update latest code from master#3

Closed

l3utterfly wants to merge 12 commits intoreact-native-modulefrom

Commits on Jun 19, 2023

Convert vector to f16 for dequantize mul mat vec (#1913)
JohannesGaessler
authored
cmake : fix build shared ggml when CUDA is enabled (#1929)

howard0su
and
ggerganov
authored
ggml : sync latest ggml repo (#1924)
ggerganov
authored
cuda : faster k-quants on older GPUs (#1930)

ikawrakow
and
Iwan Kawrakow
authored
llama : only use Q6_K for output weights if tensor size is multiple of 256 (#1932)

ikawrakow
and
Iwan Kawrakow
authored
cmake : fix trailing whitespaces
ggerganov
committed
llama : use aligned memory during ggml_init call from loading saved sessions (#1934)
l3utterfly
authored
ggml : fix bug in LBFGS optimizer (found by ggml tests)
ggerganov
committed
[Fix] Reenable server embedding endpoint (#1937)
SlyEcho
authored

Commits on Jun 20, 2023