ggml-org/llama.cpp

Pull Requests Commits

fix windows build

ngxson committed 12 days ago

1b2bd869

ngxson committed 12 days ago

92332718

if AIP_MODE is unset, do nothing

ngxson committed 12 days ago

331e4d21

ngxson committed 12 days ago

348e6088

support other AIP_* env var

ngxson committed 12 days ago

5e11eafc

ngxson committed 12 days ago

5dd6c9e5

Merge branch 'master' into xsn/vertexai

ngxson committed 12 days ago

d34f9713

server: support Vertex AI compatible API

ngxson committed 12 days ago

bfc135fe

scripts : add wc2wt.sh - create worktree from current HEAD (#22513)

ggerganov committed 12 days ago

Verified 27aef3dd

add fast matmul iquants (#22504)

SharmaRithik committed 12 days ago

Verified 45155597

spec : fix draft model checkpoints (#22521)

ggerganov committed 12 days ago

Verified 80afa33a

spec : fix vocab compat checks in spec example (#22426)

petersid2022 committed 12 days ago

Verified b42c7fa5

common : do not pass prompt tokens to reasoning budget sampler (#22488)

aldehir committed 13 days ago

Verified d7759923

hexagon: make vmem and buffer-size configurable (#22487)

max-krasnyansky committed 13 days ago

Verified 41a63be2

CUDA: fuse SSM_CONV + ADD(bias) + SILU (#22478)

anavp-nvidia committed 13 days ago

Verified 098705a2

spec : disacard last drafted token with low prob (#22506)

ggerganov committed 13 days ago

Verified 683c5acb

ggerganov committed 13 days ago

b1d5f5b4

ggml : bump version to 0.10.1 (ggml/1469)

ggerganov committed 13 days ago

4b221b7f

webui: fix slow mic stop and WAV encode (#22480)

ServeurpersoCom committed 13 days ago

Verified 59237bfb

ggml-cpu : disable tiled matmul on AIX to fix page boundary segfault (#22293)

shalinib-ibm committed 13 days ago

Verified 1cbc846e

ggml-cuda: refactor fusion code (#22468)

am17an committed 13 days ago

Verified 3142f1db

ggml-cpu: cmake: append xsmtvdotii march for SpacemiT IME (#22317)

qiurui144 committed 13 days ago

Verified b5c4227d

ggml-webgpu: Fix bug in FlashAttention support check (#22492)

reeselevine committed 13 days ago

Verified d6a50940

common: Intentionally leak logger instance to fix hanging on Windows (#22273)

rillomas committed 13 days ago

Verified 7b95ea5d

ggml : add sve tuned code for gemm_q8_0_4x8_q8_0() kernel (#21916)

hrushitfujitsu committed 13 days ago

Verified bdc9c743

TP: fix delayed AllReduce + zero-sized slices (#22489)

JohannesGaessler committed 13 days ago

Verified 739393be

ggml-cuda: Repost of 21896: Blackwell native NVFP4 support (#22196)

michaelw9999 committed 13 days ago

Verified fc2b0053

ggml-cuda: add flash-attn support for DKQ=320/DV=256 with ncols2=32 (… (#22286)

lnigam committed 14 days ago

Verified 7b8443ac

convert : add support for Nemotron Nano 3 Omni (#22481)

danbev committed 14 days ago

Verified 5d56effd

common : re-arm reasoning budget after DONE on new <think> (#22323)

BruceJillis committed 14 days ago

Verified 52e5f0a5

Older