Repository navigation
ci: decision sweep (sweep_decision.py) - #817
Merged
Merged
Conversation
tawei-amd
force-pushed
the
ci/sweep-decision
branch
3 times, most recently
from
October 19, 2026 14:04
f6b0f7e to
65d982c
Compare
tawei-amd
marked this pull request as ready for review
October 19, 2026 14:04
tawei-amd
commented
Oct 19, 2026
tawei-amd
left a comment
tawei-amd
left a comment
Collaborator
Author
There was a problem hiding this comment.
Self-review (author). Two small findings in the sweep; neither affects today's single-model catalog.
🤖 Generated with Claude Code
tawei-amd
force-pushed
the
feat/d3-lite
branch
from
October 20, 2026 16:11
fd17fb8 to
25f4d10
Compare
tawei-amd
force-pushed
the
ci/sweep-decision
branch
from
October 20, 2026 16:11
65d982c to
4090533
Compare
Per decision model, served with --decision: 15 stored requests answered by the upstream BF16 model (shape exact, top-1 agreement and mean |dp| within limits), choice / noul / score in one request, a 27-option question (two-letter code), invalid question (422), body that is not JSON (422), a prompt over the limit (max_length_exceeded) and the example of docs/docs/models/d3.md. SweepTask gains needs_decision. Kept out of the model-sweep matrix with embedding and audio, per that workflow's rule. Co-Authored-By: Claude <[email protected]>
…ON bodies - Each further decision model restarts the server with its own --decision tag; before, every model was measured against the first model's server. --no-server with more than one model is refused. - The docs case also checks that the server decides what the documented reply decides (choice, noul side, most likely score level); the shape alone let the documented numbers drift. - A reply body that is not JSON (e.g. a proxy's 5xx page) fails its case instead of raising out of the sweep. Co-Authored-By: Claude <[email protected]>
tawei-amd
force-pushed
the
feat/d3-lite
branch
from
October 22, 2026 21:59
ec777ad to
6771da0
Compare
tawei-amd
force-pushed
the
ci/sweep-decision
branch
from
October 22, 2026 21:59
ba6b6ce to
d692a6b
Compare
tawei-amd
added a commit
that referenced
this pull request
Oct 23, 2026
feat(d3): image input for d3:lite (draft; stacked on #817)
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
.github/script/sweep_decision.py: the decision-model sweep (plan F-3), stacked on #816. Per decision model, served with--decision, it runs:AA).invalid_questions, nothing runs.max_length_exceeded.docs/docs/models/d3.md, so the docs can't drift from the server.SweepTaskgainsneeds_decision. Followingmodel-sweep.yml's own rule, the sweep stays out of the matrix with embedding and audio until llm and vision have proven themselves.Test results
All 7 cases pass on this machine against the local XRT build and the HRX build.
🤖 Generated with Claude Code