ggml-org/llama.cpp

Pull Requests Commits

mtmd: add n_embd_head

ngxson committed 36 days ago

10329a2a

mtmd: add lanczos resize method [no release] (#26341)

ngxson committed 36 days ago

Verified 5f55650a

server: support inp embd to generate next token (#26313)

ngxson committed 36 days ago

Verified b4ca032a

vulkan: Support quantized concat (#25684)

jeffbolznv committed 36 days ago

Verified ea63b4d3

Test support for alternative conv layout (#25617)

pmaybank committed 37 days ago

Verified 958d9c0b

llama-context : sync pending async copies before clearing embd_seq (#25676)

o7si committed 37 days ago

Verified 432d7ffe

tests : avoid building get-model.cpp many times (#26317)

ggerganov committed 37 days ago

Verified 47f686f5

ggml-cuda: Allow transpose-free gemmv computation (#26171)

roberteg16 committed 37 days ago

Verified e1a1abb7

readme : refresh (#26280)

ggerganov committed 37 days ago

Verified 6b36c230

ggerganov committed 37 days ago

9ebfc3a8

ggml : bump version to 0.18.0 (ggml/1576)

ggerganov committed 37 days ago

6a4c3357

CUDA: add Q2_0 support (#25707)

khosravipasha committed 37 days ago

Verified 9b2a0888

Remove custom cpu op from the M3 graph, express with stock ops (#26297)

timkhronos committed 37 days ago

Verified b2f22168

metal: fix memory unwire if model is freed without any GPU operations (#26082)

nikwen committed 37 days ago

Verified d0bfb198

ui: IndexedDB and Conversations data fixes (#26278)

allozaur committed 37 days ago

Verified 21a5f5b7

ggml : Fix issue with kleidiai ci and stringop overflow warning (#26277)

JonathanC-ARM committed 37 days ago

Verified 32703b42

[UT] enhance UT to show all real unsupported backends (#25234)

arthw committed 37 days ago

Verified a6a77bc4

mimo2: address MTP review feedback (#26228)

tnhnyzc committed 37 days ago

Verified 64d528be

fix: increase greeting spacing on md screens (#26287)

allozaur committed 38 days ago

Verified 3018a11e

llama: move suppress_tokens handling to common/sampling (#26276)

ngxson committed 38 days ago

Verified afeebe10

ggml-cuda : disable MMQ on devices with less than 48 KiB shared memory (#26141)

KakaruHayate committed 38 days ago

Verified caa596ab

sycl: contiguous fast path + 32-bit index math for unary elementwise ops (#25946)

Titaniumtown committed 38 days ago

Verified 11b068d0

vendor: update BoringSSL to 0.20260728.0 (#26241)

cabelo committed 38 days ago

Verified e2f59ed7

server : add trace logging for slot similarity checking (#26271)

ggerganov committed 38 days ago

Verified 992c3253

conversion: fix Qwen2.5-Omni mmproj conversion regression (#26262)

kaben committed 38 days ago

Verified e1af89a6

RPC: add tensor_memset (#25912)

am17an committed 38 days ago

Verified f5b9bd39

add rdna3.5, and 3 to mmq configs so they can be tuned independently. (#26199)

Geramy committed 38 days ago

Verified 60bccc37

model: add NextN/MTP speculative decoding support for GLM_DSA (GLM-5.2) (#25980)

satindergrewal committed 38 days ago

Verified 7be2c65d

model: Add Laguna-S-2.1 LLM_TYPE (#26233)

crusaderky committed 38 days ago

Verified e9fa0781

ggml-webgpu: Fix some binding alias issues to support all archs, fix recurrent-state-rollback test (#25931)

reeselevine committed 38 days ago

Verified bc71c24c

Older