llama.cpp

35feac65 - ggml : sync (mem align to header + conv_transpose_2d fixes + ggml_alloc) (#2852)

ggml : sync (mem align to header + conv_transpose_2d fixes + ggml_alloc) (#2852) * ggml : sync (mem align to header + conv_transpose_2d fixes) ggml-ci * ggml-alloc : minor fix * ggml-alloc : sync more fixes