Open multimodal and language models built from Gemini research, including Gemma, PaliGemma, CodeGemma, RecurrentGemma, and ShieldGemma.
Pioneering bidirectional encoder representations from transformers for natural language understanding.
Text-to-Text Transfer Transformer models for unified natural language processing tasks.
A pretrained time-series foundation model for time-series forecasting.
Open models for medical image and clinical text comprehension to accelerate healthcare applications.
Google DeepMind technology that watermarks and identifies AI-generated content across text, audio, image, and video.
On-device machine learning using Google AI Edge tools, runtimes, and documentation.
Low-code frameworks and customizable solutions for common vision and audio tasks.
Run pretrained or custom models on-device with LiteRT (previously known as TensorFlow Lite).
Convert PyTorch models to LiteRT and author high-performance on-device LLMs.
Visualize model graphs and architectures with Model Explorer documentation guide.
High-performance numerical computing and machine learning library for accelerators.
Flexible neural network library for JAX designed for scale and research experimentation.
Scalable, high-performance LLM training and inference engine written in JAX.
Multi-backend deep learning framework for modular and flexible model authoring.
End-to-end open-source platform for machine learning and production deployment.
Enable PyTorch models and training workloads on Google TPUs and XLA devices.
Open-source machine learning compiler for GPUs, CPUs, and accelerators.
Seamless TPU acceleration for Hugging Face transformers and training stacks.
Kubernetes deployment patterns for multi-host AI/ML inference workloads.
Deep Learning Containers for training and deploying Hugging Face models on Cloud.
A collection of AI examples, best-practices, and prebuilt Kubernetes solutions.
Deploy Hugging Face Hub models on Google Cloud CPU, GPU, or TPU infrastructure.
Hugging Face Deep Learning Containers optimized for Google Cloud training and serving.
Securely manage Hugging Face authentication tokens in Colab environment secrets.
Deploy zero-configuration Hugging Face generative AI microservices on Google Cloud.