Overview

The Dynamo VLLM runtime container is a pre-built, Docker-based environment designed to run NVIDIA Dynamo with the vLLM backend for high-performance, distributed large language model (LLM) inference. It packages all necessary dependencies, runtime components, and optimizations to streamline deployment and ensure consistency across development and production environments.

Key Components

• vLLM Backend: Provides fast, efficient LLM inference, leveraging vLLM’s optimized attention and KV cache management.

• Dynamo Core Services: Includes the HTTP API server, request router, and worker processes for prefill and decode phases.

• Supporting Services: Integrates with etcd and NATS for distributed coordination and messaging.

• OpenAI-Compatible Frontend: Exposes an HTTP API compatible with OpenAI’s endpoints for easy integration.

For more information about Dynamo features, please refer to the Github repository

Getting Started

• Select the Tags tab and locate the container image release that you want to run.

• In the Pull Tag column, click the icon to copy the docker pull command.

• Open a command prompt and paste the pull command. The pulling of the container image begins. Ensure the pull completes successfully before proceeding to the next step.

• Start required services (etcd and NATS) using Docker Compose:

  docker compose -f deploy/docker-compose.yml up -d

• Run the container image and start dynamo via:

  dynamo run out=vllm deepseek-ai/DeepSeek-R1-Distill-Llama-8B

  For more examples, please refer to the examples directory in the repository.

Support Matrix

Please refer to the following support matrix to learn more about the current hardware & architecture support. Dynamo currently only provides pre-built x86_64 containers.

License

NVIDIA Dynamo is released under an open-source license, Apache-2.0, making it freely available for development, research, and deployment.

Technical Support

GitHub Issues: Dynamo GitHub Issues