ggerganov/llama.cpp

Pull Requests Commits

ggerganov committed 126 days ago

Verified 65202d29

ggml : fix ggml_gallocr_ptr type (ggml/1205)

slaren committed 127 days ago

Verified db1ff5b6

cuda : fix unused variable compile warning (whisper/0)

ggerganov committed 127 days ago

Verified 610df4cc

CUDA: batched+noncont MMQ, refactor bs>1 MoE code (#13199)

JohannesGaessler committed 127 days ago

Verified e1e8e099

arg : -hf do not fail if url mismatch (#13219)

ngxson committed 127 days ago

Verified 6f67cf1f

fix typo: `n_ctx_pre_seq` -> `n_ctx_per_seq` (#13221)

ddh0 committed 127 days ago

Verified 16a457fa

convert : improve model arch handling (#13122)

ngxson committed 127 days ago

Verified 3e168bed

llava : remove duplicate include (#13207)

tattn committed 127 days ago

Verified ceda28ef

common : add -jf / --json-schema-file flag (#12011)

ochafik committed 127 days ago

Verified 3b127c73

vulkan: use uint array index to avoid glslang bug (#13193)

jeffbolznv committed 127 days ago

Verified e5007a5e

ggml : fix ppc64le build (#13176)

shalinib-ibm committed 127 days ago

Verified 41631377

convert : correct typo image_mean --> image_std (#13208)

ngxson committed 127 days ago

Verified 07c2e2f7

feat(ggml-cpu): enable z17 compile (#13182)

taronaeo committed 127 days ago

Verified 44cd8d91

arg : allow using -hf offline (#13202)

ngxson committed 127 days ago

Verified 5933e6fd

docker : do not build tests (#13204)

ngxson committed 127 days ago

Verified da84c04d

rpc : fix cache directory initialization (#13188)

hbuxiaofei committed 128 days ago

Verified a0f7016d

scripts: n_depth for compare-llama-bench [no ci] (#13201)

JohannesGaessler committed 128 days ago

Verified 19e899ce

server : Prefilling assistant message in openai compatible API (#13174)

matteoserva committed 128 days ago

Verified e2e1ddb9

sampling : when top-k <= 0 -> noop (#13173)

ggerganov committed 128 days ago

Verified d9d398f8

llama-bench: fixed size of fields to correctly map to values (#13183)

Alcpz committed 128 days ago

Verified 5a639801

CUDA: fix non-cont. inputs for batched mat mul (#13155)

JohannesGaessler committed 128 days ago

Verified cdf76586

llama : llm_type order by size (#13177)

CISC committed 128 days ago

Verified 7d3af70b

mtmd : add qwen2vl and qwen2.5vl (#13141)

ngxson committed 128 days ago

Verified 00e3e5a1

llama : set qwen3 model type sizes (#13175)

CISC committed 128 days ago

Verified e98b3692

llama-graph : fix text position for mrope (#13159)

ngxson committed 129 days ago

Verified b6ce7430

model : Nomic Embed Text V2 with Mixture-of-Experts (MoE) architecture (#12466)

manyoso committed 129 days ago

Verified 5f5e39e1

clip : fix model size display (#13153)

ngxson committed 129 days ago

Verified eaea3253

fix(rpc): Improve input validation and error handling (#13069)

thevilledev committed 129 days ago

Verified 43ddab6e

llama-bench: add `-d` depth arg (#13096)

thevishalagarwal committed 129 days ago

Verified 1831f538

mtmd : fix glm-edge redundant token count (#13139)

ngxson committed 129 days ago

Verified 4e87962e

Older