Daniel J Walsh
Daniel J Walsh
Not stale
Does `ramalama run --oci-runtime krun --image quay.io/ramalama/ramalama smollm:135m` Work? We do have a hard coded case of --runtime vllm pulling a particular image, so this is not unprecedented. It would...
So if we change default to Vulkan, then the case we would need to worry about is cuda?
@flouthoc WDYT?
You could also check previous cuda image to see if image update broke: ramalama --image quay.io/ramalama/cuda:0.9.0 run cogito
This usually means that you have a mismatch between the libraries on the host system and those within the container? If you get the podman command and remove the "-t"...
What OS and Version are you on? please submit ramalama info What is the podman command shown when you run with --debug?
Does the following command show what crun is complaining about? `podman --label ai.ramalama.runtime=llama.cpp --label ai.ramalama.command=run --device /dev/dri --device nvidia.com/gpu=all -e CUDA_VISIBLE_DEVICES=0 --network none --runtime /usr/bin/nvidia-container-runtime --security-opt=label=disable --cap-drop=all --security-opt=no-new-privileges --pull newer...
I believe the issue is a mismatch between the cuda libraries on the host and those in the container. This is an issue I have had in the past getting...
https://github.com/containers/ramalama/pull/1517#issuecomment-2969410318https://github.com/containers/ramalama/pull/1517#issuecomment-2969410318 We use crun, for podman OCI Runtime. You did drop the ` -t ` option correct?