Hello,
I am experiencing an issue when running PostgresML with GPU support on my system. Here are the details:

• Host GPU: NVIDIA GeForce RTX 5090
• Driver Version: 580.105.08
• CUDA Version (nvidia-smi): 13.0
• Installed CUDA version (nvcc): 12.6
• PostgresML Docker Image:ghcr.io/postgresml/postgresml:2.10.0
• PostgresML Container Name:pgml-gpu

When I try to run a GPU text-generation task using thepgml.transform function:

SELECTpgml.transform(    task=>'{      "task": "text-generation",      "model": "bigscience/bloom-560m",      "token": "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",      "device": 0    }'::jsonb,    inputs=> ARRAY ['How many methods have pgml ?']);

I receive the following error:

RuntimeError: CUDA error: no kernel image is available for execution on the deviceCUDA 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=1Compile with TORCH_USE_CUDA_DSA to enable device-side assertions.

It seems that there is a mismatch between the installed CUDA toolkit and the driver or the PyTorch/CUDA binary inside the container, causing GPU tasks to fail. CPU tasks work correctly.

Could you please advise how to resolve this issue so that PostgresML can utilize the GPU properly?

Thank you for your assistance.