ggml-org/llama.cpp

Pull Requests Commits

qwen4exp: add hc ops

am17an committed 42 days ago

26276bc9

ggml-cpu : disable PCH and fix CACHE_LINE_SIZE ambiguity to fix heap corruption (#28882)

ggerganov committed 42 days ago

Verified 2f539596

ci : trigger self-hosted CI on changes to ci/run.sh (#28859)

ggerganov committed 42 days ago

Verified 89fe2424

ci : remove gg_sum summary logic (#28857)

ggerganov committed 42 days ago

Verified 15d8f2d5

sycl : fix oneDNN scratchpad breaking the pool free order (#28704)

lslusarczyk committed 42 days ago

Verified 661643e4

common : move llama_n_rs_seq to before llama_decode (#28749)

danbev committed 43 days ago

Verified 093a2f86

ggml-cuda: fallback to F32 on device without BF16 hardware acceleration (#28846)

thelittlefireman committed 43 days ago

Verified ad6c6683

grammar : coalesce find + insert into a single insert and adjust move/copy mechanics (#26885)

HanClinto committed 43 days ago

Verified 7a16a6ce

tests : exclude HY_V4 from WebGPU test-llama-archs tests (#28855)

fairydreaming committed 43 days ago

Verified 5f436ddd

models : guard the expert FFN size fallback in nemotron-h against a zero divisor (#28779)

YanissAmz committed 43 days ago

Verified e49d2c27

ggml-cpu(s390x): guard VXE-only repack helpers (#28775)

bernardladenthin committed 43 days ago

Verified 69780529

tests : fix typo in test-quant-type-selection for nemotron 3 nano (#28835)

mctylr-gh committed 43 days ago

Verified 243a3082

sycl : Fix get mem error (#28227)

arthw committed 43 days ago

Verified b6b003d2

ci : run editorconfig and code-style checks on ubuntu-slim (#28854)

ggerganov committed 43 days ago

Verified c95f8e47

pi : prefer PI_MODEL_NAME env var for model disclosure (#28853)

ggerganov committed 43 days ago

Verified bc52a12b

tests : reduce FA test sizes (#28842)

ggerganov committed 43 days ago

Verified 4a899373

ci : remove leftover command (#28839)

CISC committed 43 days ago

Verified 37b3a9e0

ci : cap test-backend-ops parallel jobs at 2 and add a 3600s timeout (#28833)

ggerganov committed 43 days ago

Verified 002a12ad

vulkan: workaround NV queuesubmit driver bug (#28830)

jeffbolznv committed 43 days ago

Verified f1e44dcc

opencl: apply the noshuffle row-alignment rule to q4_K, q5_K and q8_0, not just q6_K (#28575)

wanghqc committed 43 days ago

Verified 56b9eb28

chat : improve parsing of complex types in qwen3-coder (#28742)

aldehir committed 44 days ago

Verified 790cf51a

common: add LOG_JSON macro to log structured data (#28586)

ngxson committed 44 days ago

Verified 8e330954

common : implement common_schema internal representation for JSON schemas (#28736)

aldehir committed 44 days ago

Verified acecd560

jinja : support dot property integer literals (#28817)

CISC committed 44 days ago

Verified ae9afff8

cmake: leave the timestamp out of precompiled headers on clang (#28816)

ServeurpersoCom committed 44 days ago

Verified 737e0980

ui : add cache (#28802)

angt committed 44 days ago

Verified 3057bb66

server : allow model downloads at model limit fix issue #26809 (#28530)

chenmiaoming committed 44 days ago

Verified 56381e40

ggml-cuda: hip add specific config table for AMD GCN (#27841)

thelittlefireman committed 44 days ago

Verified c8edceb0

server : add missing headers (#28795)

angt committed 44 days ago

Verified e192abb4

vendor : update cpp-httplib to 0.56.0 (#28787)

cabelo committed 44 days ago

Verified 718f7b41

Older