model_analyzer and model_navigator
These are complementary tools: Model Analyzer profiles and benchmarks model performance characteristics, while Model Navigator optimizes and converts models into deployment-ready formats, typically used sequentially in a model preparation pipeline.
About model_analyzer
triton-inference-server/model_analyzer
Triton Model Analyzer is a CLI tool to help with better understanding of the compute and memory requirements of the Triton Inference Server models.
This tool helps machine learning engineers and MLOps professionals optimize how their AI models run on NVIDIA's Triton Inference Server. It takes your model files and hardware specifications to generate configurations that balance performance, latency, and resource usage. The output includes detailed reports showing the trade-offs of different settings, helping you choose the best setup for your production environment.
About model_navigator
triton-inference-server/model_navigator
Triton Model Navigator is an inference toolkit designed for optimizing and deploying Deep Learning models with a focus on NVIDIA GPUs.
This tool helps machine learning engineers and MLOps specialists streamline the deployment of deep learning models and pipelines, especially for inference on NVIDIA GPUs. It takes models built in PyTorch, TensorFlow, or ONNX, optimizes them, and outputs highly performant models ready for serving on Triton Inference Server or PyTriton.
Scores updated daily from GitHub, PyPI, and npm data. How scores work