ggerganov/llama.cpp

Pull Requests Commits

mtmd: cap max_image to ubatch for non_causal models

ngxson committed 10 days ago

808c5ee2

jinja : support coerced array attributes (#29574)

CISC committed 10 days ago

Verified ca2e2037

ggml-et : remove useless alloca() (#29663)

angt committed 10 days ago

Verified bdeb855b

ci : fix Models Backend Check by shortening the hrm_text fixture (#29744)

ServeurpersoCom committed 10 days ago

Verified 3b3d022b

llama: llama_prefetch_rows (#29599)

am17an committed 11 days ago

Verified 185103dc

ggml : add BF16 unary, GLU, binary and scale ops (CPU, CUDA) (#29675)

am17an committed 11 days ago

Verified 2090f60f

cpu: accept BF16 in src1 of mul_mat (#28937)

ServeurpersoCom committed 11 days ago

Verified 90c908d0

model-conversion : add --add-bos to run org model script (#29558)

danbev committed 11 days ago

Verified 8df332de

ui : shared model display primitives (#29644)

allozaur committed 11 days ago

Verified 4a096b8f

ui : model download pipeline (#27959)

allozaur committed 11 days ago

Verified 8664eaea

ui : model memory-fit estimation (#27957)

allozaur committed 11 days ago

Verified 4cfb6d1c

ui : Hugging Face Hub data layer (#27947)

allozaur committed 11 days ago

Verified 9b433361

ui : model id grammar for sidecars, quants and capability parsing (#27946)

allozaur committed 11 days ago

Verified f6532504

ui : type-safe API types, fetch helpers and download-ready models store plumbing (#29582)

allozaur committed 11 days ago

Verified fa2bde55

openvino: serve GET_ROWS on a weight view from the base Constant (#28381)

ServeurpersoCom committed 11 days ago

Verified 25747b08

ci : fix Fusion / metal by adding glm5-next to MTL.csv (#29712)

ServeurpersoCom committed 11 days ago

Verified db00347a

musa : define __CUDA_ARCH__ for device passes (#29508)

yeahdongcn committed 11 days ago

Verified 272aad8b

ci : add models backend check (#29651)

CISC committed 11 days ago

Verified 72db1e02

SYCL: reduce tensor allreduce sync with pinned host buffers (#29604)

Captain-Tripps committed 11 days ago

Verified 2a53ace3

add GLM-5.3-Flash (GLM5-Next) support (#27773)

timkhronos committed 11 days ago

Verified 649dcb10

vendor: update BoringSSL to 0.20260929.0 (#29669)

cabelo committed 11 days ago

Verified 931351ea

ggml-zdnn: impl buffer reset, fix memory leaks (#29637)

taronaeo committed 11 days ago

Verified eae11d22

Hexagon f16 activation ops (#29209)

cqderek committed 11 days ago

Verified 19e28a27

gguf : reject tensor size that wraps after padding (#26979)

x14ngch3n committed 11 days ago

Verified a6ea155d

ggml : check row bounds in get_rows_back (#29575)

angt committed 11 days ago

Verified d3954b93

model : support classifier_pooling for rerankers (#29627)

boshjerns committed 11 days ago

Verified 48de2a1b

hexagon: optimize concat op (#29673)

trivikram-reddy1 committed 11 days ago

Verified 7fee1784

CUDA: bitonic argsort handles rows wider than one block (#28957)

ServeurpersoCom committed 11 days ago

Verified 6a2743f0

ggml-cuda: HIP: optimize packed byte subtraction (`__vsubss4` -> `__vsub4`) (#29478)

thelittlefireman committed 11 days ago

Verified 748d4225

ci: add zdnn backend build but not test (#29541)

taronaeo committed 11 days ago

Verified cee37ffe

Older