ngxson/llama.cpp

Pull Requests Commits

rm migration script and ifdef

ngxson committed 54 days ago

4b65442c

ngxson committed 54 days ago

445b8439

Merge branch 'master' into xsn/model_def_self_contained

ngxson committed 54 days ago

24b6205e

webui: restore missing settings (#22666)

ntowle committed 55 days ago

Verified fa8feaed

docs : update speculative decoding parameters after refactor (#22397) (#22539)

ggerganov committed 55 days ago

Verified 846262d7

vulkan: delete dead GGML_VK_MAX_NODES def (#22621)

Atomic-Germ committed 55 days ago

Verified 6dcd824f

ggml-webgpu: add layer norm ops (#22406)

Constannnnnt committed 55 days ago

Verified d4b0c22f

common : determine generation prompt using longest common prefix (#22657)

aldehir committed 55 days ago

Verified e48034df

convert : Mistral format yarn apply_scale support (#22612)

juliendenize committed 55 days ago

Verified 048a490f

convert : apply Q/K RoPE permutation in NVFP4 repack path (#22611)

jmrobles committed 55 days ago

Verified db44417b

Merge branch 'master' into xsn/model_def_self_contained

ngxson committed 55 days ago

28f09f32

fix: CUDA device PCI bus ID de-dupe OOMing (ignoring other 3 gpus entirely) (#22533)

lucyknada committed 56 days ago

Verified d05fe1d7

server : avoid checkpoint data host copies (#22558)

ggerganov committed 56 days ago

Verified 0754b7b6

ggml-virtgpu: fix circular dependency in headers (#22557)

Juste-Leo2 committed 56 days ago

Verified 09294365

convert : disable uint types (#18908)

csabakecskemeti committed 57 days ago

Verified 63d93d17

opencl: Adreno optimization for MoE - MxFP4 (#22301)

shawngu-quic committed 57 days ago

Verified c5a3bc39

Github: update issue templates (#22594)

JohannesGaessler committed 57 days ago

Verified 9dbb3726

ggerganov committed 57 days ago

228e8363

ggml : bump version to 0.10.2 (ggml/1474)

ggerganov committed 57 days ago

ed23489f

ggerganov committed 57 days ago

457e2288

ggml : try fix win32 build (whisper/0)

ggerganov committed 57 days ago

e8ec7ab0

hexagon: hmx flash attention (#22347)

njsyw1997 committed 57 days ago

Verified 1a03cf47

llama-quant : fix `--tensor-type` when default `qtype` is overriden (#22572)

ddh0 committed 57 days ago

Verified b97ebdc9

hexagon: enable non-contiguous row tensor support for unary ops (#22574)

aparmp-quic committed 57 days ago

Verified 2098fd61

webui: Spring Cleaning Refactor v1 (#22505)

allozaur committed 57 days ago

Verified ab6120cd

ggml-webgpu: Fix vectorized handling in mul-mat and mul-mat-id (#22578)

yomaytk committed 57 days ago

Verified c3c15053

vulkan: Support asymmetric FA in coopmat2 path (#21753)

jeffbolznv committed 57 days ago

Verified 05e141a6

ggml-webgpu: add the upscale shader (#22419)

Constannnnnt committed 58 days ago

Verified aab68217

ggml-webgpu: Improve performance of mat-vec and mat-mat for MUL_MAT_ID (#22464)

yomaytk committed 58 days ago

Verified a95a11e5

Update llama-mmap to use ftello/fseeko (#22497)

reeselevine committed 58 days ago

Verified 5cbfb180

Older