feat(server): add Gemma4 draft residency support#291
Open
weicj wants to merge 3 commits into
Open
Conversation
weicj
force-pushed
the
feat-gemma4-draft-residency-followup
branch
from
8e54210 to
25f0efa
Compare
![cubic-dev-ai[bot]](https://avatars.githubusercontent.com/in/1082092?s=60&v=4){kind=small}
Contributor
There was a problem hiding this comment.
1 issue found across 10 files
Reply with feedback, questions, or to request a fix.
Re-trigger cubic
easel
pushed a commit
to easel/lucebox-hub
that referenced
this pull request
May 31, 2026
Merge PR Luce-Org#291's draft residency policy and Gemma4 draft lifecycle support into the current auto-integration stack while preserving passthrough PFlash options, transitive compression overrides, status/cancellation behavior, and existing server unit coverage. PR Luce-Org#290 is represented by this superset head.
easel
pushed a commit
to easel/lucebox-hub
that referenced
this pull request
May 31, 2026
Record the Luce-Org#291/Luce-Org#290 draft-residency integration, newly non-draft Luce-Org#321/Luce-Org#325 classification, validation, and retained worktree/transcript paths for the May 31 13:30 UTC run.
easel
pushed a commit
to easel/lucebox-hub
that referenced
this pull request
May 31, 2026
Record the post-push Luce-Org#324 head advance and conflict resolution after integrating Luce-Org#291/Luce-Org#290.
easel
pushed a commit
to easel/lucebox-hub
that referenced
this pull request
May 31, 2026
Integrate current Luce-Org/lucebox-hub PR Luce-Org#291 head into the auto-integration patch stack.
easel
pushed a commit
to easel/lucebox-hub
that referenced
this pull request
May 31, 2026
Record the exact Luce-Org#290/Luce-Org#291 merges, current Luce-Org#321/Luce-Org#325 classification, retained worktrees, and validation for the 2026-05-31 13:57 integration run.
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
1 participant
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
This follow-up PR extends the shared draft residency policy from #290 with Gemma4 draft-only park/unpark support.
#290 keeps Gemma4 out of scope to keep the policy change small. Gemma4 needs a slightly larger internal reshaping because its draft-load path currently lives inside
Gemma4Backend::init(), so draft-only residency requires extracting that path into reusable load/free helpers.Changes
Gemma4Backend::init()into a reusable helper.Gemma4Backend::park("draft")andGemma4Backend::unpark("draft")to release/reload the draft without unloading the target model.