ggml-org/llama.cpp

Pull Requests Commits

server: simplify get_info result handling

ServeurpersoCom committed 97 days ago

3dfb7c7f

server: harden get_info probe result handling

ServeurpersoCom committed 97 days ago

c74747ce

fix --rpc in docs

ngxson committed 97 days ago

5f0af44f

server: add get_info tool

ngxson committed 97 days ago

371e8b0a

ggml: use dynamic allocation for split graph inputs (#22789)

AgoraPete committed 97 days ago

Verified dbadb68e

opencl: route large q6_K lm_head to the flat GEMV (#26427)

wanghqc committed 97 days ago

Verified 39eab74a

graph : fix unused input tensors in minimax m3 graph (#26519)

ggerganov committed 97 days ago

Verified c50b34a1

model: M3: Move MSA into a new memory implementation (#26338)

timkhronos committed 97 days ago

Verified 67d5978b

llama : allocate indexer cache only in "full" indexer layers (#26474)

fairydreaming committed 97 days ago

Verified 563dec81

CUDA: Add backend sampler for penalties sampler (#25262)

kmorennv committed 97 days ago

Verified 96278e39

CUDA: Fix data-races when reusing SMEM in block_reduce (#26385)

ORippler committed 97 days ago

Verified 9bd4c09e

server: add notice for upcoming default port change 8080 --> 9931 (#26508)

ngxson committed 97 days ago

Verified 0b14b87d

server: (tools) add x-tool-cwd header (#26420)

ngxson committed 97 days ago

Verified f2b52a87

model: MTP support for Qwen3-Next (#25589)

yomaytk committed 97 days ago

Verified 4ed2b13f

llama : MTP support for DeepSeek V3.2 (#26457)

fairydreaming committed 97 days ago

Verified 2b63e061

metal: implement DSv4 Lightning Indexer (#25893)

tarruda committed 97 days ago

Verified 1464c62d

metal : add SILU_BACK (#25982)

Blackcyan30 committed 97 days ago

Verified 221f0f63

metal : add F16 support for bin ops (#26465)

ggerganov committed 97 days ago

Verified 9d21b57f

opencl: limit local workgroup size for GLU operation (#26383)

mgroeber9110 committed 97 days ago

Verified 0ab9d6fe

metal: implement DeepSeek V4 hyper-connections (#26459)

ggerganov committed 97 days ago

Verified fffbcbdb

common: support the DSpark sidecar resolution (#26458)

ServeurpersoCom committed 97 days ago

Verified bb4e0e1b

convert: add option to create separate dspark GGUF (#26452)

am17an committed 98 days ago

Verified 3581ba0c

opencl: bugfix increment ref_count in ggml_backend_opencl_init() (#26162)

akleine committed 98 days ago

Verified c745be2a

DeepseekV4 MTP + DSpark (#25784)

am17an committed 98 days ago

Verified 596a5795

chat : add qwen3 specialized parser (#26252)

aldehir committed 98 days ago

Verified f5919bf4

sycl: fix classification of iGPUs (#26105)

KyleHagy committed 98 days ago

Verified 272700b3

model : load MiMo V2 MTP tensors only if used (#26412)

CISC committed 98 days ago

Verified 75587a05

ggml-webgpu: add support for f16 repeat (#26307)

yomaytk committed 98 days ago

Verified 7a2db1a0

test: fix some CI errors (#26415)

ngxson committed 98 days ago

Verified 11924d4c

vulkan: extend topk_moe fusion to support sqrt(softplus) (#26124)

jeffbolznv committed 98 days ago

Verified a7a6d0d2

Older