Pull Requests ggerganov/llama.cpp

fix(quantize): add NVFP4 default type mapping and scale tensors examples

#22897 opened 2026-05-10 08:34 by t-timms

[ggml] Fix Vulkan-Hpp handle usage on 32-bit targets. Vulkan ggml

#22892 opened 2026-05-10 04:06 by miyanyan

vendor : update cpp-httplib to 0.43.4 script python merge ready

#22888 opened 2026-05-09 22:37 by cabelo

vulkan: Switch MUL_MAT_VEC to 4 K per iteration for F16/32 Vulkan ggml

#22887 opened 2026-05-09 22:10 by TheBlueMatt

feat: add MiMo v2.5 vision examples python

#22883 opened 2026-05-09 20:52 by AesSedai

HIP: RDNA3 mma FA, faster AMD transpose, tune AMD Nvidia GPU ggml

#22880 opened 2026-05-09 19:43 by JohannesGaessler

docs: fix metrics endpoint description in server README examples server

#22879 opened 2026-05-09 18:25 by willjoha

Optimise memory usage by evicting weights after processing each layer examples

#22877 opened 2026-05-09 17:34 by EAddario

opencl: fix crash when warming up MoE on Adreno ggml OpenCL

#22876 opened 2026-05-09 17:23 by lhez

cli: exit conversation mode on stdin EOF examples

#22874 opened 2026-05-09 16:54 by YuWei-CH

server : fix n_predict=-2 (generate until context full) examples python server

#22873 opened 2026-05-09 16:13 by kimjune01

ggml-cpu: Add IME2 Instruction Support for the SpacemiT Backend documentation build devops ggml

#22863 opened 2026-05-09 06:47 by alex-spacemit

SYCL: implement ggml_sycl_pool_vmm ggml SYCL

#22862 opened 2026-05-09 06:40 by sanmai

ggml-cpu: scope KleidiAI compile flags per-target via OBJECT library ggml

#22861 opened 2026-05-09 05:36 by shreyanshp

security: fix critical integer overflow (CWE-190) in tensor allocation ggml

#22857 opened 2026-05-08 23:45 by programacionlogicT900r1000

opencl: add q4_1 MoE for Adreno ggml OpenCL

#22856 opened 2026-05-08 22:47 by shawngu-quic

vulkan: fuse snake activation (mul, sin, sqr, mul, add) Vulkan ggml

#22855 opened 2026-05-08 21:41 by ServeurpersoCom

Add --continue-after-failure support to llama-bench for resilient benchmark sweeps examples

#22854 opened 2026-05-08 21:29 by ssam18

docs: fix suggested cmake flag in `build.md` for including rocwmma documentation

#22852 opened 2026-05-08 20:48 by Elliot-Roberts

Add preemptive priority scheduling server/webui examples python server

#22851 opened 2026-05-08 20:11 by lowlyocean

webui: Add max image size option server/webui examples server

#22849 opened 2026-05-08 19:06 by stduhpf

cli: Add quiet mode examples

#22848 opened 2026-05-08 19:02 by nh2

test-backend-ops: add more fields to csv output. testing

#22844 opened 2026-05-08 16:13 by Exile333

server: Improve get_datetime tool to return structured ISO-8601 JSON examples server

#22843 opened 2026-05-08 15:29 by yamc12

ggml: fixed Arm SVE usage bug in vec.h, vec.cpp ggml

#22841 opened 2026-05-08 14:36 by martin-klacer-arm

spec : parallel drafting support examples server

#22838 opened 2026-05-08 12:44 by ggerganov

ggml-cpu : add STQ1_0 ternary quantization with ARM NEON vec_dot kernel examples python ggml

#22836 opened 2026-05-08 11:07 by sjl623

convert_hf_to_gguf: fix Qwen3.5 linear_num_value_heads overridden by AutoConfig defaults python

#22835 opened 2026-05-08 10:56 by LucaSforza

ggml : add ggml_conv_1d_grouped testing ggml

#22833 opened 2026-05-08 10:15 by Juste-Leo2

convert : add split() to LoraTorchTensor in LoRA converter testing python

#22832 opened 2026-05-08 08:42 by jesus-talavera-ibm