Quick StartΒΆ How to Support a New LLM: Step-by-Step Step 1: Acquire the Model Assets Step 2: Convert to mllm Format Step 3: (Optional) On-Device Quantization with mllm-quantizer Step 4: Implement Core C++ Files Step 5: Create an Example Application Step 6: Open a Pull Request Conclusion Appendix How to Add a New Operator in MLLM Overview Step 1: Define the Operator Type Step 2: Create the Operator Interface Step 3: Implement Backend Support Step 4: Register the Operator Factory Step 5: Add to IR (Intermediate Representation) Usage Example Best Practices Conclusion How to run modules async Basic Usage The fork function The wait function Notification Complete Example How to perf modules Prerequisites Basic Usage Example Analyzing Results Performance Categories Best Practices