Skip to content

Pull requests: Luce-Org/lucebox

New pull request New
47 Open 398 Closed
47 Open 398 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

Restore DeepSeek4 HIP backend safely
#514 opened Jul 12, 2026 by howard0su Contributor Loading…
11
perf(ggml/hip): add batched ROCmFPX MMQ
#513 opened Jul 12, 2026 by weicj Collaborator • Draft
perf(deepseek4): enable causal Flash prefill
#512 opened Jul 11, 2026 by weicj Collaborator Loading…
16
fix(ds4): propagate sampler config in layer split
#508 opened Jul 11, 2026 by Graffioh Contributor • Draft
fix(ds4): correct HC shard handoff and scope runtime caches per shard
#507 opened Jul 11, 2026 by Graffioh Contributor • Draft
1
perf(dflash): split-chain fast rollback with F32 SSM checkpoints (+29% decode)
#506 opened Jul 10, 2026 by Amidwestnoob Loading…
10
feat(ipc): add DeepSeek4 DSpark shared worker
#505 opened Jul 10, 2026 by davide221 Contributor • Draft
perf(cuda-hip): partial-top-k + hipCUB large-N for argsort/top-k on ROCm
#504 opened Jul 10, 2026 by geometric Bot • Draft
fix/feat(ds4): ROCmFPX byte-safe decode + fused-decode path + ds4 optimizations
#503 opened Jul 8, 2026 by davide221 Contributor Loading…
10
feat(ggml): vendor ROCmFPX quant support for DeepSeek V4 Flash
#502 opened Jul 8, 2026 by davide221 Contributor Loading…
18
chore(deps): bump the github-actions group across 1 directory with 7 updates
#498 opened Jul 7, 2026 by dependabot Bot Loading…
1
feat(deepseek4): add DSpark speculative runtime
#496 opened Jul 7, 2026 by davide221 Contributor • Draft
2
feat(server): split tools from chat so PFlash works on agent turns
#492 opened Jul 5, 2026 by davidmroth Contributor Loading…
26
fix(dflash): scale DeepSeek4 graph sizing for large chunks
#490 opened Jul 5, 2026 by weicj Collaborator • Draft
3
feat(dflash): add direct DeepSeek4 pure-UMA runtime path
#489 opened Jul 5, 2026 by weicj Collaborator Loading…
16
feat(ipc): optional GPU isolation for target-shard daemon (same-backend splits)
#483 opened Jul 3, 2026 by davide221 Contributor • Draft
4 tasks
feat(server): add MoE expert-split multi-target foundation
#480 opened Jul 2, 2026 by weicj Collaborator • Draft
fix(laguna) LagunaBackend ABI mismatch from KvFlashPager
#477 opened Jul 1, 2026 by davide221 Contributor • Draft
feat(server): add Qwen35MoE target layer-split adapter
#475 opened Jul 1, 2026 by weicj Collaborator Loading…
6
perf(qwen35): use in-place GDN state write in pure AR
#473 opened Jun 30, 2026 by dusterbloom Collaborator Loading…
5
perf(qwen35): opt-in skip of graph FWHT K-rotation for q4_0 (DFLASH_SKIP_WHT)
#471 opened Jun 30, 2026 by dusterbloom Collaborator Loading…
3
fix(server): grow exact prefix-cache snapshot chains
#468 opened Jun 30, 2026 by dusterbloom Collaborator Loading…
7
feat(kvflash): pager serialize/deserialize + QK residency library
#466 opened Jun 30, 2026 by dusterbloom Collaborator Loading…
3
perf(laguna): add DDTree verify graph experiments
#462 opened Jun 29, 2026 by davide221 Contributor • Draft
Add switchable draft LoRA variants for Laguna
#460 opened Jun 29, 2026 by davide221 Contributor • Draft
ProTip! no:milestone will show everything without a milestone.