ggerganov/llama.cpp

Pull Requests Commits

enroll glm5.3-next too

am17an committed 2 days ago

8a862123

llama: create new fused op for QSA scoring

am17an committed 2 days ago

d0c83b42

graph : add precision handling for MELLUM architecture (#30584)

Xarbirus committed 2 days ago

Verified 5e03a1ff

hexagon: tensor repacking rework and support for tiled and flat quantized and fp weights (#30635)

max-krasnyansky committed 2 days ago

Verified b797b9ca

llama : share the k-pool input views between layers (#30613)

am17an committed 2 days ago

Verified 6afa9a31

Limit max CUDA pinned host buffer size to 4 GB (#30449)

gaugarg-nv committed 2 days ago

Verified b6334e37

sycl: support reordered Q8_0 MoE weights (#29337)

newjordan committed 2 days ago

Verified e707790c

[SYCL] Enable ssl in build script and CI (#30456)

arthw committed 2 days ago

Verified 43d57fb8

mtmd: add mtmd_audio_stream + support voxtral-realtime (#30629)

ngxson committed 2 days ago

Verified 2851b3bc

convert : fix Granite Switch block_count for MultiSwitch checkpoints (#30621)

planetf1 committed 2 days ago

Verified 59d5fad7

convert : fix regression in mapping of Granite speech tensors (#30628)

mrutkows committed 2 days ago

Verified e287bd87

mtmd: lightonocr does not require longest edge to always be 1540, native resolution is the default (#30600)

staghado committed 2 days ago

Verified 2f729cfa

model: add Muse Spark Open (text, vision and audio) (#30616)

ServeurpersoCom committed 2 days ago

Verified ba3c05f7

pi : expand AI disclosure with hardware and stats (#30615)

ggerganov committed 2 days ago

Verified b76b1337

llama.cpp : bump version to 0.7.0 (#30603)

ggerganov committed 3 days ago

Verified 620d2225

models : add Nemotron Super 3.5 VL and Nemotron Bolt 74B A9B support (#30520)

frozenblade1224 committed 3 days ago

Verified cd0a5b6a

scripts : add ci-status.sh to list active CI runs (#30586)

ggerganov committed 3 days ago

Verified a11b6530

server: fix the off-by-two in the SWA reuse threshold (#30580)

ServeurpersoCom committed 3 days ago

Verified 61ad80bf

model : Mellum - read MoE gating params from GGUF, unify RoPE (#30578)

Xarbirus committed 3 days ago

Verified a6b91d8c

server: save the draft state in the slot file (#30574)

ServeurpersoCom committed 3 days ago

Verified ea736c1d

server: prefer empty slots and keep states overwritten by child tasks (#30567)

ServeurpersoCom committed 3 days ago

Verified b0d23916

ggerganov committed 3 days ago

c0cb53ac

ggml : bump version to 0.27.0

ggerganov committed 3 days ago

4ed1c850

ci : bump ty to 0.0.88 (#30587)

CISC committed 3 days ago

Verified 711780ff

fix(ui): take the agentic turns default from DEFAULT_AGENTIC_CONFIG (#30582)

allozaur committed 3 days ago

Verified fd55cb98

models : add maion-coder (#29778)

mailkv23 committed 3 days ago

Verified 3e4dfa8f

jinja : improve context copy constructor performance (#30562)

Ninja3047 committed 3 days ago

Verified 08698c81

jinja : write nan and inf in tojson as json.dumps does (#30483)

ykhrustalev committed 3 days ago

Verified 82e1e2fd

vulkan : fix concurrent first device init (#29766)

losewayy committed 3 days ago

Verified 47f5a64a

sycl : replace oneAPI package by DLE+oneDNN package (#30131)

arthw committed 3 days ago

Verified e36df604

Older