ggerganov/llama.cpp

Pull Requests Commits

add check script

ngxson committed 23 days ago

19f25590

add --synthid-keys

ngxson committed 23 days ago

8b46bb96

add synthid sampler

ngxson committed 23 days ago

73c900a9

llama: fix whole source code rebuilt on each new commit (#28278)

ngxson committed 24 days ago

Verified d230ddd7

common/json-schema : fix GBNF grammar generation for empty object schemas (#28279)

SergeySklyarov committed 24 days ago

Verified c5a5535e

opencl: quant lm_head / decode GEMV and medium-batch GEMM optimizations (speculative decoding/MTP) (#26477)

wanghqc committed 24 days ago

Verified 95ef7fc1

tune MMVQ to MMQ crossover for SM87 (#28285)

kbenkhaled committed 24 days ago

Verified 8c1a2516

snapdragon: ci updates to use new run script (#28293)

max-krasnyansky committed 24 days ago

Verified d30500b8

ops: add Hexagon to ops.md and update main README.md (#28263)

tboinovski1 committed 24 days ago

Verified e107984b

server : use pytest-xdist for server tests (#28298)

danbev committed 24 days ago

Verified 42f0225f

mtmd: propagate const to preproc class (#28310)

ngxson committed 24 days ago

Verified de8656bd

metal : add sparse FA (#28098)

ggerganov committed 24 days ago

Verified 7bb0fc18

metal : fix glu dispatch with ne00 = 1 (#28306)

ggerganov committed 24 days ago

Verified 0df017d6

mtmd : add `const` in various places (#28307)

madsmtm committed 24 days ago

Verified f45576aa

CUDA: Allow concurrent streams per split for multi-GPU (#28198)

tannerbruhn committed 24 days ago

Verified 0ba6499c

vulkan: fix FA dequant path engagement (#28190)

Nathanw1014 committed 24 days ago

Verified c7bda030

sycl : enhance the api to support peer-to-peer copy (#27550)

arthw committed 24 days ago

Verified 0df974d7

convert : skip bias_vl tensor in DeepSeek-V4 DSpark conversion (#28294)

ggerganov committed 24 days ago

Verified d646c9d1

misc : prevent RAM peaking at model loading stage (#27483)

tdakhran committed 24 days ago

Verified 5ec4eab6

sycl: reduce redundant work in Q4_K multi-column MMVQ (#27062)

Eurekatic committed 24 days ago

Verified 4aa6ffba

model: add NVIDIA Nemotron-3-Puzzle-75B-A9B (NemotronHPuzzle) support (#25444)

YanissAmz committed 24 days ago

Verified c61b98b8

mtmd: fix idefics3 preproc (#28273)

ngxson committed 24 days ago

Verified 67a17c17

finetune: fix no KV cache (#27199)

ngxson committed 25 days ago

Verified 159b7414

server : accept data: URLs for input_video and input_audio (#27735)

geckguy committed 25 days ago

Verified 9cffdcc8

ggml-hexagon: add F16 support for unary ops (#28228)

cqderek committed 25 days ago

Verified f027c4f1

mtmd: add mtmd_tokenize_from_parts() (#28250)

ngxson committed 25 days ago

Verified 73390547

metal : add fa-vec tunings for M3 (#28236)

init-22 committed 25 days ago

Verified 9cc33944

metal : fix memory query under low-memory conditions (#27701)

madsmtm committed 25 days ago

Verified 8c0b9cd0

ci : check for missing autoreleasepools (#27884)

nikwen committed 25 days ago

Verified 03dbcc53

Update ROCm to 10.0.0 release (#27803)

superm1 committed 25 days ago

Verified cff18443

Older