Cross-platform algorithm testing method and system

Abstract

The invention provides a cross-platform algorithm testing method and a system thereof. The method comprises steps: after a user logs on a testing client, if the testing authority is verified to be passed, the testing client obtaining a testing task corresponding to the user ID from a Web server, a test case and test data for the user to select; obtaining the test task selected by the user, as a target test task, the target test data and the target test platform corresponding to the target test task being obtained for any target test task, the target test instructions corresponding to the target test task being generated, and the target test data and the target test instructions being sent to the target test platform for testing; receiving the test results reported by each target test platform, and making statistical analysis on all the test results. The method and the system can support a plurality of platforms and a plurality of algorithms, and can dynamically expand and adapt to newrequirements. To a certain extent, it reduces the burden of operation and storage of the test platform, and is conducive to the sharing of resources among the test platforms.

Description

technical field [0001] The invention relates to the technical field of algorithm testing, and more specifically, to a cross-platform algorithm testing method and system. Background technique [0002] At present, most software functions are implemented based on algorithms. For example, face detection software is implemented based on face algorithms. All kinds of algorithms need to be tested before practical application, especially for the emerging algorithms based on deep learning, which usually need to be tested through a huge test set after deep learning. That is, there are testing requirements for various algorithms in the prior art, and each type of algorithm often needs to be tested with a large amount of test data. [0003] Existing algorithm evaluation tools are all platform-specific, dedicated test tools for a single algorithm, lack support for multiple algorithms on multiple platforms, and cannot dynamically expand and adapt to new requirements. Moreover, this spec...

AI Technical Summary

Problems solved by technology

[0005] In order to overcome the defects that the existing algorithm evaluation tools are all based on a specific platform for a single algorithm, lack of support for multiple platforms and multiple algorithms, and cannot dynamically expand and adapt to new requirements, etc., the present invention provides a cross-platform Platform algorithm testing method and system

Images