tensorrt-cu12-libs

What it is and what it does

tensorrt-cu12-libs is a binary distribution of NVIDIA TensorRT's core libraries compiled for CUDA 12 environments. It provides the runtime components needed to load, optimize, and execute deep learning models on NVIDIA GPUs with low latency and high throughput. The package is part of TensorRT 11.X, which introduced breaking changes from the 10.X series, including removal of weakly-typed networks, implicit quantization, and IPluginV2 APIs in favor of their strongly-typed and explicit counterparts.

This is a library package intended for deployment environments where model inference is the primary goal. It does not include model training capabilities or the full TensorRT build toolchain. Installation requires pre-existing CUDA 12 support on the system and is most commonly used in conjunction with model conversion tools (ONNX, Torch-TensorRT, or the Network Definition API) to prepare models for inference. The high install friction reflects the size and specificity of the binary payload.

Use it for:

• Deploy pre-optimized deep learning models for inference on NVIDIA GPUs in production environments
• Accelerate inference latency for computer vision, NLP, and audio models on CUDA 12 systems
• Build inference servers or microservices that require low-latency model execution
• Optimize model throughput for batch inference workloads on NVIDIA data center GPUs
• Integrate GPU-accelerated inference into Python applications targeting CUDA 12 infrastructure

Worth the install?

AI-flagged interpretation of the facts on this page — verify before relying

Provides NVIDIA TensorRT libraries for CUDA 12 environments, enabling optimized deep learning model inference on NVIDIA GPUs.

Yes, if you need GPU-accelerated inference on CUDA 12 systems and accept the proprietary license terms. The package is actively maintained, has no known vulnerabilities, and is widely used (top 15000 PyPI packages). Install friction is high due to binary size, but that is inherent to the use case. Verify CUDA 12 compatibility and licensing requirements before production deployment.

Install

tensorrt-cu12-libs on PyPI

pip install tensorrt-cu12-libs

uv add tensorrt-cu12-libs

poetry add tensorrt-cu12-libs

Installing tensorrt-cu12-libs

Before you install

High install friction due to large binary dependencies (tensorrt_cu12_libs-11.2.1.2.tar.gz). Active maintenance with recent releases; repository shows strong community engagement (13250 stars) and current development activity (last commit 2026-08-04).

License in practice

Licensed as Proprietary with unclear treatment. Users should verify licensing terms with NVIDIA before deploying in production or commercial contexts, as proprietary restrictions may apply.

Quickstart

pip install tensorrt-cu12-libs

import tensorrt as trt

logger = trt.Logger(trt.Logger.WARNING)
builder = trt.Builder(logger)

Requires CUDA 12.x runtime and compatible NVIDIA GPU drivers; Python >= 3.10 and <= 3.14.x; TensorRT GA build v11.2.1.2 may need to be downloaded separately depending on installation method.

Verify before relying

• Whether this package includes the full TensorRT runtime or only CUDA 12 library bindings
• Exact Python version support (description mentions 3.10–3.14.x but package metadata does not specify)
• Whether CUDA 12.9 or 13.3 is required or if other 12.x versions are supported
• Licensing restrictions for commercial or proprietary use beyond the 'unclear' classification

Package facts

Evidence: tensorrt_cu12_libs-11.2.1.2.tar.gz

Keywords: nvidia, tensorrt, deeplearning, inference

Tags

More Artificial Intelligence packages

Further reading