ggerganov/llama.cpp

Pull Requests Commits

graph : reuse recurrent graphs

ggerganov committed 10 hours ago

Verified 638e2c23

graph : reuse hybrid graphs

ggerganov committed 10 hours ago

Verified 245f3915

server : host-memory prompt caching (#16391)

ggerganov committed 11 hours ago

Verified d00cbea6

No markdown in cot (#16483)

ServeurpersoCom committed 11 hours ago

Verified 8328fd4b

model-conversion : add support for SentenceTransformers (#16387)

danbev committed 14 hours ago

Verified 56b47958

ci: add ARM64 Kleidiai build and test support (#16462)

sudhiarm committed 18 hours ago

Verified 2c0d875a

CANN: Improve ACL graph matching (#16166)

noemotiovon committed 19 hours ago

Verified aa4711d3

kleidiai: kernel interface refactoring (#16460)

chaxu01 committed 19 hours ago

Verified d80d6d24

[SYCL] refactor soft_max, add soft_max_back (#16472)

NeoZhangJianyu committed 19 hours ago

Verified b2602137

model: EmbeddingGemma Adding Support for SentenceTransformers Dense Modules (#16367)

sfallah committed 20 hours ago

Verified e08db425

refactor: centralize CoT parsing in backend for streaming mode (#16394)

ServeurpersoCom committed 1 day ago

Verified 12bbc3fa

Disable CUDA host buffers on integrated GPUs (#16308)

ai-fonsi committed 1 day ago

Verified 9d088284

server : fix cancel pending task (#16467)

issixx committed 1 day ago

Verified d2ee056e

metal : mark FA blocks (#16372)

ggerganov committed 1 day ago

Verified b2c08c9e

server : improve context checkpoint logic (#16440)

ggerganov committed 1 day ago

Verified 7fdd16b4

ggml webgpu: profiling, CI updates, reworking of command submission (#16452)

reeselevine committed 2 days ago

Verified 74b8fc17

llama : support LiquidAI LFM2-MoE hybrid model (#16464)

tdakhran committed 2 days ago

Verified aeaf8a36

server : add `/v1/health` endpoint (#16461)

ggerganov committed 2 days ago

Verified df1b612e

webui : added download action (#13552) (#16282)

srogmann committed 2 days ago

Verified 4e0388aa

presets : fix pooling param for embedding models (#16455)

ggerganov committed 2 days ago

Verified ef4c5b87

rpc : update documentation (#16441)

rgerganov committed 2 days ago

Verified c61ae20d

memory : use sequential equal splits for recurrent modules (#16442)

ggerganov committed 2 days ago

Verified 0123ff38

metal : add support for non-padded FA KV (#16148)

ggerganov committed 2 days ago

Verified 0a319bb7

tests : add -INF blocks to the KQ mask in the FA tests (#16380)

ggerganov committed 2 days ago

Verified 1d6092fc

metal : various optimizations + refactoring (#16446)

ggerganov committed 2 days ago

Verified 8ae32dc9

llama : add --no-host to disable host buffers (#16310)

Gadflyii committed 3 days ago

Verified 3df2244d

chat : Granite Docling stopping (#16438)

gabe-l-hart committed 3 days ago

Verified c08002a1

ci : refactor sdk caching to minimize storage (#16414)

CISC committed 3 days ago

Verified 3a002afa

ggml : fix unaligned access in AMX code (#16315)

ggerganov committed 3 days ago

Verified a23b9bdb

ci : remove missing reranker model files (#16444)

danbev committed 3 days ago

Verified 04e632a4

Older