ggerganov/llama.cpp

Pull Requests Commits
`sync`: minja (#11641)
ochafik committed 214 days ago
Verified 9f4cc8f8
CUDA: non-contiguous (RMS) norm support (#11659)
JohannesGaessler committed 214 days ago
Verified fd08255d
HIP: force max threads per block to be 1024 (#11621)
fxzjshm committed 214 days ago
Verified 3ec9fd4b
server : add try..catch to places not covered by set_exception_handler (#11620)
ngxson committed 214 days ago
Verified 3962fc1a
arg : list RPC devices first when using --list-devices (#11655)
rgerganov committed 214 days ago
Verified 1bef571f
`tool-call`: command r7b fix for normal responses (#11608)
ochafik committed 215 days ago
Verified db288b60
readme : add llm_client Rust crate to readme bindings (#11628)
ShelbyJenkins committed 215 days ago
Verified 106045e7
swift : fix llama-vocab api usage (#11645)
jhen0409 committed 215 days ago
Verified f117d84b
metal : use residency set for other platforms (#11648)
jhen0409 committed 215 days ago
Verified 534c46b5
authors : update
ggerganov committed 215 days ago
Verified 387a1598
sync : ggml
ggerganov committed 215 days ago
Verified 7c9e0ca5
cmake: Add ability to pass in GGML_BUILD_NUMBER (ggml/1096)
ckastner committed 215 days ago
Verified 8f8290ad
ci : do not stale-close roadmap issues
ggerganov committed 215 days ago
Verified b34aedd5
`tool-call`: allow `--chat-template chatml` w/ `--jinja`, default to chatml upon parsing issue, avoid double bos (#11616)
ochafik committed 215 days ago
Verified cde38332
server : (webui) revert hacky solution from #11626 (#11634)
ngxson committed 215 days ago
Verified b3451785
server : (webui) allow typing and submitting during llm response (#11626)
woof-dog committed 215 days ago
Verified 1d1e6a90
server : remove CPPHTTPLIB_NO_EXCEPTIONS define (#11622)
danbev committed 216 days ago
Verified 5598f475
sync : ggml
ggerganov committed 216 days ago
Verified 8ec05832
CUDA: fix Volta FlashAttention logic (#11615)
JohannesGaessler committed 216 days ago
Verified 21c84b5d
server : (webui) Fix Shift+Enter handling (#11609)
mashdragon committed 216 days ago
Verified d92cb67e
HIP: fix flash_attn_stream_k_fixup warning (#11604)
JohannesGaessler committed 216 days ago
Verified 6eecde3c
CUDA/HIP: add support for selectable warp size to mmv (#11519)
IMbackK committed 216 days ago
Verified 396856b4
HIP: add GGML_CUDA_CC_IS_* for amd familys as increasing cc archtectures for amd gpus are not supersets of eatch other (#11601)
IMbackK committed 216 days ago
Verified 4d0598e1
nit: more informative crash when grammar sampler fails (#11593)
ochafik committed 216 days ago
Verified 90f9b88a
CUDA: use mma PTX instructions for FlashAttention (#11583)
JohannesGaessler committed 216 days ago
Verified 864a0b67
Name colors (#11573)
ericcurtin committed 217 days ago
Verified 84ec8a58
`tool-call`: support Command R7B (+ return tool_plan "thoughts" in API) (#11585)
ochafik committed 217 days ago
Verified bfcce4d6
Fix exotic ci env that lacks ostringstream::str (#11581)
ochafik committed 217 days ago
Verified 69804487
sampling : support for llguidance grammars (#10224)
mmoskal committed 217 days ago
Verified ff227703
llama : add support for GLM-Edge and GLM-Edge-V series models (#10573)
piDack committed 217 days ago
Verified 0cec062a
Terms Of Use Privacy Policy Refund Policy Impressum