Skip to content

Pull requests: ggml-org/llama.cpp

New pull request New
802 Open 9,417 Closed
802 Open 9,417 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

Mention ANV_SYS_MEM_LIMIT in Vulkan/Linux section of build.md documentation Improvements or additions to documentation
#20670 opened Mar 17, 2026 by adapt-L Loading…
ggml-webgpu: Update the RMS_NORM preprocessor and add L2_NORM documentation Improvements or additions to documentation ggml changes relating to the ggml tensor library for machine learning WebGPU
#20665 opened Mar 17, 2026 by yomaytk Loading…
ggml-webgpu: Add supports for DIAG and TRI documentation Improvements or additions to documentation ggml changes relating to the ggml tensor library for machine learning WebGPU
#20664 opened Mar 17, 2026 by yomaytk Loading…
1
vulkan: change gated_delta_net to shard a column across a subgroup ggml changes relating to the ggml tensor library for machine learning Vulkan Issues specific to the Vulkan backend
#20662 opened Mar 17, 2026 by jeffbolznv Loading…
1
Fix chat parser regressions: inference crashes/frozen; output backtracked testing Everything test related
#20660 opened Mar 17, 2026 by jpohhhh Loading…
7
vulkan: dequantize iq4_xs 4 at a time ggml changes relating to the ggml tensor library for machine learning Vulkan Issues specific to the Vulkan backend
#20657 opened Mar 16, 2026 by netrunnereve Loading…
Make debug build possible on Windows with HIP backend. ggml changes relating to the ggml tensor library for machine learning
#20655 opened Mar 16, 2026 by Exile333 Loading…
Models: Add control vector functions to qwen3.5 and qwen-next models model Model specific
#20653 opened Mar 16, 2026 by GreyWorks Loading…
1
2
tests: set dist_sampling seq_id to 0 testing Everything test related
#20645 opened Mar 16, 2026 by taronaeo Loading…
3
ggml-cuda: Add NVFP4 dp4a kernel ggml changes relating to the ggml tensor library for machine learning Nvidia GPU Issues specific to Nvidia GPUs python python script changes
#20644 opened Mar 16, 2026 by michaelw9999 Loading…
9
mtmd, server : add Voxtral Realtime 4B support with /v1/audio/transcriptions endpoint examples model Model specific python python script changes server
#20638 opened Mar 16, 2026 by didlawowo Loading…
[CUDA] Use a single warp per element instead of a single block per element if the K-dimension is small ggml changes relating to the ggml tensor library for machine learning Nvidia GPU Issues specific to Nvidia GPUs
#20635 opened Mar 16, 2026 by gaugarg-nv Loading…
8
ggml-cpu: add 128-bit RVV implementation for Quantization Vector Dot ggml changes relating to the ggml tensor library for machine learning
#20633 opened Mar 16, 2026 by rehan-10xengineer Loading…
9
model : add QKV weight fusion for LLaMA, Qwen2, and Qwen3 model Model specific python python script changes script Script related
#20628 opened Mar 16, 2026 by JoursBleu • Draft
7
ggml-cpu: simd_gemm implementation for riscv vector extension ggml changes relating to the ggml tensor library for machine learning
#20627 opened Mar 16, 2026 by rehan-10xengineer Loading…
[OpenVINO backend] add func is_splited_model() ggml changes relating to the ggml tensor library for machine learning
#20626 opened Mar 16, 2026 by zhaixuejun1993 Loading…
1
MiroThinker tool call parser
#20624 opened Mar 16, 2026 by hksdpc255 Loading…
9
Speed up claude-code by prevent adding system message that starts with 'x-anthropic-' examples server
#20623 opened Mar 16, 2026 by hksdpc255 Loading…
2
Replace UINT64_MAX with LONG_TIMEOUT in blocking case to isolate possibility of Dawn/llvm-pipe bug ggml changes relating to the ggml tensor library for machine learning
#20621 opened Mar 16, 2026 by nikhilJain17 • Draft
kleidiai : fix MUL_MAT support for batched (3D) inputs ggml changes relating to the ggml tensor library for machine learning
#20620 opened Mar 16, 2026 by jabr Loading…
1
3
ggml webgpu: Move to no timeout for WaitAny in graph submission to avoid deadlocks ggml changes relating to the ggml tensor library for machine learning
#20618 opened Mar 16, 2026 by reeselevine Loading…
1
devops: add an intel mkl container devops improvements to build systems and github actions
#20616 opened Mar 16, 2026 by kannon92 Loading…
ggml: MXFP flash attention with SoA layout (CPU scalar reference) examples ggml changes relating to the ggml tensor library for machine learning OpenCL Issues specific to the OpenCL backend SYCL https://en.wikipedia.org/wiki/SYCL - GPU programming language testing Everything test related
#20609 opened Mar 15, 2026 by timothyeburke • Draft
12
ggml blas: set mkl threads from thread context devops improvements to build systems and github actions ggml changes relating to the ggml tensor library for machine learning
#20602 opened Mar 15, 2026 by kannon92 Loading…
vulkan: allow graphics queue only through env var ggml changes relating to the ggml tensor library for machine learning Vulkan Issues specific to the Vulkan backend
#20599 opened Mar 15, 2026 by 0cc4m Loading…
1
2
ProTip! no:milestone will show everything without a milestone.