refactor: optimize the handling of LoRA models (#1070) #1590

Job	Run time
macOS-latest-cmake	4m 28s
ubuntu-latest-cmake	2m 40s
windows-latest-cmake (noavx, -DGGML_NATIVE=OFF -DGGML_AVX=OFF -DGGML_AVX2=OFF -DGGML_FMA=OFF -DSD...	2m 17s
windows-latest-cmake (vulkan, -DSD_VULKAN=ON -DSD_BUILD_SHARED_LIBS=ON)	5m 50s
windows-latest-cmake (cuda12, -DSD_CUDA=ON -DSD_BUILD_SHARED_LIBS=ON -DCMAKE_CUDA_ARCHITECTURES='...	3h 45m 49s
windows-latest-cmake (avx2, -DGGML_NATIVE=OFF -DGGML_AVX2=ON -DSD_BUILD_SHARED_LIBS=ON)	2m 9s
windows-latest-cmake-hip	4m 25s
windows-latest-cmake (avx, -DGGML_NATIVE=OFF -DGGML_AVX=ON -DGGML_AVX2=OFF -DSD_BUILD_SHARED_LIBS...	2m 6s
windows-latest-cmake (avx512, -DGGML_NATIVE=OFF -DGGML_AVX512=ON -DGGML_AVX=ON -DGGML_AVX2=ON -DS...	2m 16s
release	59s
	4h 12m 59s

Provide feedback