Add code path for LLAMA_CPP plugins to load models directly from file #25836

Job	Run time
Smart_CI	23s
Build	14m 3s
Conformance (OP)	16m 45s
Conformance (API)	2m 38s
ONNX Runtime Integration	0s
GPU_Stub	1s
NVIDIA plugin	8m 54s
OpenVINO tokenizers extension	5m 33s
Debian Packages / Debian Packages	2m 25s
Samples / Samples	8m 5s
OpenVINO JS API / OpenVINO JS API	41s
CPU functional tests / CPU functional tests	18m 44s
C++ unit tests / C++ unit tests	14m 22s
ONNX Models Tests / ONNX Models tests	25m 22s
TensorFlow TF Hub Models tests	0s
TensorFlow Hugging Face Models tests	0s
Python unit tests / Python unit tests	39m 50s
PyTorch Models tests / PyTorch Models tests	22m 32s
TensorFlow Models tests / TensorFlow Models tests	9m 50s
ci/gha_overall_status	1s
	3h 10m 9s

Provide feedback