ggml-org/llama.cpp

Pull Requests Commits

contrib: allow all AI-generated code in general

ngxson committed 30 days ago

4a94f386

hexagon: activation ops update (#25974)

tboinovski1 committed 30 days ago

Verified 0278d836

mtmd: use RAII for setting and resetting non-causal attention (#25723)

nikwen committed 30 days ago

Verified e0833bf6

feat(ui): add symbolic math support to JS sandbox via nerdamer (#25948)

rankaiyx committed 30 days ago

Verified 61328e6a

minor: fix reasoning preserve var for DS4 [no ci] (#25999)

pwilkin committed 30 days ago

Verified e8e6c7af

common: infer the speculative type from the draft repo sidecars (#25989)

ServeurpersoCom committed 30 days ago

Verified 6d5a910c

Fix DeepSeek4 crafted template (#25414)

pwilkin committed 30 days ago

Verified f534da26

ggml: enable PowerPC backend variants on AIX (#25983)

shalinib-ibm committed 30 days ago

Verified 3ce7da2c

ci : fix SYCL package shared library lookup (#25987)

KyleHagy committed 30 days ago

Verified b4d6c7d8

webgpu : add CONV_2D_DW (depthwise conv2d) kernel (#25847)

m1el committed 30 days ago

Verified 7347430f

cuda: GET_ROWS quants (#25962)

ServeurpersoCom committed 30 days ago

Verified c5a4a0bb

llama-arch: fix DeepSeek4 APE tensor op (#25945)

helanfxz committed 31 days ago

Verified 67b9b0e7

Add support for Laguna XS.2 & M.1 (#25165)

joerowell committed 31 days ago

Verified 1f66c3ce

convert: fix handle HunyuanVL XD-RoPE config (#25514)

wendadawen committed 31 days ago

Verified 66e4bf7e

mtmd : use align_corners for qwen3vl vision position embedding interpolation (#25781)

gerbenvv committed 31 days ago

Verified b4aa7dd4

hexagon: check tensor type when reusing descriptors (#25968)

w1049 committed 31 days ago

Verified 71102a73

cuda: add sqrt_softplus in topk-moe for dsv4 (#25896)

am17an committed 31 days ago

Verified 846e991e

kleidiai : warn once when a weight type has no KleidiAI kernel (#25701)

kkjjkamal123 committed 31 days ago

Verified fb0e6b62

common: resolve draft repo to its requested sidecar (#25955)

ServeurpersoCom committed 31 days ago

Verified 60f6a177

server: return 400 instead of 500 on validation error with X-Conversation-Id (#25760)

ServeurpersoCom committed 31 days ago

Verified fd41bf65

server : properly handle null llama_context (#25868)

fairydreaming committed 31 days ago

Verified 40b740ad

vulkan: Refactor vk_queue to use per-instance mutexes and unique handles (#23570)

winstonma committed 31 days ago

Verified f0480101

ggml-openvino: Add GGML_BACKEND_DL_IMPL invocation for OpenVINO backend (#25795)

seijikun committed 31 days ago

Verified 5735e10c

CUDA: vectorize same-type get_rows with int4 copy (#25929)

pwilkin committed 31 days ago

Verified 305ba519

hexagon: add CLAMP op (#25934)

tboinovski1 committed 32 days ago

Verified 76f46ad2

ui: Sidebar Conversations Bulk Action + Improved Settings logic/UI (#25815)

allozaur committed 32 days ago

Verified 2beefef6

llama_dsv4: write only used rows in state (#25325)

am17an committed 32 days ago

Verified 91d2fc38

ui: fix collapsed user bubble with markdown rendering (#25869)

ServeurpersoCom committed 32 days ago

Verified 4ee6a9af

UI: fix Settings/Display tool call content toggle (#25783)

ServeurpersoCom committed 32 days ago

Verified 43b5e635

ui: enable the agentic flow when only the JS sandbox is active (#25865)

ServeurpersoCom committed 32 days ago

Verified 1521a9ac

Older