Skip to content

Pull requests: ggml-org/llama.cpp

New pull request New
812 Open 9,639 Closed
812 Open 9,639 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

model: add Falcon OCR support
#21045 opened Mar 26, 2026 by avirajBevli Loading…
1
llama : rotate activations for better quantization
#21038 opened Mar 26, 2026 by ggerganov • Draft
Send reasoning content back to the model across turns via the reasoning_content API field examples server
#21036 opened Mar 26, 2026 by ServeurpersoCom Loading…
1
1
webui: Add option to pre-encode conversation for faster next turns examples server
#21034 opened Mar 26, 2026 by allozaur • Draft
1
1
model: add support for nvidia/gpt-oss-puzzle-88B model Model specific python python script changes
#21032 opened Mar 26, 2026 by smpurkis Loading…
1
3
rpc : proper handling of data pointers to CPU buffers ggml changes relating to the ggml tensor library for machine learning
#21030 opened Mar 26, 2026 by rgerganov Loading…
1
@rgerganov
3
vulkan: add FA dequant for q4_1, q5_0, q5_1, iq4_nl ggml changes relating to the ggml tensor library for machine learning Vulkan Issues specific to the Vulkan backend
#21029 opened Mar 26, 2026 by mkoker Loading…
1
fix: mtmd "v.patch_embd" quant and unsupported im2col ops on Metal for deepseek-ocr python python script changes
#21027 opened Mar 26, 2026 by sfallah Loading…
7
WebUI: Replace illegal nested button elements examples server
#21026 opened Mar 26, 2026 by bluemoehre Loading…
Vulkan Q4_0 Repack PoC ggml changes relating to the ggml tensor library for machine learning Vulkan Issues specific to the Vulkan backend
#21024 opened Mar 26, 2026 by 0cc4m • Draft
webui: Conversation forking + branching improvements examples server
#21021 opened Mar 26, 2026 by allozaur Loading…
1
11
hexagon: support for IQ4_NL and MXFP4 ggml changes relating to the ggml tensor library for machine learning Hexagon
#21018 opened Mar 26, 2026 by njsyw1997 Loading…
2
contrib: update AI policy to allow private repo
#21007 opened Mar 25, 2026 by ngxson Loading…
4
grammar: increase MAX_REPETITION_THRESHOLD + make it configurable via envvar testing Everything test related
#21003 opened Mar 25, 2026 by pwilkin Loading…
ci: introduce audits for self-hosted runners devops improvements to build systems and github actions
#21002 opened Mar 25, 2026 by taronaeo • Draft
7
webui: Improve Chat Messages initial scroll + auto-scroll logic + add lazy loading with transitions to content blocks examples server
#20999 opened Mar 25, 2026 by allozaur Loading…
CUDA: Add Flash Attention Support for Head Dimension 512 ggml changes relating to the ggml tensor library for machine learning Nvidia GPU Issues specific to Nvidia GPUs python python script changes
#20998 opened Mar 25, 2026 by anavp-nvidia Loading…
5
server : add LLAMA_KV_KEEP_ONLY_ACTIVE to clear idle slots' KV examples python python script changes server
#20993 opened Mar 25, 2026 by yychyo Loading…
8
Update Android link in README.md
#20992 opened Mar 25, 2026 by fmatosqg Loading…
ci: add riscv64 to release binaries
#20991 opened Mar 25, 2026 by gounthar Loading…
1
5
ggml : skip null devices in backend registry
#20983 opened Mar 25, 2026 by changgeunJi Loading…
llama-model-loader: use pinned memory for tensor overrides
#20978 opened Mar 25, 2026 by am17an Loading…
16
mtmd: Add DeepSeekOCR 2 Support
#20975 opened Mar 25, 2026 by sfallah • Draft
4
common : inhibit lazy grammar sampler while reasoning is active examples server testing Everything test related
#20970 opened Mar 25, 2026 by aldehir Loading…
5
Install libraries into LLAMA_LIB_INSTALL_DIR
#20966 opened Mar 25, 2026 by WhyNotHugo Loading…
2 tasks done
ProTip! no:milestone will show everything without a milestone.