Package Details: python-tensorrt 10.8.0.43-2

Git Clone URL: https://aur.archlinux.org/tensorrt.git (read-only, click to copy)
Package Base: tensorrt
Description: A platform for high-performance deep learning inference on NVIDIA hardware (python bindings and tools)
Upstream URL: https://developer.nvidia.com/tensorrt/
Keywords: ai artificial intelligence nvidia
Licenses: Apache-2.0, LicenseRef-custom, LicenseRef-NVIDIA-SOFTWARE-DEVELOPMENT-KITS
Provides: python-onnx-graphsurgeon, python-polygraphy, python-tensorflow-quantization
Submitter: dbermond
Maintainer: dbermond
Last Packager: dbermond
Votes: 20
Popularity: 0.111089
First Submitted: 2018-07-29 16:17 (UTC)
Last Updated: 2025-02-09 18:55 (UTC)

Dependencies (19)

Sources (12)

Latest Comments

« First ‹ Previous 1 2 3 4 5 6 7 8 9 10 11 Next › Last »

gilvbp commented on 2023-10-22 19:19 (UTC)

Has anyone been successful in constructing this version (9.1.0.4-1)?

dbermond commented on 2023-09-02 13:28 (UTC)

@damian101 package currently fails to build with cuda 12.2. If you want to build it currently, you need to downgrade cuda to 12.1.

damian101 commented on 2023-08-27 22:28 (UTC) (edited on 2023-08-27 22:28 (UTC) by damian101)

I always get a bunch of errors when compiling, until it finally fails:

https://pastebin.com/fsVVDBZP

I have no clue how to fix.

abdallahalswaiti commented on 2023-08-26 17:29 (UTC)

hi every body i successful install this package after using cuda 12.1 i use this repo https://gitlab.archlinux.org/archlinux/packaging/packages/cuda/-/tags/12.1.1-4 -makepkg -si then download https://developer.nvidia.com/downloads/compute/machine-learning/tensorrt/secure/8.6.0/tars/TensorRT-8.6.0.12.Linux.x86_64-gnu.cuda-12.0.tar.gz put it in ./tensorrt/ then makepkg -si wish this helpfull

dbermond commented on 2023-07-12 16:35 (UTC)

@chengscott This is not needed, as cuda already uses the correct gcc version. The package is currently failing to build, not due to gcc, but most probably due to cuda 12.2. At the time of writing, there is no upstream download supporting cuda 12.2, so the solution is to downgrade cuda to 12.1 for the time being.

chengscott commented on 2023-07-12 15:18 (UTC) (edited on 2023-07-12 15:19 (UTC) by chengscott)

Need to set CXX=/opt/cuda/bin/g++ when building.

as of now, CUDA12 does not support g++13 (system g++)

dbermond commented on 2023-07-07 21:56 (UTC)

@mindbound Nice, good to know.

mindbound commented on 2023-07-07 13:21 (UTC)

Thanks! Solved, I had to rebuild python-onnx.

dbermond commented on 2023-07-06 00:17 (UTC)

@mindbound You can start by searching for which packages need protobuf and/or python-protobuf as a dependency. For example, with 'pacman -Qi protobuf', and looking at the 'Required By' and 'Optional For' lines. I would bet that it's python-onnx, since it depends on protobuf. Every time that protobuf gets a soname bump you should rebuild python-onnx (and any other AUR packages that depend on it), and it's user responsibility to keep track of this.

« First ‹ Previous 1 2 3 4 5 6 7 8 9 10 11 Next › Last »