Skip to content

Conversation

@syandroo
Copy link
Contributor

just a proof of concept, we infer the image tags instead of having users manually set an image tag

chipping away at all the required fields back to reasonable defaults

tried to make it as un-jank as possible, and added errors so users can just fall back to image tags

try:
ver=torch.__version__.split("+")
torch_version=ver[0]
cuda_version=ver[1][2:] iflen(ver) >1else"113"
Copy link
ContributorAuthor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

can also default to "latest" if parsing goes awry but i'm hesitant cuz upstream build issues are worse than client side errors

Sign up for freeto join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants

@syandroo@seanshi-scale