ochafik/llama.cpp

Pull Requests Commits

app : allow --version, --licenses & --help (#25054)

angt committed 12 days ago

Verified 050ee92d

sched : reintroduce less synchronizations during split compute (#20793)

aendk committed 12 days ago

Verified 3fc4e105

devops : add llama in all docker images (#25035)

angt committed 12 days ago

Verified 5d8ccdf9

arg: fix handling --spec-draft-hf and --hf-repo-v (#25043)

ngxson committed 12 days ago

Verified 024930c6

openvino: Update to OV 2026.2.1, self-contained release packages, operator improvements (#24974)

ravi9 committed 12 days ago

Verified 5397c361

ggerganov committed 12 days ago

e7ea94af

ggml : bump version to 0.15.3 (ggml/1550)

ggerganov committed 12 days ago

96183e98

vulkan: opt mul_mat_vecq for mi50 (#22933)

chraac committed 12 days ago

Verified 487a6cc1

vulkan: add INTEL_XE1 arch enum and enable coopmat1 on Intel Xe-LPG Plus (#24404)

fish-jiang committed 12 days ago

Verified 5a6a0dd7

ui: fix accessibility for hover-gated interactive elements assisted by claude(in debugging and tests) (#24727)

sanjayahari committed 12 days ago

Verified ded1561b

vulkan: Workaround compiler bug in conv2d coopmat2 path (#24924)

jeffbolznv committed 12 days ago

Verified 9df06805

CUDA: add cublasSgemmBatched mapping for HIP/MUSA vendor headers (#25033)

leonardHONG committed 12 days ago

Verified 2f18fe13

ggml-cpu: fix SVE leftover path in ggml_vec_dot_f32 (#24699)

tdakhran committed 12 days ago

Verified c16c35b8

server + ui: SSE Replay Buffer (#23226)

ServeurpersoCom committed 12 days ago

Verified 1a87dcdc

sycl : clamp softmax input to avoid underflow (#24941)

Jassieluo committed 12 days ago

Verified e7e3f350

mtmd: add more validations (#25013)

ngxson committed 12 days ago

Verified b11f7c16

CUDA: batch out_prod broadcast (dps2>1) path with cublasSgemmBatched (#24426)

leonardHONG committed 12 days ago

Verified f818065d

mamba2: remove hardcoded 2x expansion factor and invalid d_inner % d_state check (#23082)

limloop committed 12 days ago

Verified 960d628f

opencl: flush profiling batch at shutdown for incomplete batches (#25016)

shaofeiqi committed 12 days ago

Verified 5c7c22c3

xcframework : disable mtmd video on i/tv/visionos (#25018)

CISC committed 13 days ago

Verified beac5309

model : Add label for LFM2.5-230M (#25008)

tdakhran committed 13 days ago

Verified 9d5d882d

CUDA: Various fixes to `cpy.cu` (#25000)

ORippler committed 13 days ago

Verified 1ec44d17

misc: fix labeler (#25012)

ngxson committed 13 days ago

Verified c7cddefc

server: use status code 403 for disabled features (#24970)

ngxson committed 13 days ago

Verified e9d1b76d

misc: update lables (#24920)

ngxson committed 13 days ago

Verified 099bf069

common: refactor model handling (#24980)

ngxson committed 13 days ago

Verified 60bc8866

docs : Eagle3 qwen3 draft model support (#24977)

kashif committed 13 days ago

Verified e8ecce53

app : add the llama download subcommand (#24982)

angt committed 13 days ago

Verified 683b04cc

ggml : address integer overflows in binary ops CUDA implementation (#24706)

fairydreaming committed 13 days ago

Verified f728adab

ui: fix always-show-sidebar-on-desktop setting after navigation refactor (#24979)

ServeurpersoCom committed 13 days ago

Verified 3e61ea0e

Older