Back to library
Build the GameFrameworkProprietaryFree
NVIDIA TensorRT
TensorRT is NVIDIA's proprietary inference SDK with graph optimization, precision calibration, CUDA kernel selection, and C++ or Python runtimes.
Visit official siteContent updated Aug 7, 2026Automated check reached the official site · Checked Aug 24, 2026
Why use it
Generative AI, speech, vision, and animation models can reduce latency and memory on NVIDIA PCs or servers.
Where it fits
Find engines, frameworks, code libraries, version control, and production automation.
aioptimization
What to check
It targets NVIDIA hardware under SDK terms; test model conversion, accuracy loss, and each hardware generation.