Cortex.cpp: Inference, Params & Error Handling

[dan-homebrew]

Goal

• Cortex.cpp routes Model Inference to respective engines
• Cortex.cpp should expose full set of Inference Params from engines (e.g. llama.cpp)
• Cortex.cpp should handle Inference Errors gracefully, and show informative error messages that enable user to recover

Key Decisions

Architecture

Related

• feat: Jan supports most llama.cpp params jan#3508