I developed TAO TOOLKIT 5.1 with RTX 3080. I would like to know if it can be used with RTX 4080 or higher GPUs.
I would like to know what are the recommended versions of CUDA and CUDNN for using TAO TOOLKIT 5.1 with the 4080 series.
I also want to configure the triton server, which version is the most optimal.
Your answers will be appreciated.
You can refer to the versions in the tao-deploy docker. Or you can directly run the engine inside the docker. nvcr.io/nvidia/tao/tao-toolkit:5.1.0-deploy.
For triton server, you can refer to https://github.com/NVIDIA-AI-IOT/tao-toolkit-triton-apps/blob/main/docker/Dockerfile.
Thanks for the answer.
I would appreciate it if you could tell me if the 4080 series I asked about will have no problem learning TAO 5.1.0 and what version is recommended.
There is no update from you for a period, assuming this is not an issue anymore. Hence we are closing this topic. If need further support, please open a new one. Thanks
Yes, there is no problem. Searing in this TAO forum, you can find several user are running TAO with 4080 series.