Skip to content

Pull requests: antirez/ds4

New pull request New
383 Open 226 Closed
383 Open 226 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

ROCm: improve gfx1151 IQ2 prefill
#992 opened Sep 6, 2026 by kyuz0 Contributor • Draft
Qwen3.8 flash next
#991 opened Sep 6, 2026 by ivanfioravanti Contributor Loading…
2
Add an experimental Qwen 3.8 Flash Next port on Metal
#990 opened Sep 6, 2026 by andrea82 • Draft
rocm: fix strix-halo build, rsqrtf() is device-only under HIP
#988 opened Sep 5, 2026 by MiklosPathy Loading…
Introduce welcome screen with DwarfStar logo
#987 opened Sep 5, 2026 by m3ftwz Loading…
kvstore: age the eviction value of unused checkpoints
#986 opened Sep 5, 2026 by JordiPosthumus Contributor Loading…
Preserve live KV cache across normalized tool replay
#985 opened Sep 5, 2026 by rratmansky Loading…
1
cuda: reduce long-context visual-attention scratch
#984 opened Sep 5, 2026 by JordiPosthumus Contributor Loading…
kvstore: stamp checkpoints with a behavioral tokenizer fingerprint
#983 opened Sep 5, 2026 by nazerim Loading…
3
Fix agent tool results on text-only DeepSeek engines
#982 opened Sep 5, 2026 by zhangjun8823-gif Loading…
Fix progressive JPEG AC refinement ZRL handling in Iris decoder
#981 opened Sep 5, 2026 by korale77 Loading…
cuda: reduce GB10 Q8 attention-output prefill overhead
#979 opened Sep 5, 2026 by JordiPosthumus Contributor Loading…
cuda: keep exact Q8 MMA loads within tensor bounds
#978 opened Sep 5, 2026 by JordiPosthumus Contributor Loading…
server: unify text and vision KV prefix reuse
#977 opened Sep 4, 2026 by HariboApfel Loading…
5 tasks done
Add WebP image decoding
#976 opened Sep 4, 2026 by Bendr0id Loading…
cuda: avoid unstable 512-thread top-k exchange
#975 opened Sep 4, 2026 by JordiPosthumus Contributor Loading…
agent: fork the current KV session without re-prefill
#972 opened Sep 4, 2026 by rinaldofesta Contributor Loading…
server: opt-in auto-reduction of over-budget image histories (DS4_VISION_KEEP_IMAGES)
#971 opened Sep 4, 2026 by nazerim Loading…
Steering: read GLP (GGUF Layer Projection) vectors
#970 opened Sep 4, 2026 by msuiche Loading…
Fix text tool observation path for non-GLM multimodal messages
#969 opened Sep 4, 2026 by pan-patryk Loading…
server: thinking-frontier bridge + vision encoder result cache
#968 opened Sep 4, 2026 by nazerim Loading…
7
Fix CUDA build: restore g_tp_block_ctx and add missing TP/kv-norm CUDA stubs
#967 opened Sep 4, 2026 by Matthley Loading…
2 of 3 tasks
metal: bypass DSpark speculation once a request proves unprofitable
#965 opened Sep 3, 2026 by rinaldofesta Contributor Loading…
Metal: accelerate GLM 5.3 Flash decode and prefill
#964 opened Sep 3, 2026 by trueimage Loading…
17
ds4-agent: fix every tool call failing on DeepSeek without --vision (+ regression test)
#963 opened Sep 3, 2026 by rinaldofesta Contributor Loading…
1
ProTip! Type g i on any issue or pull request to go back to the issue listing page.