Replace the default IModelLoader implementation in Aspose.LLM for .NET to control the full model-loading pipeline....pipeline — file resolution, native llama.cpp model creation, inference...// 3. Delegate to the SDK's native loading path, or call llama...