the latest release of jetpak6.1 is not supporting pytorch, jetson-inference developed by Mr.Dusty and also there is no suitable version of Torch vision to run yolov8.
help immediately please.
Thanks,
– Hems.
Hi,
Have you solved the problems? I also need to the corresponding torch version.
Hey guys, sorry for the delay with the support for JetPack 6.1 - the wheels for PyTorch, torchvision and a bunch of the ML/LLM stuff are up awhile here for CUDA 12.6 / L4T R36.4 along with some of the containers. Also built and pushed ROS Humble desktop image awhile.
For jetson-inference, I’m still backed up on finishing the migrations and testing of that to TRT10, but for the time being was able to run that in a container with previous version of TRT.
Thanks for your patience as we work through all the parts, at this point most things you can build them from jetson-containers if I haven’t uploaded yet. If you encounter errors and encounter other things that need fixing, let me know 👍
@Dusty-nv , Thank you very much. we will wait for jetson-inference . Please let me know where is 'ROS Humble desktop image ’ as you mentioned above ‘Also built and pushed ROS Humble desktop image awhile’.
thank you. , hems.
Hi, The problem has yet to be solved. All hope is on Mr Dustin Franklin. I’ve been waiting for two weeks.
Apologies hemsingh, here was the ROS Humble container for JP 6.1: dustynv/ros:humble-desktop-l4t-r36.4.0
This topic was automatically closed 14 days after the last reply. New replies are no longer allowed.