search

NVIDIA / TensorRT

NVIDIA® TensorRT™ is an SDK for high-performance deep learning inference on NVIDIA GPUs. This repository contains the open source components of TensorRT.
https://developer.nvidia.com/tensorrt
Apache License 2.0
10.15k stars 2.08k forks source link

No me permite instalar TENSORrt #3946

Open MAFLIXD opened 2 weeks ago

MAFLIXD commented 2 weeks ago

Description

Environment

TensorRT Version:

NVIDIA GPU:

NVIDIA Driver Version:

CUDA Version:

CUDNN Version:

Operating System:

Python Version (if applicable):

Tensorflow Version (if applicable):

PyTorch Version (if applicable):

Baremetal or Container (if so, version):

Relevant Files

Model link:

Steps To Reproduce

Commands or scripts:

Have you tried the latest release?:

Can this model run on other frameworks? For example run ONNX model with ONNXRuntime (polygraphy run <model.onnx> --onnxrt):

Essa79 commented 2 weeks ago

I'd be glad to assist you with installing TensorRT, even without the specific error message #3946. Here's a comprehensive guide that incorporates insights from previous responses and addresses potential issues:

Prerequisites:

NVIDIA GPU: TensorRT requires an NVIDIA GPU with compute capability 3.0 or higher. You can check your GPU model using nvidia-smi in the terminal (Linux/macOS) or NVIDIA Control Panel (Windows). CUDA Toolkit and NVIDIA Driver: Install the latest CUDA Toolkit and NVIDIA Driver version that's compatible with your GPU from the NVIDIA website (https://developer.nvidia.com/cuda-toolkit). Installation Steps:

Choose Installation Method:

Package Manager (Recommended): For Linux distributions like Ubuntu, Debian, or CentOS, you can often use the package manager: Bash sudo apt-get install libnvinfer-dev libnvinfer-plugins-dev # For Ubuntu/Debian sudo yum install cuda-toolkit nvidia-tensorrt # For CentOS/RHEL From Source (Advanced): If packages aren't available or you need a specific version, download the source code from the NVIDIA NGC catalog (https://developer.nvidia.com/tensorrt) and follow the build instructions. Verify Installation:

Open a Python interpreter and run import tensorrt. If there's no error, TensorRT is installed correctly. Troubleshooting Common Issues:

Incorrect CUDA or Driver Version: Ensure they're compatible with TensorRT. Refer to the TensorRT documentation for supported versions. Missing Dependencies: Install required dependencies like python3-dev or build-essential (Linux) using the package manager. Permissions: Make sure you have the necessary permissions to install packages. Use sudo if required. Cache Issues: Sometimes deleting package cache files can help: Bash sudo apt-get clean # Ubuntu/Debian sudo yum clean all # CentOS/RHEL Use code with caution. content_copy Additional Tips:

If you encounter specific errors, search online for solutions related to that error code. Check TensorRT documentation and forums for known issues and troubleshooting steps (https://docs.nvidia.com/deeplearning/tensorrt/). If you're still facing issues after trying these steps:

Provide More Details: Share more information about your environment (OS, GPU model, CUDA version, etc.) and the exact error message you're encountering. This will help in providing more tailored assistance. Consider Using Docker: Docker can simplify installation and environment management by creating a container with pre-configured dependencies.

lix19937 commented 2 weeks ago

trt install doc see https://docs.nvidia.com/deeplearning/tensorrt/install-guide/index.html#installing
You can choose between the following installation options when installing TensorRT; Debian or RPM packages, a Python wheel file, a tar file, or a zip file.