Deploying an Image Classification Service

Deploying a Service

You can deploy a model as a real-time service that provides a real-time test UI and monitoring capabilities. After model training is complete, you can deploy a version with the ideal accuracy and in the Successful status as a service. The procedure is as follows:

On the phase execution page, after the service deployment status changes to Awaiting input, double-click Deploy Service. On the configuration details page, configure resource parameters.
On the service deployment page, select the resource specifications used for service deployment.
- AI Application Source: defaults to the generated AI application.
- AI Application and Version: The current AI application version is automatically selected, which is changeable.
- Resource Pool: defaults to public resource pools.
- Traffic Ratio: defaults to 100 and supports a value range of 0 to 100.
- Specifications: Select available specifications based on the list displayed on the console. The specifications in gray cannot be used in the current environment. If there are no specifications after you select a public resource pool, no public resource pool is available in the current environment. In this case, use a dedicated resource pool or contact the administrator to create a public resource pool.
- Compute Nodes: an integer ranging from 1 to 5. The default value is 1.
- Auto Stop: enables a service to automatically stop at a specified time. If this function is not enabled, the real-time service continuously runs and fees are incurred accordingly. Auto stop is enabled by default and its default value is 1 hour later.
  The auto stop options are 1 hour later, 2 hours later, 4 hours later, 6 hours later, and Custom. If you select Custom, enter any integer from 1 to 24 in the text box on the right.
  
  You can choose the package that you have bought when you select specifications. On the configuration fee tag, you can view your remaining package quota and how much you will pay for any extra usage.
After configuring resources, click Next. Wait until the status changes to Executed. The AI application has been deployed as a real-time service.

Testing a Service

After the service is deployed, click Instance Details to go to the real-time service details page. Click the Prediction tab to test the service. You can also use code to test a service. For details, see Accessing Real-Time Services.

Figure 1 Testing the service

The following describes the procedure for performing a service test after the image classification model is deployed as a service on the ExeML page.

After the model is deployed, click Instance Details in the service deployment phase to go to the service page. On the Prediction tab page, click Upload and select a local image for test.

Click Prediction to conduct the test. After the prediction is complete, label sunflowers and its detection score are displayed in the prediction result area on the right. If the model accuracy does not meet your expectation, add images on the Label Data tab page, label the images, and train and deploy the model again. Table 1 describes the parameters in the prediction result. If you are satisfied with the model prediction result, call the API to access the real-time service as prompted. For details, see Accessing Real-Time Services.

Only JPG, JPEG, BMP, and PNG images are supported.

Figure 2 Prediction result

**Table 1** Parameters in the prediction result
Parameter	Description
predicted_label	Image prediction label
scores	Prediction confidence of top 5 labels

A running real-time service continuously consumes resources. If you do not need to use the real-time service, stop the service to stop billing. To do so, click Stop in the More drop-down list in the Operation column. If you want to use the service again, click Start.
If you enable the auto stop function, the service automatically stops after the specified time and no fee is generated.