⬆️ Update ggerganov/llama.cpp #3078

Job	Run time
core-image-build (cublas, 12, 4, linux/amd64, false, -cublas-cuda12-ffmpeg-core, true, core, ubun... / reusable_image-build	1h 21m 5s
core-image-build (linux/amd64, false, -ffmpeg-core, true, core, ubuntu-latest, ubuntu:22.04, --jo... / reusable_image-build	19m 38s
core-image-build (sycl_f16, linux/amd64, false, quay.io/go-skynet/intel-oneapi-base:latest, ubunt... / reusable_image-build	39m 32s
core-image-build (vulkan, linux/amd64, false, -vulkan-ffmpeg-core, true, core, ubuntu-latest, ubu... / reusable_image-build	22m 6s
extras-image-build (cublas, 12, 4, linux/amd64, false, -cublas-cuda12-ffmpeg, true, extras, arc-r... / reusable_image-build	2h 7m 34s
extras-image-build (hipblas, linux/amd64, false, -hipblas, false, extras, rocm/dev-ubuntu-22.04:6... / reusable_image-build	6h 0m 21s
extras-image-build (sycl_f16, linux/amd64, false, quay.io/go-skynet/intel-oneapi-base:latest, ubu... / reusable_image-build	1h 15m 40s
extras-image-build (linux/amd64, false, -ffmpeg, true, extras, arc-runner-set, ubuntu:22.04, --jo... / reusable_image-build	1h 9m 30s
	13h 15m 26s

Provide feedback