Skip to content

Pull requests: antirez/ds4

New pull request New
143 Open 136 Closed
143 Open 136 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

Add served model name option for server discovery
#456 opened Jun 25, 2026 by RiccardoFiorentini Loading…
1
Metal: keep selected-address SSD prefill opt-in by default
#454 opened Jun 25, 2026 by andreaborio • Draft
1
2
Fix typo in README
#453 opened Jun 25, 2026 by mwbini Loading…
Support SSD streaming for Q4_K routed experts on ROCm
#451 opened Jun 24, 2026 by kmc6042 Loading…
4
Protect incoming KV prefix during live miss
#448 opened Jun 23, 2026 by JordiPosthumus Loading…
2
Fix ROCm Q8->F16 cache reserve starving session tensors on large models (q4q2)
#446 opened Jun 23, 2026 by alantsev Contributor Loading…
AGENTS.md rename (and server performance improvements?)
#443 opened Jun 21, 2026 by OPS-NeoRetro Loading…
2
Add Quickstart section to README
#438 opened Jun 18, 2026 by sethconvex Loading…
cuda: generalize router-select for arbitrary expert count (fixes Pro on CUDA, #427)
#435 opened Jun 17, 2026 by newjordan • Draft
1
1
Fix quality-score link after streaming refactor
#434 opened Jun 17, 2026 by andreaborio Loading…
Fix server JSON duplicate-field cleanup
#433 opened Jun 17, 2026 by 539hex Loading…
Add reverse distributed topology with coordinator-owned output suffix
#430 opened Jun 16, 2026 by lobanov Loading…
1
Handle modified Enter as newline in multiline linenoise
#426 opened Jun 16, 2026 by ljubomirj • Draft
Fix: ds4-server rejects HTTP requests using Transfer-Encoding: chunked
#423 opened Jun 16, 2026 by moritzburgard Loading…
agent: reject edit calls whose new= text contains [upto]
#421 opened Jun 16, 2026 by aledesogusbusiness-hue Loading…
1
Metal: protect tensor alloc/free byte counters with a mutex
#420 opened Jun 16, 2026 by aledesogusbusiness-hue Loading…
1
server: expose only the loaded model in /v1/models
#419 opened Jun 16, 2026 by aledesogusbusiness-hue Loading…
1
2
Metal: FP8-packed compressed-KV cache + long-context memory optimizations
#418 opened Jun 16, 2026 by aledesogusbusiness-hue Loading…
1
Metal: FP8-packed compressed-KV cache + long-context memory optimizations
#416 opened Jun 15, 2026 by lixiangnlp Loading…
docs(speed-bench): add generated benchmark summary
#413 opened Jun 15, 2026 by dutifulbob Loading…
1
Fix bug with impact on DeepSeek V4 Pro MTP Drafter usage
#411 opened Jun 14, 2026 by Deviad Loading…
2
rocm: fix distributed inference on unified-memory APUs (strix halo / gfx1151)
#407 opened Jun 13, 2026 by kyuz0 Loading…
1
[3/N] add prefetch support for CUDA backend : running ds4 for any GPU with cache (2.75 x faster!)
#402 opened Jun 12, 2026 by yiakwy-xpu-ml-framework-team Loading…
2
Local gen with dist prefill
#401 opened Jun 12, 2026 by lobanov • Draft
2
Extract dashboard into standalone ds4_dashboard module
#400 opened Jun 12, 2026 by EdwQ Loading…
ProTip! Type g i on any issue or pull request to go back to the issue listing page.