ggml-org/llama.cpp

Pull Requests Commits

ci : add self-hosted ui workflow

ggerganov committed 27 days ago

Verified 87f18f76

ci : move python requirements check to CPU runners

ggerganov committed 27 days ago

Verified cf285e19

ci : add comment about UI jobs

ggerganov committed 27 days ago

Verified 07ec9fd8

ci : back to 3.11

ggerganov committed 27 days ago

Verified 5a2e7684

ci : move lint back to 3.11

ggerganov committed 27 days ago

Verified 5a727def

ci : try to bump 3.11 -> 3.13

ggerganov committed 27 days ago

Verified f0bbb1a9

ci : prevent cmake pkg to run on dedicated fast runners

ggerganov committed 27 days ago

Verified a0a98e70

ci : prevent heavy CPU jobs from running on fast runners

ggerganov committed 27 days ago

Verified 8c75e6ee

ci : slim -> self-hosted

ggerganov committed 27 days ago

Verified 651afdb4

ci : remove tag from build-self-hosted.yml

ggerganov committed 27 days ago

Verified 5f0e5348

ci : update build-self-hosted.yml (#23616)

ggerganov committed 27 days ago

Verified 549b9d84

convert : minor fixes for numpy 2.x (#23571)

CISC committed 27 days ago

Verified 5d246a79

cmake : fix ui build (#23592)

aldehir committed 27 days ago

Verified 63248fc3

server: add margin for draft model for `fit` (#23485)

am17an committed 27 days ago

Verified 83eebe9d

TP: fix entirely zero-sized slices per device (#23525)

JohannesGaessler committed 27 days ago

Verified fff63b51

opencl: batch profiling to improve speed and prevent memory leaks (#23495)

shaofeiqi committed 27 days ago

Verified f3061116

hexagon: apply repl optimization in flash attn softmax as #22993 (#23455)

njsyw1997 committed 27 days ago

Verified 1c0f6db5

snapdragon: update windows toolchain to use hsdk v6.6.0.0 (#23552)

aparmp-quic committed 27 days ago

Verified cec51c7a

cmake/ui : refactor the build (#23352)

aldehir committed 28 days ago

Verified b22ff4b7

requirements : bump torch to 2.11.0 (#23503)

adityasingh2400 committed 28 days ago

Verified c0c7e147

model : add NVFP4 MTP scale tensors (#23563)

michaelw9999 committed 28 days ago

Verified b0df4c0c

ggml : Check the right iface method before using the fallback 2d get (#23514)

dskwe committed 28 days ago

Verified a4974763

vulkan: fix windows find_package of SPIRV-Headers (#23215)

jeffbolznv committed 28 days ago

Verified 95405ac6

opencl: generalize Adreno MoE kernels on M (#23449)

shawngu-quic committed 28 days ago

Verified 0f3cb3fc

server: only parse empty msg if continuing an assistant msg (#23506)

aldehir committed 29 days ago

Verified 1acee6bf

perplexity : fix integer overflow (#23496)

fairydreaming committed 29 days ago

Verified ef570f63

SYCL: improve MoE prefill throughput (#23142)

sanmai committed 29 days ago

Verified cc9e3312

sycl : Level Zero detection in ggml_sycl_init (#23097)

sanmai committed 29 days ago

Verified bcfd1989

SYCL : gated_delta_net K>1 (#23174)

karavayev committed 29 days ago

Verified 56f16f23

SYCL: add BF16 to DMMV kernel path (~4x tg speedup on Intel Arc) (#21580)

PMZFX committed 29 days ago

Verified 8cc67efc

Older