Description
Local large-language-model inference can run with extra quantization formats and CPU acceleration through OpenBLAS.
It is useful for users experimenting with offline AI models without sending prompts to a remote service. Local models can still process sensitive text and may generate wrong output, so data handling and result review matter.