ggml-org/llama.cpp

Pull Requests Commits

ngxson committed 10 hours ago

8f637a66

common: load adapters before context creation

ngxson committed 10 hours ago

f20b386a

rm redundant header

ngxson committed 1 day ago

771a4062

ngxson committed 1 day ago

11c4867c

keep track n_lora_nodes from llama_model

ngxson committed 1 day ago

d66c5cd4

ngxson committed 1 day ago

ac6392db

3 nodes per weight

ngxson committed 1 day ago

fe2f7fc7

lora: count lora nodes in graph_max_nodes

ngxson committed 1 day ago

e5f1e38b

webui: fix prompt progress ETA calculation (#18468)

ngxson committed 1 day ago

Verified 51a48720

Webui/prompt processing progress (#18300)

ServeurpersoCom committed 1 day ago

Verified c9a3b40d

CUDA: fix replacment of bad archs in CMake (#18457)

JohannesGaessler committed 1 day ago

Verified 0bd1212a

server : Cmdline arg -to changes http read timeout from current 600sec default (#18279)

wbtek committed 1 day ago

Verified 5b1248c9

contributing: tighten AI usage policy (#18388)

ngxson committed 1 day ago

Verified 3595ae59

android: routine maintenance - Dec 2025 (#18338)

naco-siren committed 1 day ago

Verified c1366056

server : handle closed connection for tasks (#18459)

ggerganov committed 1 day ago

Verified 2a85f720

model-conversion : add device option to embd run orig model (#18386)

danbev committed 1 day ago

Verified 7cbec34a

retrieval : use at most n_seq_max chunks (#18400)

hectorem2 committed 1 day ago

Verified 0c898640

common: fix return value check for setpriority (#18412)

o7si committed 1 day ago

Verified daa242df

CUDA: Blackwell features for non-native builds (#18436)

JohannesGaessler committed 1 day ago

Verified e70e640d

cuda: fix race condition in cumsum (#18448)

am17an committed 1 day ago

Verified 5fa66c6e

ci : re-enable rocm build on amd64 (#18439)

TimNN committed 2 days ago

Verified 382808c1

HIP: Use mmq on MFMA devices for MUL_MAT_ID in cases where a lot of splits would be generated (#18202)

IMbackK committed 2 days ago

Verified 4ffc47cb

model : Plamo3 support (#17304)

mmnga committed 2 days ago

Verified 9c675c71

Revert "ggml-cuda: use CMAKE_CUDA_ARCHITECTURES if set when GGML_NATIVE=ON (#18413)" (#18426)

am17an committed 2 days ago

Verified 07a0c4ba

rpc: fix segfault on invalid endpoint format (#18387)

o7si committed 2 days ago

Verified 60f17f56

llama-fit-params: fix step size for last device (#18415)

JohannesGaessler committed 2 days ago

Verified f8d561eb

github: update issue templates [no ci] (#18410)

JohannesGaessler committed 2 days ago

Verified e59efe6a

mtmd: clarify that we no longer accept AI-generated PRs (#18406)

ngxson committed 2 days ago

Verified cffa5c46

cmake: Added more x86_64 CPU backends when building with `GGML_CPU_ALL_VARIANTS=On` (#18186)

bberberov committed 2 days ago

Verified 94de74e7

ggml-cuda: use CMAKE_CUDA_ARCHITECTURES if set when GGML_NATIVE=ON (#18413)

QDelta committed 2 days ago

Verified 4fd59e84

Older