Skip to content

Conversation

@wkobielx
Copy link

What does this PR do?

Accept custom installed from branch/sha versions - this should allow to use at least 2.19.0 dev0 builds

@HuggingFaceDocBuilderDev

The docs for this PR live here. All of your documentation changes will be reflected on that endpoint. The docs are available until 30 days after the last update.

Copy link
Collaborator

@nikita-savelyevv nikita-savelyevv left a comment

Choose a reason for hiding this comment

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

Thanks for the quick fix!

Copy link
Collaborator

@nikita-savelyevv nikita-savelyevv left a comment

Choose a reason for hiding this comment

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

On the second look, using is_nncf_version(">=", "2.19.0.dev0") would allow any custom commit between NNCF 2.18 and 2.19 which is not ideal. This won't lead to any failures, but still if someone has such "in-between" version installed, they should be prompted to upgrade to the latest NNCF version.

I wonder if instead validation can use custom commit of optimum-intel and not the latest version?

import nncf

if is_nncf_version("<", "2.19"):
if not is_nncf_version(">=", "2.19.0.dev0"):
Copy link
Collaborator

@rkazants rkazants Dec 11, 2025

Choose a reason for hiding this comment

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

I think we can just cut off the current version value to have in format xx.yy.zz and after that we can do comparation like is_nncf_version("<", "2.19"). It should fix the problem

Copy link
Collaborator

Choose a reason for hiding this comment

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

Thanks for the suggestion! Unfortunately that won't work because in NNCF we increase dev version to the next one right after the release, i.e. the version was updated to 2.19.0.dev0 back in August openvinotoolkit/nncf@5ba2f29. That's why we'll have the same problem as described in #1550 (review).

@nikita-savelyevv
Copy link
Collaborator

@wkobielx Feel free to close this PR, if another approach was found. I will address the problem separately after some discussion with the team.

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.

4 participants