metal : simplify kernel arguments using a struct (ggml/3229) (llama/12194)

metal : fix default.metallib build (llama/12224)

ggml : fix GGMLMetalClass ODR (llama/12200)

cuda/cpu: Increase support for fp16 unary operations (ggml/1125)

metal : copy kernels for quant to F32/F16 conversions (llama/12017)

metal : fix the crash caused by the lack of residency set support on Intel Macs. (llama/11904)

metal : optimize dequant q6_K kernel (llama/11892)

repo : update links to new url (llama/11886)

metal : avoid breaking build when metal API predates TARGET_OS_VISION (llama/11690)

metal : adjust support conditions for norm operators (llama/11671)

CUDA: non-contiguous (RMS) norm support (llama/11659)

metal : use residency set for other platforms (llama/11648)

metal: Handle null returned from MTLCreateSystemDefaultDevice() (llama/11441)

metal : use residency sets (llama/11427)

metal : fix out-of-bounds write (llama/11314)

ggml : do not install metal source when embed library (ggml/1054)

metal : avoid uint (llama/11019)

llama : add Qwen2VL support + multimodal RoPE (llama/10361)

metal : Extend how Llama.cpp locates metal resources (llama/10676)

ggml: add `GGML_SET` Metal kernel + i32 CPU kernel (ggml/1037)

ggml : add `GGML_PAD_REFLECT_1D` operation (ggml/1034)

ggml : move AMX to the CPU backend (llama/10570)

metal : small-batch mat-mul kernels (llama/10581)

metal : fix group_norm support condition (llama/0)

metal : enable mat-vec kernels for bs <= 4 (llama/10491)

ggml : add support for dynamic loading of backends (llama/10469)

metal : minor code formatting

feat: add `GGML_UNARY_OP_ARGMAX` Metal kernel (ggml/1019)

metal : add `GGML_OP_CONV_TRANSPOSE_1D` kernels (ggml/1026)

ggml : sync resolve (skip) (#0)

metal : fox offset integer overflows in im2col (ggml/1015)

metal : add `GGML_UNARY_OP_ELU` kernel (ggml/1018)

metal : refactor kernel args into structs (llama/10238)

ggml: new optimization interface (ggml/988)

sync : leftovers (ggml/0)

ggml : build backends as libraries (llama/10256)