release : v1.7.4

release : v1.7.3

release : v1.7.2

release : v1.7.1

release : v1.7.0

revert : cmake : set MSVC to use UTF-8 on source files (#2346)

cmake : set MSVC to use UTF-8 on source files (#2346)

cmake : allow external ggml

cmake : remove install of llama convert script [no ci] (#2266)

whisper : reorganize source code + improve CMake (#2256)

build : update make / cmake

cmake : fix sycl build (#0)

ggml : remove OpenCL (#0)

cuda : enable CUDA graphs (#0)

cmake : fix CUDA build (#0)

whisper : add CUDA-specific computation mel spectrograms (#2206)

release : v1.6.2

release : v1.6.1

examples : add support for decoding input with ffmpeg (Linux) (#2133)

release : v1.6.0

cmake : fix HIP/ROCm build (#2102)

cmake : fix metal embed sources path (#2110)

build : improve disabling AVX-512 (#2129)

release : v1.5.5

build : fix embedded Metal library generation (#2045)

build : detect AVX512 in Makefile, add AVX512 option in CMake (#2043)

cmake : support for CPU BLAS build via Intel MKL (#2024)

cmake : create solution folders (#2004)

build : use pkg-config for OpenBLAS (#1778)

sync : ggml (#2001)

cmake : copy ggml-common.h to bin

cmake : add library versioning (#1352)

whisper : add SYCL support (#1863)

build : update CBLAS flags + fix unused var warning (#0)

metal : option to embed MSL source into compiled binary (#1842)

cmake : make libwhisper.so position independent (#1792)

cmake : temporary remove VLA check (#1795)

release : v1.5.4

release : v1.5.3

examples : fix WASM Stack Overflow (#1713)

sync : ggml (VMM, sync-ggml-am, dotprod ARM fixes, CUDA fixes) (#1691)

cmake : set default CUDA architectures (#1667)

release : v1.5.2

cmake : install required ggml.h header (#1568)

release : v1.5.1

release : v1.5.0

release : v1.4.3

sync : ggml (backend v2, k-quants, CUDA opts, Metal opts, etc.) (#1422)

cmake : Abort the build if a requested feature could not be configured (#1350)

cmake : Prefer pkg-config while looking for BLAS (#1349)