ggml-org/llama.cpp

Pull Requests Commits

cpu: fix ARM NEON nvfp4 vec dot

0cc4m committed 38 days ago

a30369d5

hexagon: slight optimization for argosrt output init (#21463)

YardenTal44 committed 38 days ago

Verified 25eec6f3

llama : correct platform-independent loading of BOOL metadata (#21428)

anchortense committed 38 days ago

Verified 58190cc8

model : add HunyuanOCR support (#21395)

richarddd committed 39 days ago

Verified af76639f

ci : use default RISE RISC-V Runners (#21263)

luhenry committed 39 days ago

Verified 761797ff

server : fix logging of build + system info (#21460)

ddh0 committed 39 days ago

Verified 5d3a4a7d

ci: lower cuda12 floor to 12.8.1 for broader host compatibility (#21438)

M1DNYT3 committed 39 days ago

Verified c08d28d0

ci: fix vulkan workflow referencing non-existent action (#21442)

nisparks committed 39 days ago

Verified 661e9acb

common : add gemma 4 specialized parser (#21418)

aldehir committed 40 days ago

Verified b8635075

server: Fix undefined timing measurement errors in server context (#21201)

thedanhoffman committed 40 days ago

Verified 9c699074

common : respect specified tag, only fallback when tag is empty (#21413)

angt committed 40 days ago

Verified d01f6274

llama-model: read final_logit_softcapping for Gemma 4 (#21390)

ssam18 committed 40 days ago

Verified 650bf14e

llama: add custom newline split for Gemma 4 (#21406)

am17an committed 40 days ago

Verified b7ad48eb

ggml-webgpu: move from parameter buffer pool to single buffer with offsets (#21278)

reeselevine committed 41 days ago

Verified d0068583

ci: Add Windows Vulkan backend testing on Intel (#21292)

rillomas committed 41 days ago

Verified e4397009

server: save and clear idle slots on new task (`--clear-idle`) (#20993)

yychyo committed 41 days ago

Verified 50e0ad08

common/parser: fix call ID detection (Mistral parser mostly) + atomicity for tag-json parsers (#21230)

pwilkin committed 41 days ago

Verified f1f793ad

common : fix tool call type detection for nullable and enum schemas (#21327)

sacredvoid committed 41 days ago

Verified af5c1384

docker : bump cuda12 to 12.9.1 (#20920)

M1DNYT3 committed 41 days ago

Verified 277ff5ff

docs: Update build.md: HSA_OVERRIDE_GFX_VERSION clarification (#21331)

jeromew committed 41 days ago

Verified 384c0076

jinja: coerce input for string-specific filters (#21370)

CISC committed 41 days ago

Verified 1f34806c

ci: add more binary checks (#21349)

taronaeo committed 41 days ago

Verified 887535c3

fix: remove stale assert (#21369)

pwilkin committed 41 days ago

Verified d3416a4a

HIP: build eatch ci build test for a different architecture (#21337)

IMbackK committed 41 days ago

Verified 43a4ee4a

fix: add openssl to nix dependencies (#21353) (#21355)

Tillerino committed 41 days ago

Verified f851fa5a

ggml-zendnn : add MUL_MAT_ID op support for MoE models (#21315)

z-vishal committed 41 days ago

Verified f1ac8411

vocab: fix Gemma4 tokenizer (#21343)

pwilkin committed 41 days ago

Verified b069b10a

rpc : reuse compute graph buffers (#21299)

rgerganov committed 41 days ago

Verified 0c58ba33

chat : avoid including json in chat.h (#21306)

ggerganov committed 41 days ago

Verified 57ace0d6

(revert) kv-cache : do not quantize SWA KV cache (#21332)

ggerganov committed 41 days ago

Verified 39b27f0d

Older