Deploy and Iterate Deep Learning Model
In this phase, you can deploy the trained deep learning model to the production line for inference, enabling quality inspection. After the model has been deployed, if its performance declines, you can iterate and optimize the model.
Deploy the Deep Learning Model
The AI-based quality inspection solution supports integrating the trained deep learning model into the quality control system (software) of the production line by using Mech-DLK SDK.
You can use the APIs provided by Mech-DLK SDK to build the deep learning inference component within the existing quality control system (software) in C#, C++, and C languages.
-
For more information about how to install Mech-DLK SDK, see Mech-DLK SDK Installation Guide.
-
For more information about integration and development, see Sample Usage Guide.
-
For more information about how to use Mech-DLK SDK for inference, see Mech-DLK SDK Get Started and Mech-DLK SDK API Reference.
In addition, the AI-based quality inspection solution allows you to deploy deep learning models into the machine vision software Mech-Vision. For more information, see Use Models in Mech-Vision.
Iterate the Deep Learning Model
When a model is deployed on the production line for some time, it might not cover certain scenarios. At this point, you need to iterate and optimize the model to ensure the model maintains satisfactory inference performance.
Mech-DLK allows you to use the Model Finetuning function for model iteration so as to maintain accuracy and save time.
-
To iterate a single algorithm module, refer to General Model Iteration.
-
To iterate cascaded modules, refer to Cascaded Model Iteration.