NVIDIA Triton Inference Server

Achieve high-throughput inference.

Triton executes multiple models from the same or different frameworks concurrently on a single GPU or CPU. In a multi-GPU server, Triton automatically creates an instance of each model on each GPU to increase utilization.

It also optimizes serving for real-time inferencing under strict latency constraints with dynamic batching, supports batch inferencing to maximize GPU and CPU utilization, and includes built-in support for audio and video streaming input. Triton supports model ensemble for use cases that require a pipeline of multiple models to perform end-to-end inference, such as conversational AI.

Models can be updated live in production without restarting Triton or the application. Triton enables multi-GPU, multi-node inference on very large models that cannot fit in a single GPU’s memory.

Scale inference with ease.

Available as a Docker container, Triton integrates with Kubernetes for orchestration, metrics, and autoscaling. Triton also integrates with Kubeflow and KServe for an end-to-end AI workflow and exports Prometheus metrics for monitoring GPU utilization, latency, memory usage, and inference throughput. It supports the standard HTTP/gRPC interface to connect with other applications like load balancers and can easily scale to any number of servers to handle increasing inference loads for any model.

Triton can serve tens or hundreds of models through a model control API. Models can be loaded and unloaded into and out of the inference server based on changes to fit in GPU or CPU memory. Supporting a heterogeneous cluster with both GPUs and CPUs helps standardize inference across platforms and dynamically scales out to any CPU or GPU to handle peak loads.