Skip to content

Pull requests: Luce-Org/lucebox-hub

New pull request New
28 Open 317 Closed
28 Open 317 Closed
Author
Filter by author
Loading
Label
Filter by label
Loading
Use alt + click/return to exclude labels
or ⇧ + click/return for logical OR
Projects
Filter by project
Loading
Milestones
Filter by milestone
Loading
Reviews
Filter by reviews
No reviews Review required Approved review Changes requested
Assignee
Filter by who’s assigned
Assigned to nobody Loading
Sort
Sort by
Newest Oldest Most commented Least commented Recently updated Least recently updated Best match
Most reactions
👍 👎 😄 🎉 😕 ❤️ 🚀 👀

Pull requests list

perf(kvflash): async page-out D2H and page-in H2D via pinned copy stream
#408 opened Jun 16, 2026 by cheese-cakee Contributor Loading…
3
feat(server): support KV Flash with same-backend target layer split
#391 opened Jun 16, 2026 by weicj Collaborator Loading…
4
spec-decode: eliminate replay pass via fast-rollback
#390 opened Jun 15, 2026 by howard0su Contributor • Draft
perf(server): reduce MoE expert-compute IPC overhead
#388 opened Jun 15, 2026 by weicj Collaborator • Draft
[codex] fix prefix cache for user-first prompts
#387 opened Jun 15, 2026 by easel Collaborator • Draft
feat(kvflash): target-QK residency scorer (--kvflash-policy qk)
#385 opened Jun 14, 2026 by dusterbloom Collaborator Loading…
3
[codex] unify cache capacity config
#381 opened Jun 14, 2026 by easel Collaborator • Draft
2
Wire prefill cache slots
#380 opened Jun 14, 2026 by easel Collaborator Loading…
2
feat(server): add cross-backend MoE expert compute foundation
#375 opened Jun 12, 2026 by weicj Collaborator • Draft
6
qwen35: sampled-verify — speculative decoding with an active sampler
#374 opened Jun 12, 2026 by Rhonstin Contributor Loading…
6
feat(dflash): add DeepSeek V4 Flash backend
#353 opened Jun 9, 2026 by howard0su Contributor • Draft
test(server): CPU-only HTTP server test rig (stub backend + scenarios)
#343 opened Jun 4, 2026 by easel Collaborator Loading…
4 tasks
3
feat(server): card-driven thinking control + reasoning_content channel + /props schema-4
#341 opened Jun 3, 2026 by easel Collaborator Loading…
6 tasks
18
feat(server): soft-close thinking termination (qwen35 + gemma4)
#339 opened Jun 3, 2026 by easel Collaborator Loading…
5 tasks
9
feat(luce-bench): in-tree bench harness + multi-turn agent_recorded + LLM judge
#337 opened Jun 3, 2026 by easel Collaborator Loading…
4 tasks
51
feat(lucebox): hub CLI + autotune/sweep/profile + harness adapters + shell wrapper
#335 opened Jun 3, 2026 by easel Collaborator Loading…
6 tasks
88
Add LLM Auto Context Compaction
#304 opened May 29, 2026 by howard0su Contributor • Draft
fix(server): Qwen3.6-27B tool calling for claude-code Anthropic path
#276 opened May 25, 2026 by dusterbloom Collaborator Loading…
5 of 7 tasks
4
feat(pflash): prefill compress up to 128k -> 2-12× prefill (content-dependent), decode at parity
#274 opened May 24, 2026 by dusterbloom Collaborator Loading…
38
feat(pflash): add dynamic threshold for chunk selection and lower auto-mode trigger
#249 opened May 22, 2026 by howard0su Contributor • Draft
feat(dflash): linear native MTP integrated decode CLI (stacked on #153)
#154 opened May 11, 2026 by javierpazo Contributor Loading…
2
feat(dflash): native Qwen3.6 MTP (NextN) runtime + contract test
#153 opened May 11, 2026 by javierpazo Contributor Loading…
3
Add Gemma 4 RTX 4090 backend helpers
#152 opened May 11, 2026 by adybag14-cyber Loading…
@dusterbloom
14
feat(dflash): accept FP16 safetensors drafter alongside BF16
#142 opened May 9, 2026 by javierpazo Contributor Loading…
chore(dflash): enforce sm_89 user override and keep BSA enabled
#137 opened May 9, 2026 by javierpazo Contributor Loading…
4
ProTip! Exclude everything labeled bug with -label:bug.