Skip to content

Pull requests: ggml-org/llama.cpp

New pull request New
1,142 Open 11,671 Closed
1,142 Open 11,671 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

chat: trim messages sent to StepFun parser (fixes long reasoning loops)
#25238 opened Jul 2, 2026 by pwilkin Member Loading…
1
common: Set optimal default thread count for ppc ( linux as well as AIX)
#25237 opened Jul 2, 2026 by shalinib-ibm Contributor Loading…
[SYCL] support OP cross_entropy_loss, cross_entropy_loss_back 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
#25236 opened Jul 2, 2026 by arthw Contributor Loading…
common,server : fix custom preset dedup against cached models server
#25235 opened Jul 2, 2026 by angt Member Loading…
[UT] enhance UT to show all real unsupported backends testing Everything test related
#25234 opened Jul 2, 2026 by arthw Contributor Loading…
chat: sanitize invalid UTF-8 before peg-native parsing devops improvements to build systems and github actions
#25233 opened Jul 2, 2026 by iaa2005 Loading…
llama : clear error when MTP draft shares KV cache across backends
#25232 opened Jul 2, 2026 by liminfei-amd Contributor Loading…
1 task done
[SYCL] fix unsupported UT cases of CONT & CPY 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
#25231 opened Jul 2, 2026 by arthw Contributor Loading…
Ensure unique node names and add org_src to track the org tensor for OpenVINO backend ggml changes relating to the ggml tensor library for machine learning testing Everything test related
#25230 opened Jul 2, 2026 by zhaixuejun1993 Contributor Loading…
3
vulkan: when using transfer queue for async copies, sync on event_wait to avoid race ggml changes relating to the ggml tensor library for machine learning Vulkan Issues specific to the Vulkan backend
#25229 opened Jul 2, 2026 by 0cc4m Contributor Loading…
CUDA: Support CUDA Virtual Devices CUDA Related to the CUDA backend ggml changes relating to the ggml tensor library for machine learning
#25228 opened Jul 2, 2026 by anavp-nvidia Contributor Loading…
4
server : don't list cached models when a preset is used server
#25226 opened Jul 2, 2026 by angt Member Loading…
2
ui: Improve performance when streaming server/ui
#25225 opened Jul 2, 2026 by ntowle Contributor Loading… UI: Refactoring & Code Quality
1
[SYCL] Flash Attention with XMX engine via oneDNN graph API (SDPA) on KV f16; Qwen3.6-27b-Q8_0 prefill speed up x1.21 at p=512 and x4.26 at p=80k ggml changes relating to the ggml tensor library for machine learning SYCL https://en.wikipedia.org/wiki/SYCL - GPU programming language
#25222 opened Jul 2, 2026 by hmscider Loading…
8
common : add missing <fstream> include in common.h
#25220 opened Jul 2, 2026 by zhangrunda Loading…
1 task done
1
1
vendor : update cpp-httplib to 0.49.0 vendor
#25218 opened Jul 2, 2026 by cabelo Contributor Loading…
sycl: add fused top-k MoE ggml changes relating to the ggml tensor library for machine learning SYCL https://en.wikipedia.org/wiki/SYCL - GPU programming language
#25217 opened Jul 2, 2026 by newjordan Contributor Loading…
9
hexagon: add VISION RoPE support ggml changes relating to the ggml tensor library for machine learning Hexagon
#25216 opened Jul 2, 2026 by aparmp-quic Contributor Loading…
llama : skip K/V rotation input when its buffer is unallocated
#25215 opened Jul 2, 2026 by liminfei-amd Contributor Loading…
1 task done
1
2
server: add --no-sleep flag for GPU heartbeat on headless GPUs CUDA Related to the CUDA backend ggml changes relating to the ggml tensor library for machine learning server SYCL https://en.wikipedia.org/wiki/SYCL - GPU programming language Vulkan Issues specific to the Vulkan backend
#25214 opened Jul 1, 2026 by johnkarlhill Loading…
6
rocm: fix mmap loading of large models CUDA Related to the CUDA backend ggml changes relating to the ggml tensor library for machine learning
#25212 opened Jul 1, 2026 by pwilkin Member Loading…
1
3
Optimize RWKV7 inference by fusing some graph operators Apple Metal https://en.wikipedia.org/wiki/Metal_(API) CUDA Related to the CUDA backend ggml changes relating to the ggml tensor library for machine learning model Model specific SYCL https://en.wikipedia.org/wiki/SYCL - GPU programming language testing Everything test related Vulkan Issues specific to the Vulkan backend
#25206 opened Jul 1, 2026 by MollySophia Collaborator • Draft
3
sycl: add GGML_SYCL_FATTN_VEC_NTHREADS build option ggml changes relating to the ggml tensor library for machine learning SYCL https://en.wikipedia.org/wiki/SYCL - GPU programming language
#25205 opened Jul 1, 2026 by Titaniumtown Loading…
1
llama: fix quantized kv-cache for dsv4 model Model specific
#25202 opened Jul 1, 2026 by am17an Contributor Loading…
7
llama-cli: fix passing chat_template_kwargs and reasoning_format params examples
#25201 opened Jul 1, 2026 by percontation Contributor Loading…
ProTip! Exclude everything labeled bug with -label:bug.