fix: Remove tests and examples for deprecated models#1012
fix: Remove tests and examples for deprecated models#1012jakelorocco merged 15 commits intogenerative-computing:mainfrom
Conversation
Signed-off-by: Fred Reiss <frreiss@us.ibm.com>
Signed-off-by: Fred Reiss <frreiss@us.ibm.com>
Signed-off-by: Fred Reiss <frreiss@us.ibm.com>
Signed-off-by: Fred Reiss <frreiss@us.ibm.com>
Signed-off-by: Fred Reiss <frreiss@us.ibm.com>
Signed-off-by: Fred Reiss <frreiss@us.ibm.com>
Signed-off-by: Fred Reiss <frreiss@us.ibm.com>
|
The PR description has been updated. Please fill out the template for your PR to be reviewed. |
ajbozarth
left a comment
There was a problem hiding this comment.
A couple of items to fix or clarify before merge.
There was a problem hiding this comment.
Do we need to add tests back for a newer version of the uncertainty model later?
There was a problem hiding this comment.
That would probably be a good idea.
| try: | ||
| assert result == expected | ||
| except AssertionError as ae: | ||
| pytest.xfail(f"Known differences across platforms. Diff was: {ae}") |
There was a problem hiding this comment.
I'm fine with this to fix the version pinning / non-deterministic issues now. But some of our tests do sometimes fail because of different versions of the adapter. I'll take a look at #976 later to see what that solution is as well.
|
Upon further discussion, it appears that the actual requirement is to keep context relevance support around, but to disable it selectively for Granite 4.1 base models. This will require a different set of changes. |
Signed-off-by: Fred Reiss <frreiss@us.ibm.com>
ajbozarth
left a comment
There was a problem hiding this comment.
posted a few nits, but otherwise everything is looking good
Co-authored-by: Alex Bozarth <ajbozart@us.ibm.com>
Co-authored-by: Alex Bozarth <ajbozart@us.ibm.com>
Signed-off-by: Fred Reiss <frreiss@us.ibm.com>
Misc PR
Type of PR
Description
Testing
Attribution