Merge branch 'ggerganov:master' into master #5

Job	Run time
ubuntu-latest-cmake	2m 4s
ubuntu-focal-make	1m 33s
ubuntu-latest-cmake-mpi (mpich)	1m 53s
ubuntu-latest-cmake-sanitizer (ADDRESS, Debug)	3m 19s
ubuntu-latest-cmake-mpi (libopenmpi-dev)	1m 48s
ubuntu-latest-cmake-sanitizer (ADDRESS, Release)	5m 13s
ubuntu-latest-cmake-sanitizer (THREAD, Debug)	8m 27s
ubuntu-latest-cmake-sanitizer (THREAD, Release)	5m 59s
ubuntu-latest-cmake-sanitizer (UNDEFINED, Debug)	2m 38s
ubuntu-latest-cmake-sanitizer (UNDEFINED, Release)	4m 41s
macOS-latest-cmake	2m 9s
macOS-latest-cmake-ios	1m 12s
macOS-latest-cmake-tvos	1m 39s
macOS-latest-make	2m 16s
macOS-latest-swift (generic/platform=macOS)	2m 7s
macOS-latest-swift (generic/platform=iOS)	2m 34s
windows-latest-cmake (noavx, -DLLAMA_NATIVE=OFF -DLLAMA_BUILD_SERVER=ON -DLLAMA_AVX=OFF -DLLAMA_A...	3m 59s
macOS-latest-swift (generic/platform=tvOS)	2m 30s
windows-latest-cmake (avx2, -DLLAMA_NATIVE=OFF -DLLAMA_BUILD_SERVER=ON -DBUILD_SHARED_LIBS=ON)	3m 28s
windows-latest-cmake (avx, -DLLAMA_NATIVE=OFF -DLLAMA_BUILD_SERVER=ON -DLLAMA_AVX2=OFF -DBUILD_SH...	3m 49s
windows-latest-cmake (avx512, -DLLAMA_NATIVE=OFF -DLLAMA_BUILD_SERVER=ON -DLLAMA_AVX512=ON -DBUIL...	2m 59s
windows-latest-cmake-cublas (12.2.0, cublas)	30m 19s
windows-latest-cmake (clblast, -DLLAMA_NATIVE=OFF -DLLAMA_BUILD_SERVER=ON -DLLAMA_CLBLAST=ON -DBU...	4m 8s
windows-latest-cmake-cublas (11.7.1, cublas)	28m 48s
windows-latest-cmake (openblas, -DLLAMA_NATIVE=OFF -DLLAMA_BUILD_SERVER=ON -DLLAMA_BLAS=ON -DBUIL...	3m 26s
release	1m 7s
	2h 14m 5s

Provide feedback