-
Notifications
You must be signed in to change notification settings - Fork 12.5k
Pull requests: ggml-org/llama.cpp
Author
Label
Projects
Milestones
Reviews
Assignee
Sort
Pull requests list
flush partial stop string when <EOG> is reached in /completion endpoint in streaming mode
examples
server
#15007
opened Aug 1, 2025 by
matteoserva
Loading…
fix compile bug when the BASE_CUDA_DEV_CONTAINER is based on Ubuntu 2…
devops
improvements to build systems and github actions
#15005
opened Aug 1, 2025 by
simevo
Loading…
llama-bench: rename DB table name from test to llama_bench
examples
python
python script changes
script
Script related
#15003
opened Aug 1, 2025 by
yeahdongcn
Loading…
Add support for CogVLM model
examples
python
python script changes
#15002
opened Aug 1, 2025 by
Tianyue-Zhao
Loading…
2 of 4 tasks
server: enable token array inputs for OAI API
examples
server
#15001
opened Jul 31, 2025 by
JohannesGaessler
Loading…
Handling delta.reasoning_content in llama-server's webUI
examples
server
#14997
opened Jul 31, 2025 by
mostlygeek
Loading…
OpenCL: add initial FA support
ggml
changes relating to the ggml tensor library for machine learning
OpenCL
Issues specific to the OpenCL backend
#14987
opened Jul 31, 2025 by
rmatif
Loading…
Draft: vulkan: Use coopmat2 for conv2d
ggml
changes relating to the ggml tensor library for machine learning
Vulkan
Issues specific to the Vulkan backend
#14982
opened Jul 31, 2025 by
jeffbolznv
•
Draft
CUDA: add set
ggml
changes relating to the ggml tensor library for machine learning
Nvidia GPU
Issues specific to Nvidia GPUs
#14980
opened Jul 31, 2025 by
jeemzz147
Loading…
ggml: WebGPU backend host improvements and style fixing
ggml
changes relating to the ggml tensor library for machine learning
#14978
opened Jul 30, 2025 by
reeselevine
Loading…
ggml: initial IBM zDNN backend
devops
improvements to build systems and github actions
documentation
Improvements or additions to documentation
ggml
changes relating to the ggml tensor library for machine learning
#14975
opened Jul 30, 2025 by
taronaeo
Loading…
Optimize l2_norm_f32 op with SIMD
ggml
changes relating to the ggml tensor library for machine learning
#14970
opened Jul 30, 2025 by
TIKki43
Loading…
Implementation of GGML_NUMA_MIRROR for 64% inferencing performance gain on numa systems
examples
ggml
changes relating to the ggml tensor library for machine learning
Nvidia GPU
Issues specific to Nvidia GPUs
testing
Everything test related
chat : fix multiple tool_calls on hermes-2-pro
testing
Everything test related
#14962
opened Jul 30, 2025 by
jhen0409
Loading…
llama : enable LLAMA_SET_ROWS=1 by default
examples
testing
Everything test related
#14959
opened Jul 30, 2025 by
ggerganov
Loading…
ggml : fix field name when new ggml_backend
ggml
changes relating to the ggml tensor library for machine learning
Nvidia GPU
Issues specific to Nvidia GPUs
OpenCL
Issues specific to the OpenCL backend
SYCL
https://en.wikipedia.org/wiki/SYCL - GPU programming language
Vulkan
Issues specific to the Vulkan backend
#14944
opened Jul 29, 2025 by
aisk
Loading…
ops: add MUSA
documentation
Improvements or additions to documentation
#14941
opened Jul 29, 2025 by
yeahdongcn
Loading…
vulkan: coopmat2 mul_mat optimizations
ggml
changes relating to the ggml tensor library for machine learning
Vulkan
Issues specific to the Vulkan backend
#14934
opened Jul 29, 2025 by
jeffbolznv
Loading…
vulkan: optimizations for direct convolution
ggml
changes relating to the ggml tensor library for machine learning
Vulkan
Issues specific to the Vulkan backend
#14933
opened Jul 29, 2025 by
jeffbolznv
Loading…
mtmd : support home-cooked Mistral Small Omni
examples
#14928
opened Jul 28, 2025 by
ngxson
Loading…
repack : optimize mul_mat_id path
ggml
changes relating to the ggml tensor library for machine learning
#14918
opened Jul 28, 2025 by
ggerganov
Loading…
1 task
opencl: fixed a typo
ggml
changes relating to the ggml tensor library for machine learning
OpenCL
Issues specific to the OpenCL backend
#14908
opened Jul 27, 2025 by
l29ah
Loading…
Previous Next
ProTip!
no:milestone will show everything without a milestone.