Closed skeller88 closed 2 years ago
According to our guidelines, the community-content folder has no support provided. The original author is no longer available to support this sample.
@andrewferlitsch do you have an answer to this question about Vertex AI pipelines? https://stackoverflow.com/questions/73642322/running-custom-docker-container-with-gpu-using-vertex-ai-pipelines-how-to-inst
Expected Behavior
In the pytorch_image_classification_single_gpu_with_vertex_sdk_and_torchserve example, I would expect there to be a driver installation step for the pipeline to use the GPU.
Actual Behavior
Does the pytorch image already contain the nvidia drivers, or does it install the nvidia driver onto the host somehow? I asked a similar question here.