ggerganov/llama.cpp

Pull Requests Commits

ci: Bump CUDA Windows x64 builds to 13.4.1 (#28930)

shivamkumard-ctrl committed 35 days ago

Verified 0ecb159c

ci : fix android release (#28936)

CISC committed 35 days ago

Verified 987498f4

cuda : enable i16 and i32 for DUP (#28897)

amankarki151 committed 35 days ago

Verified 4c9233c0

cmake : use PROJECT_SOURCE_DIR instead of CMAKE_SOURCE_DIR (#28771)

danbev committed 35 days ago

Verified 69eb2506

webui: stop re-probing disabled /tools endpoint on every message (#28646)

geckguy committed 35 days ago

Verified 1bc7a5af

ci : reuse build tag name when used instead of safe one (#28911)

CISC committed 35 days ago

Verified 7cf1c54a

CI: hip-quality-check: ignore spill added in bfdc32183d57f1e35bacf35c47d6311e2028bbbc (#28909)

IMbackK committed 35 days ago

Verified 96ffdc41

HIP: fattn-mma: use fp32 accumulation on MFMA devices (#28576)

IMbackK committed 35 days ago

Verified bfdc3218

ci : add ubuntu-cuda builds to release (#28186)

ORippler committed 35 days ago

Verified 391fac16

qwen4exp: enable rms_norm + mul fusion (#28896)

am17an committed 36 days ago

Verified 41abbfd5

release : added gfx1103 to ubuntu rocm build (#28423)

apparle committed 36 days ago

Verified b4fa47d2

cmake : remove precompiled headers (#28892)

danbev committed 36 days ago

Verified f3a184b1

scripts: Add script to verify API/ABI compatibility (#28579)

ckastner committed 36 days ago

Verified dfe45163

llama.cpp : bump version to 0.4.1 (#28900)

ggerganov committed 36 days ago

Verified b29c606e

ggerganov committed 36 days ago

d9e03f10

ggml : bump version to 0.24.0 (ggml/1627)

ggerganov committed 36 days ago

eeea7316

tests : add fusion baseline README and broaden fusion CI triggers (#28893)

ggerganov committed 36 days ago

Verified bbdd9f24

models : fix incorrect uses of get_key_or_arr (#28868)

CISC committed 36 days ago

Verified 97e4ca73

models : fix mimo2 swa pattern load (#28865)

CISC committed 36 days ago

Verified 1aca1f9f

tests(s390x): add non-vxe build to tests (#28776)

taronaeo committed 36 days ago

Verified be2c6d7d

llama: add Maple 20B-A1B ternary MoE architecture (CPU) (#27000)

AlexGabbia committed 36 days ago

Verified 3d10bcd1

sycl: rfc: Use radix select for top_k (#28670)

cwriter committed 36 days ago

Verified 21f6b0d2

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

ggerganov committed 36 days ago

Verified 2f539596

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

ggerganov committed 36 days ago

Verified 89fe2424

ci : remove gg_sum summary logic (#28857)

ggerganov committed 36 days ago

Verified 15d8f2d5

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

lslusarczyk committed 36 days ago

Verified 661643e4

common : move llama_n_rs_seq to before llama_decode (#28749)

danbev committed 36 days ago

Verified 093a2f86

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

thelittlefireman committed 36 days ago

Verified ad6c6683

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

HanClinto committed 36 days ago

Verified 7a16a6ce

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

fairydreaming committed 36 days ago

Verified 5f436ddd

Newer Older