Skip to content

Pull requests: antirez/ds4

New pull request New
97 Open 106 Closed
97 Open 106 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

fix: limit direct/full-map fallbacks to layer span cover for distributed CUDA workers (refs #293)
#317 opened May 31, 2026 by hexxyan Contributor Loading…
Add ACP stdio mode to ds4-agent
#315 opened May 31, 2026 by omercelik Loading…
cuda: add DS4_CUDA_MANAGED env var for full UMA pool access on Strix Halo
#313 opened May 31, 2026 by kmc6042 Loading…
6
ds4_server: silence the warning
#312 opened May 31, 2026 by alantsev Contributor Loading…
rocm: fix gfx1151 correctness (precise expert-selection math), enable indexer, enable agent
#311 opened May 31, 2026 by alantsev Contributor Loading…
Add ds4-onboarding skill (docs-grounded agent skill tutor)
#310 opened May 31, 2026 by maeste Loading…
1
ds4_test: make metal-tensor-equivalence test deterministic on CUDA/ROCm
#308 opened May 31, 2026 by alantsev Contributor Loading…
bench: add routed expert locality profiler
#307 opened May 31, 2026 by hexxyan Contributor Loading…
5 of 6 tasks
metal: simdgroup MMA mini-GEMM for decode MoE [experimental]
#306 opened May 31, 2026 by hexxyan Contributor Loading…
5 of 8 tasks
1
cpu: add x86 AVX2/AVX512 SIMD fast paths for quantized MoE dot products
#305 opened May 30, 2026 by hexxyan Contributor Loading…
5 of 7 tasks
Structured outputs via llguidance
#302 opened May 30, 2026 by pminervini Loading…
1
server: fix client-triggerable double-free in JSON request parsing
#301 opened May 30, 2026 by ferdinandobons Loading…
fix(server): expose loaded model metadata
#287 opened May 29, 2026 by derogab Loading…
Teacher-Forced Directional Steering ideas-to-explore
#282 opened May 28, 2026 by Chida82 Loading…
4
feat: add REAP-compact GGUF support
#281 opened May 28, 2026 by ljubomirj Loading…
2
Support live Responses previous_response_id reuse
#280 opened May 28, 2026 by nhwaani Loading…
Add Q4_K MoE prefill kernels + CUDA 11.x compatibility for V100
#279 opened May 28, 2026 by eazlong Loading…
2 of 4 tasks
Tune suffix decoding defaults for M5 Max Metal
#276 opened May 27, 2026 by nhwaani Loading…
2
Add DS4_EMBED_KERNELS flag to embed Metal kernel sources into binary
#274 opened May 27, 2026 by kilork Loading…
Harden DSML tool continuations
#273 opened May 27, 2026 by harrychin Loading…
1
Makefile: use -march=native on x86_64 Darwin
#271 opened May 27, 2026 by hexxyan Contributor Loading…
download_model.sh: honor HF_ENDPOINT for Hugging Face mirrors
#270 opened May 27, 2026 by crimsondhaks Loading…
1
Add --api-key Bearer token authentication to ds4-server
#269 opened May 27, 2026 by hexxyan Contributor Loading…
1
Document wide MoE prefill tile setting
#268 opened May 27, 2026 by nhwaani Loading…
1
cuda: increase matmul rows/block from 8 to 16
#266 opened May 27, 2026 by arpesenti Loading…
1
ProTip! Updated in the last three days: updated:>2026-05-29.