Problem to run YOLOv8 with GPU in a conda env

[mestrelan]

I confess that I am new to Yolo and GPUs, but I would appreciate it if someone could help me.

I was following the conda quickstart guide but faced a problem with the step:

conda install -c pytorch -c nvidia -c conda-forge pytorch torchvision pytorch-cuda=11.8 ultralytics

Conda could not finish the installation both in Linux and Windows. I will provide more information on what went wrong and the terminal output next week, but I would like to ask first if I need to install cuda toolkit and cudnn before or anaconda install them for me in the env. Maybe someone have a detailed guide of how I could run Yolov8 with GPU (for beginners :) )

I managed to install ultralytics with: conda install -c conda-forge ultralytics but the code did not run with the GPU.

Thank you for your attention. :)

==== Edit ====
I found now the troubleshooting steps and I will soon try it. :)

The last step says that the code must specify GPU usage and this might be in the settings or configuration. How could I specify GPU usage in the code?

[Burhan-Q]

@mestrelan please make sure to open issues for the ultralytics package and YOLO on our primary repository as this one is only for publishing our package to conda-forge and is not managed by Ultralytics directly.