ggml-org/llama.cpp

Pull Requests Commits

ui : escape sidecar tokens in the regex alternation

allozaur committed 9 days ago

383408a3

ui : model id grammar for sidecars, quants and capability parsing

allozaur committed 9 days ago

b92431df

Update tools/ui/src/lib/components/app/chat/index.ts

allozaur committed 9 days ago

856b1f74

ui : document the model list index pairing, fix an em-dash

allozaur committed 9 days ago

14d78a13

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

allozaur committed 9 days ago

02dc85db

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

ServeurpersoCom committed 9 days ago

Verified 25747b08

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

ServeurpersoCom committed 9 days ago

Verified db00347a

musa : define __CUDA_ARCH__ for device passes (#29508)

yeahdongcn committed 9 days ago

Verified 272aad8b

ci : add models backend check (#29651)

CISC committed 9 days ago

Verified 72db1e02

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

Captain-Tripps committed 9 days ago

Verified 2a53ace3

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

timkhronos committed 9 days ago

Verified 649dcb10

vendor: update BoringSSL to 0.20260929.0 (#29669)

cabelo committed 9 days ago

Verified 931351ea

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

taronaeo committed 9 days ago

Verified eae11d22

Hexagon f16 activation ops (#29209)

cqderek committed 10 days ago

Verified 19e28a27

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

x14ngch3n committed 10 days ago

Verified a6ea155d

ggml : check row bounds in get_rows_back (#29575)

angt committed 10 days ago

Verified d3954b93

model : support classifier_pooling for rerankers (#29627)

boshjerns committed 10 days ago

Verified 48de2a1b

hexagon: optimize concat op (#29673)

trivikram-reddy1 committed 10 days ago

Verified 7fee1784

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

ServeurpersoCom committed 10 days ago

Verified 6a2743f0

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

thelittlefireman committed 10 days ago

Verified 748d4225

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

taronaeo committed 10 days ago

Verified cee37ffe

opencl: fix get_tensor for q5_K adreno gemm_nonshuffle kernel (#29555)

lhez committed 10 days ago

Verified 6dbbac44

vulkan: Tune GDN kernel, fix Intel performance (#29476)

0cc4m committed 10 days ago

Verified 5c200e0c

vulkan : Load F32 A matrix 2 at a time when its 2-aligned (#29254)

TheBlueMatt committed 10 days ago

Verified 83dd71f8

vulkan: MOE aware mat_mul_id tile selection (#29182)

Ankk98 committed 10 days ago

Verified 94a0ae3e

ggml : fix c++ odr by properly using GGML_COMMON_DECL_CPP (#29504)

XZiar committed 10 days ago

Verified da89bb3c

vocab : keep </s> NORMAL in PLaMo-2 and PLaMo-3 (#29580)

tokinasin committed 10 days ago

Verified a3f84faf

ggml : accumulate f16 dot products in f32 on AVX512-FP16 (#29545)

angt committed 10 days ago

Verified 284153e0

ggml : require input tensors to be GGML_OP_NONE (#29647)

ggerganov committed 10 days ago

Verified b5cf8ce0

hexagon: add FP32 GELU_ERF and GEGLU_ERF support (#29631)

aparmp-quic committed 10 days ago

Verified e904318a

Older