Conversation
| ci_arm: tlcpackstaging/ci_arm:20230504-142417-4d37a0a0 | ||
| ci_cortexm: tlcpackstaging/ci_cortexm:20230504-142417-4d37a0a0 | ||
| ci_cpu: tlcpackstaging/ci_cpu:20230504-142417-4d37a0a0 | ||
| ci_cpu: tlcpackstaging/ci_cpu:20230513-200357-e54bbc73 |
There was a problem hiding this comment.
I noticed that all images are pointing to tlcpackstaging account rather than tlcpack since #14651. Should we fix them?
There was a problem hiding this comment.
looks the images were uploaded into https://hub.docker.com/u/tlcpack already, we can replace tlcpackstaging with tlcpack for the images with tag 20230504-142417-4d37a0a0. I can update it in follow-up pr after the cpu image with tag 20230513-200357-e54bbc73 is uploaded
There was a problem hiding this comment.
According to https://tvm.apache.org/docs/contribute/ci.html#updating-a-docker-image-tag, the current protocol is to reference tlcpack in a PR that updates an image, and CI automatically fetches the corresponding image from tlcpackstaging. The images are automatically uploaded to tlcpack after successful merge.
I updated the account to tlcpack.
|
I don't get why we are still getting the |
|
Previously what I did to debug the tests related to CI changes is to reproduce it locally first, the ec2 instance type for cpu container is r5.large.
Then try to run the failed test, it should be running into same issue as ci reported |
* [CI] Update CPU image to install PyTorch * use link-static to prevent symbol conflict problem * tlcpackstaging -> tlcpack * disable tvmc pth tests for now * fixed skip marker
Follow up to #14842 to actually update the image