TensorRT
Build, optimize, and deploy fast GPU inference engines with TensorRT
All Skills in This Repository (5)
Pure Emerald Level Indicators
NVIDIAtrt-cpp-runtime-quickstart
Load and run TensorRT engines with the modern C++ Runtime API.
NVIDIAtrt-perf-analysis
Validate and analyze TensorRT performance data from paired JSON files.
NVIDIAtrt-onnx-quickstart
Convert ONNX models to TensorRT engines and verify numerical parity.
NVIDIAtrt-strong-typing-migration
Automate TensorRT build migration from weak to strong typing for Python networks.
NVIDIAtrt-torch-quickstart
Convert PyTorch models to TensorRT engines using torch_tensorrt.
Frequently Asked Questions
FAQPage SchemaHow to install TensorRT skills?▼
Run `npx skills add NVIDIA/TensorRT --all -g -y` in your terminal to install all TensorRT agent skills globally.
How to convert an ONNX model to TensorRT?▼
The trt-onnx-quickstart skill walks your AI agent through exporting, sanitizing, and building a verified FP16 engine using trtexec and Polygraphy parity checks.
How to run a TensorRT engine from C++?▼
The trt-cpp-runtime-quickstart skill generates modern load-and-run code using enqueueV3 and setTensorAddress, avoiding deprecated TensorRT 8.x APIs.
How to fix TensorRT 11 migration errors?▼
The trt-strong-typing-migration skill automatically rewrites weakly-typed Python, C++, and trtexec build code to the strongly-typed format required by TensorRT 11.
Can these skills analyze TensorRT performance?▼
Yes. The trt-perf-analysis skill validates layer and profile JSON files, finds hot layers and fusion opportunities, and generates a browser-based performance report.
Related Repositories in Software Engineering
View All in Software Engineering→
openclawopenclaw
Run a personal AI assistant across your devices and chat apps
obrasuperpowers
Gives coding agents a disciplined workflow from idea to merged code
facebookreact
AI agent skills for building, testing, and porting React core