Skip to content

Conversation

ckadner
Copy link
Collaborator

@ckadner ckadner commented Oct 2, 2025

Description

Draft PR #497 has all tests pass because it uses both tiny granite models

  • the 4-layer model from the cache
  • the 3-layer model downloaded with revision

This PR checks the test results for bypassing the cache (not using 4-layer model)

Related Issues

#497

Signed-off-by: Christian Kadner <[email protected]>
@vllm-project vllm-project deleted a comment from github-actions bot Oct 2, 2025
- name: "Restore HF models cache"
id: cache_restore
if: steps.changed-src-files.outputs.any_changed == 'true'
if: ! contains(env.model_key, 'micro')
Copy link
Collaborator

Choose a reason for hiding this comment

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

Okay I think I see what you're trying to do here

ckadner and others added 3 commits October 2, 2025 15:17
Signed-off-by: Prashant Gupta <[email protected]>
Signed-off-by: Prashant Gupta <[email protected]>
@prashantgupta24
Copy link
Collaborator

Going to push commits to this PR to test the model.revision bug fixes!

Signed-off-by: Prashant Gupta <[email protected]>
Signed-off-by: Prashant Gupta <[email protected]>
@ckadner
Copy link
Collaborator Author

ckadner commented Oct 3, 2025

Going to push commits to this PR to test the model.revision bug fixes!

Cool! All tests green 💯

@prashantgupta24
Copy link
Collaborator

So my plan at this time is to open a new PR with the latest commits from this PR which fix the model revision used in the tests. Once we get that in then we can discuss about whether we even want to revert to the 3 layer model or not

@ckadner ckadner closed this Oct 15, 2025
Sign up for free to 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