ggerganov/llama.cpp

Pull Requests Commits

llama : enable layer input extraction

ggerganov committed 63 days ago

Verified 0445829c

Hexagon: Process M-tail rows on HMX instead of HVX (#22724)

trivikram-reddy1 committed 63 days ago

Verified bbeb89d7

opencl: refactor Adreno q4_0 (#22335)

lhez committed 63 days ago

Verified ff806a11

rpc : use graph uid instead of graph cache (#22701)

rgerganov committed 64 days ago

Verified d5003b6e

common : fix missing-noreturn warnings when compiling with clang 21 (#22702)

angt committed 64 days ago

Verified 2635ac76

ggerganov committed 64 days ago

70a83091

ggml : bump version to 0.11.0 (ggml/1478)

ggerganov committed 64 days ago

c91faf99

common : only load backends when required (#22290)

angt committed 64 days ago

Verified bf76ac77

vendor : update cpp-httplib to 0.43.3 (#22686)

cabelo committed 64 days ago

Verified a09a00e5

server : validate --tools CLI argument against known tool names (#22538)

ggerganov committed 64 days ago

Verified 2bacb1eb

llama : add option to save memory in device buffers (#22679)

ggerganov committed 64 days ago

Verified d6e7b033

graph : handle non-contiguous Q/K/V in mul_mat_aux (#22630)

CISC committed 64 days ago

Verified fa595462

ggml : implement fast walsh-hadamard transform for kv rotation (#21352) (#22631)

AlrIsmail committed 64 days ago

Verified a817a22b

kleidiai : update to v1.24.0 and use release archive (#22549)

chaxu01 committed 64 days ago

Verified eff06702

CUDA: use fastdiv for batch index split in get_rows (#22650)

leonardHONG committed 64 days ago

Verified e77056f9

server: implement /models?reload=1 (#21848)

ngxson committed 64 days ago

Verified 935a3402

examples: refactor diffusion generation (#22590)

Sailaukan committed 64 days ago

Verified d8794eec

webui : fix circular dependency between chat.service.ts and models.svelte.ts (#22625)

Juste-Leo2 committed 65 days ago

Verified 36a694c9

common/autoparser: fixes for newline handling / forced tool calls (#22654)

pwilkin committed 65 days ago

Verified a4701c98

model: move `load_hparams` and `load_tensors` to per-model definition (#22004)

ngxson committed 65 days ago

Verified 994118a1

server: Add a simple get_datetime server tool (#22649)

eapache committed 65 days ago

Verified c84e6d6d

webui: restore missing settings (#22666)

ntowle committed 65 days ago

Verified fa8feaed

docs : update speculative decoding parameters after refactor (#22397) (#22539)

ggerganov committed 65 days ago

Verified 846262d7

vulkan: delete dead GGML_VK_MAX_NODES def (#22621)

Atomic-Germ committed 65 days ago

Verified 6dcd824f

ggml-webgpu: add layer norm ops (#22406)

Constannnnnt committed 65 days ago

Verified d4b0c22f

common : determine generation prompt using longest common prefix (#22657)

aldehir committed 65 days ago

Verified e48034df

convert : Mistral format yarn apply_scale support (#22612)

juliendenize committed 65 days ago

Verified 048a490f

convert : apply Q/K RoPE permutation in NVFP4 repack path (#22611)

jmrobles committed 65 days ago

Verified db44417b

fix: CUDA device PCI bus ID de-dupe OOMing (ignoring other 3 gpus entirely) (#22533)

lucyknada committed 66 days ago

Verified d05fe1d7

server : avoid checkpoint data host copies (#22558)

ggerganov committed 66 days ago

Verified 0754b7b6

Older