ggml-org/llama.cpp

Pull Requests Commits
llama : add log about loading model tensors
ggerganov committed 1 year ago
Verified 3b6a0a81
vulkan: optimize coopmat2 iq2/iq3 callbacks (#11521)
jeffbolznv committed 1 year ago
Verified 2c6c8df5
vulkan: initial support for IQ4_XS quantization (#11501)
remyoudompheng committed 1 year ago
Verified 8a7e3bf1
vulkan: use smaller combined allocations to avoid fragmentation (#11551)
jeffbolznv committed 1 year ago
Verified 1b598b30
metal : avoid breaking build when metal API predates TARGET_OS_VISION (#11690)
charles-dyfis-net committed 1 year ago
Verified 902368a0
readme : add link to Autopen under UIs (#11684)
blackhole89 committed 1 year ago
Verified c3db0480
metal : adjust support conditions for norm operators (#11671)
ggerganov committed 1 year ago
Verified d774ab3a
CUDA: support for mat. mul. with ne03 != ne13 (#11656)
JohannesGaessler committed 1 year ago
Verified fa62da9b
llava: add quantization for the visual projector LLAVA, Qwen2VL (#11644)
samkoesnadi committed 1 year ago
Verified 1ec20808
`sync`: minja (#11641)
ochafik committed 1 year ago
Verified 9f4cc8f8
CUDA: non-contiguous (RMS) norm support (#11659)
JohannesGaessler committed 1 year ago
Verified fd08255d
HIP: force max threads per block to be 1024 (#11621)
fxzjshm committed 1 year ago
Verified 3ec9fd4b
server : add try..catch to places not covered by set_exception_handler (#11620)
ngxson committed 1 year ago
Verified 3962fc1a
arg : list RPC devices first when using --list-devices (#11655)
rgerganov committed 1 year ago
Verified 1bef571f
`tool-call`: command r7b fix for normal responses (#11608)
ochafik committed 1 year ago
Verified db288b60
readme : add llm_client Rust crate to readme bindings (#11628)
ShelbyJenkins committed 1 year ago
Verified 106045e7
swift : fix llama-vocab api usage (#11645)
jhen0409 committed 1 year ago
Verified f117d84b
metal : use residency set for other platforms (#11648)
jhen0409 committed 1 year ago
Verified 534c46b5
authors : update
ggerganov committed 1 year ago
Verified 387a1598
sync : ggml
ggerganov committed 1 year ago
Verified 7c9e0ca5
cmake: Add ability to pass in GGML_BUILD_NUMBER (ggml/1096)
ckastner committed 1 year ago
Verified 8f8290ad
ci : do not stale-close roadmap issues
ggerganov committed 1 year ago
Verified b34aedd5
`tool-call`: allow `--chat-template chatml` w/ `--jinja`, default to chatml upon parsing issue, avoid double bos (#11616)
ochafik committed 1 year ago
Verified cde38332
server : (webui) revert hacky solution from #11626 (#11634)
ngxson committed 1 year ago
Verified b3451785
server : (webui) allow typing and submitting during llm response (#11626)
woof-dog committed 1 year ago
Verified 1d1e6a90
server : remove CPPHTTPLIB_NO_EXCEPTIONS define (#11622)
danbev committed 1 year ago
Verified 5598f475
sync : ggml
ggerganov committed 1 year ago
Verified 8ec05832
CUDA: fix Volta FlashAttention logic (#11615)
JohannesGaessler committed 1 year ago
Verified 21c84b5d
server : (webui) Fix Shift+Enter handling (#11609)
mashdragon committed 1 year ago
Verified d92cb67e
HIP: fix flash_attn_stream_k_fixup warning (#11604)
JohannesGaessler committed 1 year ago
Verified 6eecde3c
Terms Of Use Privacy Policy Refund Policy Impressum