ngxson/llama.cpp

Pull Requests Commits
llama : support LiquidAI LFM2-MoE hybrid model (#16464)
tdakhran committed 230 days ago
Verified aeaf8a36
server : add `/v1/health` endpoint (#16461)
ggerganov committed 230 days ago
Verified df1b612e
webui : added download action (#13552) (#16282)
srogmann committed 231 days ago
Verified 4e0388aa
presets : fix pooling param for embedding models (#16455)
ggerganov committed 231 days ago
Verified ef4c5b87
rpc : update documentation (#16441)
rgerganov committed 231 days ago
Verified c61ae20d
memory : use sequential equal splits for recurrent modules (#16442)
ggerganov committed 231 days ago
Verified 0123ff38
metal : add support for non-padded FA KV (#16148)
ggerganov committed 231 days ago
Verified 0a319bb7
tests : add -INF blocks to the KQ mask in the FA tests (#16380)
ggerganov committed 231 days ago
Verified 1d6092fc
metal : various optimizations + refactoring (#16446)
ggerganov committed 231 days ago
Verified 8ae32dc9
llama : add --no-host to disable host buffers (#16310)
Gadflyii committed 231 days ago
Verified 3df2244d
chat : Granite Docling stopping (#16438)
gabe-l-hart committed 231 days ago
Verified c08002a1
ci : refactor sdk caching to minimize storage (#16414)
CISC committed 231 days ago
Verified 3a002afa
ggml : fix unaligned access in AMX code (#16315)
ggerganov committed 231 days ago
Verified a23b9bdb
ci : remove missing reranker model files (#16444)
danbev committed 231 days ago
Verified 04e632a4
ggml-cpu : fix leftover handling in ggml_vec_scale_f32 for SVE (#16443)
danbev committed 231 days ago
Verified a80ff183
nix : removed metal for nix (#16118)
yuannan committed 232 days ago
Verified 1d49ca37
server: update readme to mention n_past_max metric (#16436)
okuvshynov committed 232 days ago
Verified c5fef0fc
model : Granite docling + Idefics3 preprocessing (SmolVLM) (#16206)
gabe-l-hart committed 232 days ago
Verified ca71fb9b
ggml webgpu: actually add softmax, fix rms_norm offset (#16400)
reeselevine committed 233 days ago
Verified 35266573
vulkan: use a more appropriate amount of threads when generating shaders (#16418)
netrunnereve committed 233 days ago
Verified 86df2c9a
rpc : check src buffer when copying tensor (#16421)
rgerganov committed 233 days ago
Verified f3928396
rpc : add support for multiple devices (#16276)
rgerganov committed 234 days ago
Verified 898acba6
vulkan : incremental shader builds (#16341)
Acly committed 234 days ago
Verified e29acf74
chat : support Magistral thinking (#16413)
ServeurpersoCom committed 234 days ago
Verified 128d522c
server : context checkpointing for hybrid and recurrent models (#16382)
ddh0 committed 234 days ago
Verified f6dcda39
metal : fix loop bound in ggml_mem_ranges (#16412)
ggerganov committed 234 days ago
Verified 606a73f5
llama : fix shapes for bert/mpt q/k norm (#16409)
CISC committed 234 days ago
Verified 946f71ed
ggml : fix graph reallocation with multiple chunks (#16396)
Acly committed 234 days ago
Verified 638d3302
Fix missing messages on sibling navigation (#16408)
allozaur committed 235 days ago
Verified 84c8e305
vulkan: Replace uses of maxMemoryAllocationSize and VK_WHOLE_SIZE (#16354)
jeffbolznv committed 235 days ago
Verified 2aaf0a2a
Terms Of Use Privacy Policy Refund Policy Impressum