Skip to content

Pull requests: ggml-org/llama.cpp

New pull request New
1,272 Open 12,271 Closed
1,272 Open 12,271 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

common : add env to override n_rs_seq
#26499 opened Aug 3, 2026 by Green-Sky Collaborator • Draft
1
tools : split Metal FA-vec tuning into a standalone tuner Apple Metal https://en.wikipedia.org/wiki/Metal_(API) documentation Improvements or additions to documentation examples ggml changes relating to the ggml tensor library for machine learning testing Everything test related
#26498 opened Aug 3, 2026 by forforever73 Contributor Loading…
Deepseek4: concat inputs to stay under limit
#26496 opened Aug 3, 2026 by am17an Contributor Loading…
2
llama-bench : continue after setup errors examples
#26495 opened Aug 3, 2026 by zcxGGmu • Draft
1
CUDA: support ncols > 1024 in bitonic argsort/top-k The original bito… CUDA Related to the CUDA backend ggml changes relating to the ggml tensor library for machine learning
#26493 opened Aug 3, 2026 by Geramy Contributor Loading…
1
Deepseek 4: -sm tensor ggml changes relating to the ggml tensor library for machine learning testing Everything test related
#26490 opened Aug 3, 2026 by am17an Contributor Loading…
1
docs: fix inaccurate non-terminal naming rules in GBNF README documentation Improvements or additions to documentation
#26489 opened Aug 3, 2026 by JingliangGao Loading…
1
1
cuda: add GGML_CUDA_BLOCKING_SYNC to eliminate 100% CPU busy-wait on full GPU offload CUDA Related to the CUDA backend ggml changes relating to the ggml tensor library for machine learning
#26487 opened Aug 3, 2026 by hksk Loading…
2 tasks done
1
ggml : vectorize rms_norm reduce and fuse the scale write ggml changes relating to the ggml tensor library for machine learning testing Everything test related
#26486 opened Aug 3, 2026 by cschanaj • Draft
2
vendor : update cpp-httplib to 0.52.0 vendor
#26485 opened Aug 3, 2026 by cabelo Contributor Loading…
docs(server): note OpenAI client base_url for multi-model gateways documentation Improvements or additions to documentation server
#26483 opened Aug 3, 2026 by seven7763 Loading…
3 tasks
1
Add Vulkan SDK setup and caching to ubuntu release workflow devops improvements to build systems and github actions
#26479 opened Aug 2, 2026 by oscarbg Contributor Loading…
opencl: quant lm_head / decode GEMV and medium-batch GEMM optimizations (speculative decoding/MTP) ggml changes relating to the ggml tensor library for machine learning OpenCL Issues specific to the OpenCL backend
#26477 opened Aug 2, 2026 by wanghqc Contributor • Draft
opencl: fix q6_K flat mul_mat for Adreno A6x/A7x GPUs with older E031 compilers ggml changes relating to the ggml tensor library for machine learning OpenCL Issues specific to the OpenCL backend
#26476 opened Aug 2, 2026 by wanghqc Contributor • Draft
llama : allocate GLM_DSA indexer cache only in "full" indexer layers
#26474 opened Aug 2, 2026 by fairydreaming Collaborator • Draft
Preserve omitted assistant content during OpenAI-compatible message round-tripping testing Everything test related
#26473 opened Aug 2, 2026 by i386 Loading…
1
Fix issue where tool arguments can appear in any JSON-valid order testing Everything test related
#26472 opened Aug 2, 2026 by i386 Loading…
ggml : fuse soft_max sweeps into fewer passes ggml changes relating to the ggml tensor library for machine learning
#26468 opened Aug 2, 2026 by cschanaj Loading…
Instella moe conversion model Model specific
#26467 opened Aug 2, 2026 by csabakecskemeti Contributor Loading…
1
ggml-cuda: HIP replace __shfl_xor_sync with dpp instructions CUDA Related to the CUDA backend ggml changes relating to the ggml tensor library for machine learning
#26466 opened Aug 2, 2026 by thelittlefireman • Draft
opencl: make the MoE expert scatter deterministic ggml changes relating to the ggml tensor library for machine learning OpenCL Issues specific to the OpenCL backend
#26464 opened Aug 2, 2026 by wanghqc Contributor • Draft
ci: prepare to onboard AMD ROCm CI devops improvements to build systems and github actions
#26461 opened Aug 2, 2026 by taronaeo Member Loading…
llama: re-create the KV cache when flash attention resolves to disabled (performance and layout bug)
#26460 opened Aug 2, 2026 by wanghqc Contributor Loading…
ggml: add gfx90c HIP support CUDA Related to the CUDA backend ggml changes relating to the ggml tensor library for machine learning
#26454 opened Aug 2, 2026 by AuroraRAS Loading…
14
docs : clarify tensor split context sizing documentation Improvements or additions to documentation
#26453 opened Aug 2, 2026 by zcxGGmu Loading…
1 task done
1
1
ProTip! Exclude everything labeled bug with -label:bug.