ggerganov/llama.cpp

Pull Requests Commits
Make logging more verbose (#11714)
ericcurtin committed 1 year ago
Verified d2fe216f
llama : fix defrag logic (#11707)
ggerganov committed 1 year ago
Verified ed926d88
vocab : ignore invalid UTF-8 input in the BPE tokenizer (#11729)
cfillion committed 1 year ago
Verified 2d219b38
llama : fix progress dots (#11730)
magicse committed 1 year ago
Verified 333820d7
vulkan: print shared memory size (#11719)
jeffbolznv committed 1 year ago
Verified c026ba3c
llama : add llama_sampler_init for safe usage of llama_sampler_free (#11727)
cfillion committed 1 year ago
Verified 7ee953a6
SYCL: remove XMX info from print devices (#11712)
qnixsynapse committed 1 year ago
Verified ec3bc827
common : add default embeddings presets (#11677)
danbev committed 1 year ago
Verified b7552cfc
ggml : optimize and build warning fix for LoongArch (#11709)
MQ-mengqing committed 1 year ago
Verified 225bbbfa
llama : fix old glm4 models (#11670)
tv1wnd committed 1 year ago
Verified 855cd073
sync : ggml
ggerganov committed 1 year ago
Verified 8a59053f
rpc: fix known RCE in rpc-server (ggml/1103)
retr0reg committed 1 year ago
Verified 1d20e53c
server : (webui) migrate project to ReactJS with typescript (#11688)
ngxson committed 1 year ago
Verified 2fb3c32a
docs: update fedora cuda guide for 12.8 release (#11393)
teihome committed 1 year ago
Verified 9ab42dc7
SYCL: Adjust support condition for norm operators (#11674)
qnixsynapse committed 1 year ago
Verified 194b2e69
llama : add log about loading model tensors (#11699)
ggerganov committed 1 year ago
Verified 9dd7a039
build : fix llama.pc (#11658)
angt committed 1 year ago
Verified c0d48432
ggml : fix LoongArch compile error with 128-bit SIMD (#11701)
junchao-loongson committed 1 year ago
Verified 8d4d2be1
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
Terms Of Use Privacy Policy Refund Policy Impressum