Dockerfile should use the correct nvidia channel to install libcu* #144
Triggered via pull request
December 11, 2024 18:33
praateekmahajan
Status
Failure
Total duration
21m 50s
Artifacts
–
gpuci.yml
on: pull_request
build-container
/
main
12s
run-gpu-tests
0s
Annotations
1 error
|
build-container / main
buildx failed with: ERROR: failed to solve: process "/bin/bash -euo pipefail -c conda create -y --name curator -c nvidia/label/cuda-${CUDA_VER} -c conda-forge python=3.10 cuda-cudart libcufft libcublas libcurand libcusparse libcusolver cuda-nvvm && source activate curator && pip install --upgrade pytest pip" did not complete successfully: exit code: 1
|