AI Model Compatibility Screening for Target Device Selection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The challenge of determining a suitable hardware device for executing artificial intelligence models efficiently, considering varying inference performance across different hardware platforms, leads to prolonged product development cycles due to the need for extensive testing and resource knowledge.
Innovation Solution
A method for determining a target device by extracting model information and device resource information, checking compatibility, and generating a candidate device list based on target resource conditions, with options for conversion, compression, or quantization to ensure compatibility.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If extensive testing and hardware knowledge are used to determine suitable AI model execution hardware, then device compatibility and model performance are improved, but product development time increases
Solution Approach 1:
The system performs preliminary actions by pre-collecting and storing device information (hardware specifications, supported frameworks, runtime capabilities) in a device database before actual model deployment. This advance preparation enables quick compatibility checking without extensive testing during product development, resolving the contradiction between ensuring device compatibility and reducing development time
Solution Approach 2:
The patent introduces an intermediary compatibility checking system that acts as a mediator between AI models and target devices. This intermediary automatically compares model requirements with device capabilities using stored device information, eliminating the need for manual extensive testing while ensuring accurate compatibility determination, thus reducing development time without sacrificing reliability
2Measurement precision
If model distribution to actual hardware is performed to check model operation, then hardware compatibility is verified, but resource consumption and testing complexity increase
Solution Approach 1:
Instead of distributing models to actual hardware for testing, the system creates and uses a digital copy of device information stored in the device database. This copying approach allows virtual compatibility verification by comparing model requirements with stored device specifications, achieving accurate hardware compatibility verification without the complexity of physical device testing
Solution Approach 2:
The patent extracts essential device information (hardware specs, supported frameworks, runtime capabilities) from actual devices and stores it in a database. This extraction creates a simplified representation that can be used for compatibility checking without requiring the physical presence of test devices, reducing testing complexity while maintaining verification accuracy
3Measurement precision
If comprehensive device information collection is performed, then compatibility checking accuracy is improved, but information processing time increases
Solution Approach 1:
The system performs preliminary action by collecting and organizing comprehensive device information in advance, storing it in a structured device database with categories such as hardware specifications, supported frameworks, and runtime capabilities. This advance organization enables fast retrieval and comparison during compatibility checking, achieving both high accuracy and efficient processing time
Solution Approach 2:
The patent segments comprehensive device information into distinct categories (hardware specs, framework support, runtime capabilities, memory information). This segmentation allows the compatibility checking system to efficiently query and compare only relevant segments based on model requirements, improving both checking accuracy and processing speed by avoiding unnecessary data analysis
Data Source
AI summary
A method for determining a target device to execute an artificial intelligence (AI) model, performed by a computing device, is disclosed. Upon receiving the AI model from a user terminal, the method extracts model-related information, including runtime, layer, memory, and file size details. The process begins with a first check, sending a signal to a device database to retrieve information about a specific device. It determines whether the model is executable on that device by comparing its runtime information with the model's requirements. If executable, a second check evaluates whether the device's resources—such as memory, file size capacity, and layer support—meet the model's resource conditions. Based on the results, the device may be added to a candidate list of recommended devices for executing the AI model. This ensures optimal device selection for AI model execution.