Error while running the command: "tao detectnet_v2 train"

• Hardware: Rtx 3080
• Network Type : Detectnet_v2
• Tao Toolkit Version: 4.0.0
• Attaching the spec file
detectnet_v2_train_resnet18_kitti.txt (6.3 KB)
• To reproduce the error please use the following command:

tao detectnet_v2 train -e /workspace/tao-experiments/detectnet_v2/specs/detectnet_v2_train_resnet18_kitti.txt -r /workspace/tao-experiments/detectnet_v2/experiment_dir_unpruned -k tlt_encode -n resnet18_detector --gpus 1

tensorflow.python.framework.errors_impl.InternalError: CUDA runtime implicit initialization on GPU:0 failed. Status: the provided PTX was compiled with an unsupported toolchain.
Complete Error Log:
nvidia_tao_log.txt (297.3 KB)

Please update nvidia driver.

Uninstall current driver:
sudo apt purge nvidia-driver-510
sudo apt autoremove
sudo apt autoclean

Install new driver.
sudo apt install nvidia-driver-520

1 Like

Thanks, it worked!

This topic was automatically closed 14 days after the last reply. New replies are no longer allowed.