Trimble Business Center

 View Only

 CUDA error with RTX 5000 series when training custom DL model

Hoang Bach Cao's profile image
Hoang Bach Cao posted 12-22-2025 21:29

Hi everyone,

I'm following the guidelines to train the demo stockpile dataset. And I ran into CUDA error with the custom DL training. I want to ask if this is CUDA/GPU compatibility or is there any workaround? I already set TBC power management in Nvidia control panel to prefer maximum performance.

Error: 

RuntimeError: CUDA error: no kernel image is available for execution on the device

CUDA kernel errors might be asynchronously reported at some other API call, so the stacktrace below might be incorrect.

For debugging consider passing CUDA_LAUNCH_BLOCKING=1

Compile with `TORCH_USE_CUDA_DSA` to enable device-side assertions.

My system:
Dual RTX 5070Ti 16GB
Windows 10 Pro

Cuda ccompiler:
nvcc: NVIDIA (R) Cuda compiler driver
Release: CUDA 13.0, V13.0.88

------------------------------
Hoang Bach Cao
------------------------------