ggml-org/llama.cpp

Pull Requests Commits

delta_net: fill the conv slot index on the host

ServeurpersoCom committed 2 days ago

10364f65

am17an committed 10 days ago

d9958cc2

delta_net: access conv state via views

am17an committed 10 days ago

2e7a5fb7

qwen: merge tensors on convert (#31028)

am17an committed 10 days ago

Verified 2280f0fd

ci : disable test-llama-context with fa off in the webgpu (#30977)

yomaytk committed 10 days ago

Verified 1a0b56ce

llama: fix exhausting CPU and MEM caused by dry sequence breaker (#31035)

trunk-trick committed 10 days ago

Verified 63412d87

llama-quant : fix 4D tensor quantization using ggml_nrows() and name-based expert detection (#30820)

anthony-wss committed 10 days ago

Verified 5c965fd8

grammar: limit recursion depth to prevent stack overflow (#29495) (#29755)

Hrushikesh-ramilla committed 10 days ago

Verified f1168ac4

llama : respect no_alloc in all memory caches (#31087)

ggerganov committed 10 days ago

Verified fa3645fe

tools : reject truncated GGUF tensor data in merge (#31044)

shreya-11733 committed 10 days ago

Verified 7790a923

CUDA: fuse FFN gate/up matmuls and GLU in MMQ with interleaved gate/up rows (#29948)

anujj committed 10 days ago

Verified c49edf09

ggml-cuda: fuse qwen MRoPE + set-rows (#31025)

am17an committed 10 days ago

Verified a55c16e4

vulkan: Add a larger tile size for conv3d (coopmat2 only) (#29920)

jeffbolznv committed 10 days ago

Verified e321bdad

vulkan : one warp tile per subgroup in the small matmul tiles (#30106)

ianloic committed 10 days ago

Verified 459a1ba5

hexagon: generic dims collapse/reshape for unary, binary and matmul ops (#31071)

max-krasnyansky committed 10 days ago

Verified f0fb5984

opencl: add topk and lightning indexer (#31026)

lhez committed 10 days ago

Verified 9a4580b6

opencl: add initial q2_k support (#30999)

lhez committed 10 days ago

Verified b083c384

ci : update CPU job assignments (#31136)

ggerganov committed 10 days ago

Verified 344be0a6

chat : reject JSON nested more than 128 levels before walking it (#30242)

ykhrustalev committed 10 days ago

Verified b3fa6995

model: DSV4.1 flash (#31086)

am17an committed 10 days ago

Verified 74bca395

jinja : fix select and selectattr argument handling (#31059)

JoeyPatricio committed 11 days ago

Verified d327bf1f

hexagon: KV-head team split for multicore HMX flash attention (#31055)

ebateni committed 11 days ago

Verified 78a8ae14

server: refactor server entrypoint (#31113)

ngxson committed 11 days ago

Verified f3269ef9

metal : remove mul_mm_id amax rescaling (#30924)

ggerganov committed 11 days ago

Verified 3f03b177

rpc : add bounds check in rdma_recv (#30399)

sohaibharraoui committed 11 days ago

Verified 9a0fde99

ui: Models Manager as Route + UI improvements (#30988)

allozaur committed 11 days ago

Verified e168e041

fix(ui): align hover card and context popup styling with dropdown menu (#30978)

allozaur committed 11 days ago

Verified ea1d7d66

ci : pass per-runner results dir to self-hosted jobs on T4 runners (#31100)

ggerganov committed 11 days ago

Verified 71255fef

minimax-m3 : forward quant scales to the graph (fixes #31063) (#31097)

samgsx committed 11 days ago

Verified 73adf89e

tests : fix test-llama-context for large n_batch (#31083)

sanmai committed 11 days ago

Verified df9beac7

Older