Skip to content

Pull requests: ggml-org/llama.cpp

New pull request New
1,206 Open 11,992 Closed
1,206 Open 11,992 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

build : add ROCm component directories to ggml-hip RPATH ggml changes relating to the ggml tensor library for machine learning
#25899 opened Jul 19, 2026 by liminfei-amd Contributor Loading…
1
OpenCL: Disable unsupported allocation of Adreno large buffers to fix compilation failure in OpenCL 2.x ggml changes relating to the ggml tensor library for machine learning OpenCL Issues specific to the OpenCL backend
#25897 opened Jul 19, 2026 by mgroeber9110 Contributor Loading…
cuda: add sqrt_softplus in topk-moe for dsv4 CUDA Related to the CUDA backend ggml changes relating to the ggml tensor library for machine learning testing Everything test related
#25896 opened Jul 19, 2026 by am17an Contributor Loading…
metal: implement DSv4 Lightning Indexer Apple Metal https://en.wikipedia.org/wiki/Metal_(API) ggml changes relating to the ggml tensor library for machine learning testing Everything test related
#25893 opened Jul 19, 2026 by tarruda Contributor • Draft
I have used int32_t for llama_memory_seq_div
#25892 opened Jul 19, 2026 by dogukandemirci-software-engineer Loading…
1
common/chat: fix DeepSeek V4 parsing/encoding jinja parser Issues related to the jinja parser testing Everything test related
#25891 opened Jul 19, 2026 by tarruda Contributor • Draft
2
common: fix ctype(3) usage
#25888 opened Jul 19, 2026 by 0-wiz-0 Contributor Loading…
fit: warn when --cpu-moe is used on integrated/UMA GPUs; docs: add Jetson MTP benchmark documentation Improvements or additions to documentation
#25883 opened Jul 19, 2026 by hernandez42 Loading…
2
sycl: fix use-after-return of the SDPA scale in the oneDNN flash-attention path ggml changes relating to the ggml tensor library for machine learning SYCL https://en.wikipedia.org/wiki/SYCL - GPU programming language testing Everything test related
#25880 opened Jul 19, 2026 by meatposes Loading…
1
quantization: complete MXFP4 and NVFP4 examples
#25878 opened Jul 19, 2026 by misfit-alpaca Loading…
server: add read_image tool server/ui server
#25877 opened Jul 18, 2026 by parabelboi Loading…
1
2
Extended SYCL oneDNN SDPA to non-FP16 KV caches (Q4_0–Q8_0 and FP32) documentation Improvements or additions to documentation ggml changes relating to the ggml tensor library for machine learning SYCL https://en.wikipedia.org/wiki/SYCL - GPU programming language
#25874 opened Jul 18, 2026 by johnkarlhill • Draft
3
llama : enforce the same K and V cache types for DeepSeek V4; enable FA if V cache is quantized
#25871 opened Jul 18, 2026 by fairydreaming Collaborator Loading…
1
1
ui: fix collapsed user bubble with markdown rendering server/ui
#25869 opened Jul 18, 2026 by ServeurpersoCom Contributor Loading…
1
1
server : properly handle null llama_context server
#25868 opened Jul 18, 2026 by fairydreaming Collaborator Loading…
Update ggml/src/gguf.cpp : Defined virtual keyword for destructor of … ggml changes relating to the ggml tensor library for machine learning
#25867 opened Jul 18, 2026 by yymin1022 Loading…
ui: enable the agentic flow when only the JS sandbox is active server/ui
#25865 opened Jul 18, 2026 by ServeurpersoCom Contributor Loading…
1
1
server : fix dry_sequence_breakers missing from GET /props generation settings server
#25864 opened Jul 18, 2026 by verma8076 Loading…
1
Fix/hip apu host buffer CUDA Related to the CUDA backend ggml changes relating to the ggml tensor library for machine learning
#25863 opened Jul 18, 2026 by Victor-Loos Loading…
vulkan: use mul_mat_vec_id for high-expert-count MoE decode ggml changes relating to the ggml tensor library for machine learning Vulkan Issues specific to the Vulkan backend
#25862 opened Jul 18, 2026 by essentialols • Draft
2
Deepseek V4: split-mode tensor ggml changes relating to the ggml tensor library for machine learning testing Everything test related
#25860 opened Jul 18, 2026 by S3kundes • Draft
5
sycl: parallelize the non-contiguous concat kernel ggml changes relating to the ggml tensor library for machine learning SYCL https://en.wikipedia.org/wiki/SYCL - GPU programming language
#25852 opened Jul 18, 2026 by Titaniumtown Contributor Loading…
2
vulkan: add TQ2_0 (ternary) support ggml changes relating to the ggml tensor library for machine learning testing Everything test related Vulkan Issues specific to the Vulkan backend
#25850 opened Jul 18, 2026 by michaeltrabalka-tech Loading…
1 task done
2
server: server tools - get_datetime in UTC ISO format server
#25848 opened Jul 17, 2026 by radupotop Loading…
webgpu : add CONV_2D_DW (depthwise conv2d) kernel documentation Improvements or additions to documentation ggml changes relating to the ggml tensor library for machine learning WebGPU
#25847 opened Jul 17, 2026 by m1el Loading…
2
ProTip! Follow long discussions with comments:>50.