AI-BMT Platform — Python Submitter Interface (macOS ARM64)

Last Updated: 2024-10-14

---

1. Environment

• ISA(Instruction Set Architecture) : ARM64(aarch64)
• OS : macOS 11.0 (Big Sur) or later
• Python Version: 3.8.X ~ 3.13.X supported

---

2. System Requirements

1. Hardware

• Apple Silicon Mac (M1, M2, M3, or later)
• 8GB RAM minimum, 16GB recommended

2. Software Dependencies

• Python 3.8 or later (conda/homebrew recommended)
• Required Python packages (automatically installed):

  pip install "numpy>=1.21" "opencv-python>=4.5" "onnxruntime>=1.16"

3. Project Description

1. Implement AI_BMT_Interface to operate with the intended AI Processing Unit (e.g., CPU, GPU, NPU).
2. Various task example codes are provided. Use these example codes as a reference to implement the interface for the AI Processing Unit.
3. This macOS ARM64 version includes all necessary Qt frameworks and libraries for standalone distribution.

---

4. Submitter Development Guide

Required Interface

submitter must subclass bmt.AI_BMT_Interface and implement the following methods:

class SubmitterImplementation(bmt.AI_BMT_Interface):

    # Load and initialize your model here
    def initialize(self, model_path: str) -> None:

    # return the implemented interface task type.
    def getInterfaceType(self) -> InterfaceType:

    #  Vision tasks: preprocessing & inference
    #  - preprocessVisionData: convert raw image file into model input format
    #  - inferVision: run inference on preprocessed data and return results
    def preprocessVisionData(self, image_path: str) -> VariantType:
    def inferVision(self, data: List[VariantType]) -> List[BMTVisionResult]:

    # LLM tasks: preprocessing & inference
    # - preprocessLLMData: convert raw input into model input format
    # - inferLLM: run inference on preprocessed data and return results
    def preprocessLLMData(self, llmData: LLMPreprocessedInput) -> VariantType:
    def inferLLM(self, data: List[VariantType]) -> List[BMTLLMResult]:

Optional Interface

submitter can optionally provide hardware/system metadata using:

class SubmitterImplementation(bmt.AI_BMT_Interface):
    def getOptionalData(self) -> Optional_Data:
        data = Optional_Data()
        data.cpu_type = "Apple M1 Pro"  # or M2, M3, etc.
        data.accelerator_type = "Apple Neural Engine"
        data.submitter = "YourCompany"
        data.cpu_core_count = "10"  # 8 performance + 2 efficiency cores
        data.cpu_ram_capacity = "16GB"  # unified memory
        data.cooling = "Air"
        data.cooling_option = "Passive"
        data.cpu_accelerator_interconnect_interface = "Unified Memory Architecture"
        data.benchmark_model = "ResNet-50"
        data.operating_system = "macOS 14.0"  # or your macOS version
        return data

5. Start BMT

using following commands in AI_BMT_GUI_Submitter_MacOS_ARM64_Python/ directory.

Method 1: Direct Python execution

python main.py