for developers, showcasing how machine learning models can operate offline in web browsers.

The article discusses the expansion of the Responsible Generative AI Toolkit, introducing new tools designed for various large language models (LLMs) like Gemma and Gemini.

The article discusses Google's advancements in multilingual AI through the introduction of Gemma 2, a lightweight model designed to empower communities to adapt AI for their languages.

The article discusses the release of Gemma 2, a new suite of open models that sets a new standard for performance and accessibility in conversational AI.

The article provides an overview of the Gemma model family architectures, detailing its lightweight, state-of-the-art open models derived from Gemini research.

The article discusses the advancements in responsible AI through the introduction of Gemma 2, which includes models with 27 billion and 9 billion parameters, emphasizing safety and accessibility.

The article presents three engaging project ideas utilizing Gemma, a family of open models for AI tasks including text generation and code completion.

This article provides a comprehensive guide on using Gemma with Ray on Vertex AI, detailing the steps to set up, fine-tune, and deploy machine learning models.

The article discusses the I/O Connect event held in Berlin, which gathered over 1300 developers from more than 100 countries.

The article discusses the release of the Gemma 2 model with 27 billion parameters, highlighting its capabilities in Keras and integration with JAX for efficient model training.

The article introduces PaliGemma, an open vision-language model, along with Gemma 2, the next generation of the Gemma models, and updates to the Responsible AI Toolkit.

This article discusses how to publish Keras models on Kaggle and Hugging Face, highlighting the ease of sharing fine-tuned models with the community.

The article introduces the expansion of the Gemma family with two new models, CodeGemma and RecurrentGemma, designed specifically for developers and researchers.

The article highlights the achievements and activities of Google Machine Learning communities in the second quarter of 2023, showcasing various training campaigns, community events, and innovative ...