ggerganov/llama.cpp

Pull Requests Commits

ui : ignore search dialog keys while an IME is composing

allozaur committed 5 hours ago

ddf2eb49

ui : replace the sidebar search mode with a search dialog

allozaur committed 5 hours ago

210d011b

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

allozaur committed 5 hours ago

Verified 0d5ca600

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

lhez committed 5 hours ago

Verified d930fe3c

opencl: initial support for q3_k (#31172)

lhez committed 5 hours ago

Verified a2b3e121

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

allozaur committed 5 hours ago

Verified 823df7d0

ui: add conversation compaction (#27401)

ServeurpersoCom committed 5 hours ago

Verified 1a97e387

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

allozaur committed 6 hours ago

Verified bc262f33

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

allozaur committed 6 hours ago

Verified f5892ca5

hexagon: add GGML_HEXAGON_VTCM_SIZE (#31221)

ebateni committed 11 hours ago

Verified 7aca0d2e

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

begcdn committed 13 hours ago

Verified 961c376a

ci : add windows behavior check (#31191)

CISC committed 13 hours ago

Verified 4244ad7a

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

wanghqc committed 17 hours ago

Verified 63b4b303

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

max-krasnyansky committed 20 hours ago

Verified caa8191d

llama : move no_alloc out of llama_hparams (#31141)

ggerganov committed 21 hours ago

Verified a4699400

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

ggerganov committed 21 hours ago

Verified 4b851fbe

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

ServeurpersoCom committed 21 hours ago

Verified 4d71b42f

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

virajwad committed 21 hours ago

Verified 70c5f263

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

Titaniumtown committed 21 hours ago

Verified 5cd62ffe

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

Titaniumtown committed 21 hours ago

Verified 38645c2a

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

Captain-Tripps committed 21 hours ago

Verified 02269b4c

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

KyleHagy committed 21 hours ago

Verified 5320c87f

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

ServeurpersoCom committed 22 hours 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

tests : refactor test-llama-archs to not link internal libllama symbols (#30880)

ggerganov committed 1 day ago

Verified fd8f8451

ci : fix race in webgpu memset (#31162)

eapache committed 1 day ago

Verified 9125a2eb

metal : stop overflow from happening in large matrix output offsets (#30072)

feni6 committed 1 day ago

Verified 9b8446b9

ui: UX improvements (#31030)

allozaur committed 1 day ago

Verified dc9e3bed

ui: Sidebar resizing improvements (#31119)

allozaur committed 1 day ago

Verified 69b3cc40

Older