ngxson/llama.cpp

Pull Requests Commits

opencl: fix couple crashes (#12795)

Henry Linjamäki committed 283 days ago

Verified edbf42ed

releases : build CPU backend separately (windows) (#13642)

slaren committed 283 days ago

Verified d643bb2c

hparams : support models for which all layers use SWA (#13682)

ggerganov committed 283 days ago

Verified 8e186ef0

server : improve error reporting (#13680)

ggerganov committed 283 days ago

Verified 5fbfe384

convert : add qwen2vl support for unsloth merges (#13686)

antichristHater committed 283 days ago

Verified c76532e7

examples : switch retrieval to llama_encode (#13685)

CISC committed 283 days ago

Verified 2aa777d8

gguf-py : display the invalid gguf type (#13687)

emmanuel-ferdman committed 283 days ago

Verified eb0f5c28

ggml : add ggml_gelu_erf() (#13667)

ngxson committed 283 days ago

Verified cf4cb59e

server : Add the endpoints /api/tags and /api/chat (#13659)

R-Dson committed 284 days ago

Verified 0d5c7421

server : fix first message identification (#13634)

doringeman committed 284 days ago

Verified 42158ae2

kv-cache : simplify the interface (#13660)

ggerganov committed 284 days ago

Verified 797f2ac0

model : disable SWA for Phi models (#13676)

ggerganov committed 284 days ago

Verified b44890df

musa: Upgrade MUSA SDK version to rc4.0.1 and use mudnn::Unary::IDENTITY op to accelerate D2D memory copy (#13647)

yeahdongcn committed 284 days ago

Verified 33983057

vulkan: fix warnings (#13626)

netrunnereve committed 284 days ago

Verified fb1cab20

mtmd-helper : bug fix to token batching in mtmd (#13650)

l3utterfly committed 284 days ago

Verified b7a17463

model : fix llama4 graph (#13663)

ggerganov committed 284 days ago

Verified be023969

llama : remove llama_kv_cache_view API + remove deprecated (#13653)

ggerganov committed 285 days ago

Verified a4090d11

CUDA: skip fully masked-out KV in FA vec kernel (#13584)

JohannesGaessler committed 285 days ago

Verified b69f1647

tests : avoid github urls due to throttling (#13654)

CISC committed 285 days ago

Verified 759e37b0

sycl: disable reorder for sycl mulmat (#13536)

sgeor255 committed 285 days ago

Verified 4245e622

Set GLM4 blk.*.attn_output.weight, kqv_out-* matmul to GGML_PREC_F32 to fix infinity values in output (#13639)

0cc4m committed 285 days ago

Verified c9c64dee

metal : fix typo in FA kernel comments (#13651)

ggerganov committed 285 days ago

Verified c00a2634

kv-cache : add SWA support (#13194)

ggerganov committed 285 days ago

Verified e298d2fb

CANN: Update CANN model support (#13162)

bachelor-dou committed 285 days ago

Verified f0adb80b

sycl : Overcoming workaround for mmap() allocation on Windows (#13482)

s-Nick committed 285 days ago

Verified f7c9429c

common : add load_progress_callback (#13617)

psocolovsky committed 285 days ago

Verified 1dfbf2cf

Vulkan: Add f32 accumulator support to quantized mul mat to fix GLM4 32B incoherence (#13607)

0cc4m committed 285 days ago

Verified 8960efd0

sycl : backend documentation review (#13544)

Alberto Cabrera Pérez committed 286 days ago

Verified 725f23f1

mtmd : add vision support for llama 4 (#13282)

ngxson committed 286 days ago

Verified 92ecdcc0

ci : upgraded oneAPI version in SYCL workflows and dockerfile (#13532)

Alberto Cabrera Pérez committed 286 days ago

Verified f71f40a2

Newer Older