ggml-org/llama.cpp

Pull Requests Commits

ngxson committed 24 days ago

ad778127

models: support vllm decision 2 qwen 3.5 variant

ngxson committed 24 days ago

0a2420a4

model: support vllm d3 models (#30383)

ngxson committed 24 days ago

Verified efe3626b

model: support vllm decision 2 model (#30378)

ngxson committed 24 days ago

Verified 097cf33d

tests : link each test against only the libraries it uses (#30299)

ggerganov committed 24 days ago

Verified ccdbe290

metal : fix mul_mat_id with duplicate ids (#30373)

ggerganov committed 24 days ago

Verified 073b7403

examples : fix draft sampling and rejection handling (#30095)

danbev committed 24 days ago

Verified 77757a34

llama : report compute graph reallocations during inference (#30329)

ggerganov committed 24 days ago

Verified 972b5fcc

ui: fix text attachment preview overflow (#29898)

shamnadxyz committed 24 days ago

Verified 2f160ba5

sycl: guard DMMV reads at row tails (#29338)

newjordan committed 24 days ago

Verified f9a95aeb

vulkan: use ternary expressions to avoid switch lowering for CM2 quant block size (#30164)

SG-Amadeus committed 25 days ago

Verified 08ffba21

ggml: Change default visibility of symbols to hidden (#30014)

ckastner committed 25 days ago

Verified a05e58c9

vulkan: fix buffer logging on 32-bit platforms after source split (#30326)

miyanyan committed 25 days ago

Verified b2f036c0

hexagon: optimize q2_k bpw and activation (#30035)

jhen0409 committed 25 days ago

Verified e937b17a

hexagon: add IM2COL_3D (#30305)

kurquhar committed 25 days ago

Verified b13a93a3

opencl: improve fa, allow dk512 for gemma-4, improve dk64 (#30266)

lhez committed 25 days ago

Verified f2cc7282

opencl: fix image limit for q4_k dense bin kernels, q4_0 bcast and rms_norm (#30310)

lhez committed 25 days ago

Verified c83f3058

rpc : turn GGML_RPC_DEBUG into a verbosity level and add logs (#29544)

ggerganov committed 25 days ago

Verified 5c8854fd

model : support for Prism Bonsai 2 27B (#29600)

bri-prism committed 25 days ago

Verified a9d27ac6

cpu: fix AVX2 CPUs not using F16C when available (#30270)

xinitrcn1 committed 25 days ago

Verified 1cf77677

vulkan: handle mul_mat_id duplicates in the prepass rather than looping (#29998)

jeffbolznv committed 25 days ago

Verified 966baae7

server: add component tests (#30312)

ngxson committed 25 days ago

Verified 62a6f74b

server: leave a busy slot untouched when a request pins it (#30295)

ServeurpersoCom committed 26 days ago

Verified 23b0202a

model : support MiniCPM-V 4.7 (#29416)

tc-mb committed 26 days ago

Verified 69f201a2

cmake(s390x): disable z17 target for unsupported compilers (#30297)

taronaeo committed 26 days ago

Verified abee0c84

args: add LLAMA_ARG_SLOT_SAVE_PATH env for --slot-save-path (#30272)

kucharskim committed 26 days ago

Verified aa94f208

spec: properly handle mtmd input for mtp (#30257)

ngxson committed 26 days ago

Verified 781dbc5a

mtmd: add build_inp_attn_mask (#30259)

ngxson committed 26 days ago

Verified 0fd868cb

cuda: always use MMVQ for MUL_MAT_ID on sm_60 (#27828)

dfriehs committed 27 days ago

Verified 1623d8ce

CUDA/HIP: fix race in flash_attn_ext_f16_process_tile when nbatch_combine != DKQ/2 (#30103)

IMbackK committed 27 days ago

Verified 1bb2b9fc

Older