ggml-org/llama.cpp

Pull Requests Commits

ui: fix system message row stretching to full prompt height in WebKit

allozaur committed 21 hours ago

4fa35c30

glm5-next : QSA without kq_mask (+ DSA) (#31009)

sanmai committed 22 hours ago

Verified d78f40d9

ci : use ccache depend mode for ROCm release builds (#31231)

sreeram-11 committed 23 hours ago

Verified fae9ac56

ci : force python 3.12 on ubuntu 26.04 (#31222)

CISC committed 23 hours ago

Verified a8276b70

ggml : relax scale to non-contiguous views (#31146)

ggerganov committed 1 day ago

Verified de5122d8

tests : fix multi-seq split replay and partial-state cases for trained models (#31216)

ggerganov committed 1 day ago

Verified 6666a733

llama : remove obsolete n_pad from memory instances (#31200)

ggerganov committed 1 day ago

Verified 7a2e5aba

ui: list org-less models first in the model selector (#31215)

allozaur committed 1 day ago

Verified 0d5ca600

opencl: refine mul mat id -1 index (#31170)

lhez committed 1 day ago

Verified d930fe3c

opencl: initial support for q3_k (#31172)

lhez committed 1 day ago

Verified a2b3e121

ui: Highlight reasoning toggle on hover and align it with model selector (#31212)

allozaur committed 1 day ago

Verified 823df7d0

ui: add conversation compaction (#27401)

ServeurpersoCom committed 1 day ago

Verified 1a97e387

ui: Open model details in a drawer without leaving the page (#31176)

allozaur committed 1 day ago

Verified bc262f33

ui: use shared confirmation dialog for conversation deletion (#31187)

allozaur committed 1 day ago

Verified f5892ca5

hexagon: add GGML_HEXAGON_VTCM_SIZE (#31221)

ebateni committed 1 day ago

Verified 7aca0d2e

llama : enable moe amax rescaling for laguna (#31147)

begcdn committed 1 day ago

Verified 961c376a

ci : add windows behavior check (#31191)

CISC committed 1 day ago

Verified 4244ad7a

opencl: write KV-cache views in set_tensor so a state restore is not dropped (#30471)

wanghqc committed 1 day ago

Verified 63b4b303

hexagon: MUL_MAT_ID row skipping optimizations and improvements (#31169)

max-krasnyansky committed 1 day ago

Verified caa8191d

llama : move no_alloc out of llama_hparams (#31141)

ggerganov committed 1 day ago

Verified a4699400

models : replace ggml_tensor** out-params with output structs (#31142)

ggerganov committed 1 day ago

Verified 4b851fbe

ggml-cpu : rewrite concat as one row kernel for all types (#31179)

ServeurpersoCom committed 1 day ago

Verified 4d71b42f

vulkan: rms norm optimization using subgroup reductions (#29882)

virajwad committed 1 day ago

Verified 70c5f263

sycl: template the plain MMVQ launch triples (#31013)

Titaniumtown committed 1 day ago

Verified 5cd62ffe

sycl: template the per-quant concat impls (#31008)

Titaniumtown committed 1 day ago

Verified 38645c2a

sycl: add set_tensor_2d for faster tensor-split loading (#30358)

Captain-Tripps committed 1 day ago

Verified 02269b4c

sycl : bound reorder temp memory to 32 MiB (#31062)

KyleHagy committed 1 day ago

Verified 5320c87f

common: pin model resolution on real HF repo layouts (#30548)

ServeurpersoCom committed 1 day ago

Verified 74b104c9

mtmd: add mtmd_helper_get_capabilities (#31183)

ngxson committed 1 day ago

Verified c6fb3b5b

ggml-cpu : fix unused variables in arm arch (#31180)

CISC committed 1 day ago

Verified 8b28c239

Older