-
Notifications
You must be signed in to change notification settings - Fork 3
Pull requests: ohdearquant/lattice
Author
Label
Projects
Milestones
Reviews
Assignee
Sort
Pull requests list
fix(inference): fail-closed integrity guard for Qwen embedding cache (#504 first slice)
#544
opened Jul 2, 2026 by
ohdearquant
Owner
Loading…
fix(inference): thread seed into mtp_verify_draft for reproducible MTP-verify (#329)
#543
opened Jul 2, 2026 by
ohdearquant
Owner
Loading…
test(inference): regression coverage for PagedKVCache Lru overflow (#292)
#542
opened Jul 2, 2026 by
ohdearquant
Owner
Loading…
fix(inference): correct partial-block asymmetric Q4 scale (#346)
#541
opened Jul 2, 2026 by
ohdearquant
Owner
Loading…
fix(inference): bound IncrementalDetokenizer retention to the decode-boundary tail window (#324)
#538
opened Jul 2, 2026 by
ohdearquant
Owner
Loading…
feat(inference): W3 3-bit MLP-only weight path (#420) [FOUNDER-GATED: silent-quality-loss review]
#515
opened Jul 1, 2026 by
ohdearquant
Owner
•
Draft
feat(tune): surface-B GDN LoRA weight gradients + train_grad_full fields
#202
opened Jun 22, 2026 by
ohdearquant
Owner
•
Draft
ProTip!
Exclude everything labeled
bug with -label:bug.