Gemma is a set of lightweight, generative artificial intelligence (AI)open models. Gemma models are available to run in yourapplications and on your hardware, mobile devices, or hosted services. You canalso customize these models using tuning techniques so that they excel atperforming tasks that matter to you and your users. Gemma models arebased onGemini models and are intendedfor the AI development community to extend and take further.

Fine-tuning can help improve a model's performance in specific tasks. Becausemodels in the Gemma model family are open weight, you can tune any ofthem using the AI framework of your choice and the Vertex AI SDK.You can open a notebook example to fine-tune the Gemma model usinga link available on the Gemma model card in Model Garden.

The following Gemma models are available to use with Vertex AI.To learn more about and test the Gemma models, see theirModel Garden model cards.

The following are some options for where you can use Gemma:

Use Gemma with Vertex AI

Vertex AI offers a managed platform for rapidly building and scalingmachine learning projects without needing in-house MLOps expertise. You can useVertex AI as the downstream application that serves theGemma models. For example, you might port weights from the Kerasimplementation of Gemma. Next, you can use Vertex AI toserve that version of Gemma to get predictions. We recommend usingVertex AI if you want end-to-end MLOps capabilities, value-added MLfeatures, and a serverless experience for streamlined development.

To get started with Gemma, see the following notebooks:

• Serve Gemma 3n in Vertex AI

• Serve Gemma 3 in Vertex AI

• Serve Gemma 2 in Vertex AI

• Serve Gemma in Vertex AI

• Fine-tune Gemma 3 using PEFT and then deploy to Vertex AI from Vertex

• Fine-tune Gemma 2 using PEFT and then deploy to Vertex AI from Vertex

• Fine-tune Gemma using PEFT and then deploy to Vertex AI from Vertex

• Fine-tune Gemma using PEFT and then deploy to Vertex AI from Huggingface

• Fine-tune Gemma using KerasNLP and then deploy to Vertex AI

• Fine-tune Gemma with Ray on Vertex AI and then deploy to Vertex AI

• Run local inference with ShieldGemma 2 with Hugging Face transformers

• Serve MedGemma in Vertex AI

• Serve MedSigLIP in Vertex AI

• Run local inference with T5Gemma with Hugging Face transformers

Use Gemma in other Google Cloud products

You can use Gemma with other Google Cloud products, such asGoogle Kubernetes Engine and Dataflow.

Use Gemma with GKE

Google Kubernetes Engine (GKE) is the Google Cloud solutionfor managed Kubernetes that provides scalability, security, resilience, and costeffectiveness. We recommend this option if you have existing Kubernetesinvestments, your organization has in-house MLOps expertise, or if you needgranular control over complex AI/ML workloads with unique security, datapipeline, and resource management requirements. To learn more, see the followingtutorials in the GKE documentation:

• Serve Gemma with vLLM
• Serve Gemma with TGI
• Serve Gemma with Triton and TensorRT-LLM
• Serve Gemma with JetStream

Use Gemma with Dataflow

You can use Gemma models with Dataflow forsentiment analysis.Use Dataflow to run inference pipelines that use theGemma models. To learn more, seeRun inference pipelines with Gemma open models.

Use Gemma with Colab

You can use Gemma with Colaboratory to create your Gemmasolution. In Colab, you can use Gemma with frameworkoptions such as PyTorch and JAX. To learn more, see:

• Get started with Gemma using Keras.
• Get started with Gemma using PyTorch.
• Basic tuning with Gemma using Keras.
• Distributed tuning with Gemma using Keras.

Gemma model sizes and capabilities

Gemma models are available in several sizes so you can buildgenerative AI solutions based on your available computing resources, thecapabilities you need, and where you want to run them. Each model is availablein a tuned and an untuned version:

• Pretrained - This version of the model wasn't trained on any specific tasksor instructions beyond the Gemma core data training set. We don'trecommend using this model without performing some tuning.

• Instruction-tuned - This version of the model was trained with human languageinteractions so that it can participate in a conversation, similar to a basicchat bot.

• Mix fine-tuned - This version of the model is fine-tuned on a mixture ofacademic datasets and accepts natural language prompts.

Lower parameter sizes means lower resource requirements and more deploymentflexibility.

Gemma has been tested using Google's purpose built v5e TPUhardware and NVIDIA's L4(G2 Standard), A100(A2 Standard),H100(A3 High) GPU hardware.

What's next

• SeeGemma documentation.