Deploy Model and Get Prediction Results

0 %

Deploy Model and Get Prediction Results

Details

Deploy Model and Get Prediction Results

Juliana Morais

December 20, 2020

Created by

December 14, 2020

Beginner

20 min.

Perform the final steps to train your own Business Entity Recognition custom model to get machine learning entity predictions for the text you submit to the service.

You will learn

How to deploy your own Business Entity Recognition machine learning model
How to send an inference request to the service and get machine learning entity predictions for unstructured text

The core functionality of Business Entity Recognition is to automatically detect and highlight any given type of named entity in unstructured text and classify it in accordance with predefined categories. In this tutorial, you will learn how to use the service APIs to deploy your own machine learning model to get named entity predictions for the texts you submit to the service.

Step 1: List models

To see all your models, you can use the GET /models endpoint to see the list of them.

Click the endpoint name to expand it.
Click Try it out.
Click Execute.

You should receive a response like the following:

The response includes the Business Entity Recognition pre-trained machine learning models (sap_email_business_entity and sap_invoice_header), and the new custom model you have created, in this case, Tutorial_custom_model.

Done

Log on to answer question

Step 2: List model versions

Use the GET /models/{modelName}/versions endpoint to see the available versions and other details about a specific model.

Click the endpoint name to expand it.
Click Try it out.
Enter the modelName (Tutorial_custom_model in this case).
Click Execute.

You should receive a response like the following:

Done

Log on to answer question

Step 3: Deploy model

Finally, to make actual predictions, you have to deploy your model. You can do so by using the POST /deployments endpoint.

Click the endpoint name to expand it.
Click Try it out.
In payload, enter the modelName (Tutorial_custom_model in this case.), and the modelVersion (1 in this case).
Click Execute.
Copy the deploymentId from the Response body to check the status of the deployment in the next step.

Done

Log on to answer question

Step 4: See deployment status

As with training jobs, you now have to check the status of the deployment every now and then. Use the GET /deployments/{deploymentId} endpoint to do so.

Click the endpoint name to expand it.
Click Try it out.
Enter the deploymentId obtained in the previous step.
Click Execute.

You should receive a response like below. Note that, compared to training jobs, deployments do not have a status SUCCEEDED. The status RUNNING indicates that the deployment is live and the model can be used for predictions.

Which deployment status indicates that deployment is done and the model can be used for predictions? SUCCEEDED RUNNING PENDING FAILED

Step 5: Enter inference text

View Full Instructions

To make a prediction, or in machine learning terms an inference, you use the POST /inference/jobs endpoint to submit a text from which your model should extract entities.

Click the endpoint name to expand it.
Click Try it out.

In payload, enter the text you want to extract named entities from, modelName, and modelVersion. You may use the following example:

{
   "text":"Mrs. Hardi Shah, Email id: hardi.shah@yuhu.com, hdshah@xy.com Contact No: 9000900090 Objective: Designation: Sales & Marketing, Digita/ To work with an organization where my strong work ethics, Marketing dedication, organizational skills, communication skills and educational Education: expertise can be utilized in the growth of organization as well as mine, MBA as a manager. Gujarat Technical University Work Experience: Year: 1.2 Professional Achievements: Employer: MAP project on Awareness regarding Dry waste and Blue Star Company (1 Year) wet waste segregation AT Ahmedabad Municipal Assistant Sales Manager Corporation. A comprehensive project report on \"a study on reach Software and Tool: to the patients through socialmedia marketing by the MS Excel/Power Point Hobbies: Extra-Curricular Activities: Reading Books, Travelling PPO in MBA SEMESTER -2 Operating System: Attend Workshop Classroom to Corporate for 7 days. Accomplished \"The Fundamentals of Digital Marketing at Google Linux, Windows Digital Garage on 27/04/2020. Marita/ Status: Single ACADEMIC CREDENTIALS: Nationality: NO. DEGREE INSTITUTE PASSING PERCENTAGE Indian YEAR Mailing Address: 1 MBA GTU 2019 8.5( CGPI) Naranpura, Ahmedabad. India 2 BBA GU 2017 67% PIN: 380013 Language Known: 3 HSC Gujarat 2014 70% Board English, Hindi, Gujarati 4 SSC Gujarat 2012 50% Board DECLARATION: I declare that the above details which I have given are correct.",
   "modelName":"Tutorial_custom_model",
   "modelVersion":1
}

Click Execute.
Copy the id from the Response body to see the result of the extraction in the next step. Please also note the remark regarding limits in Trial. See Trial Account Input Limits.

Done

Log on to answer question

Step 6: Get extraction results

View Full Instructions

Use the GET /inference/jobs/{jobId} endpoint to see the text extraction results and the confidence level of the Tutorial_custom_model custom model.

Click the endpoint name to expand it.
Click Try it out.
Enter the jobId obtained in the previous step and click Execute.

You should receive a response like the following:

In the response, you will find the prediction for the extracted entities. The prediction is made with a probability indicated by the confidence field which represents how certain the model is about the value provided. A confidence of 1 means that the model is 100% sure about its prediction. The model tries to provide a value for all its capabilities. Thus, you may see a different result depending on your model’s capabilities. In case the model cannot identify an entity in the text, it returns an empty value.

Below, you find an example of a full prediction:

{
  "data": {
    "id": "5a183e26-4e5c-4139-ae44-9d17d09648eb",
    "status": "SUCCESS",
    "result": [
      {
        "title": [],
        "firstname": [
          {
            "value": "Mrs. Hardi",
            "confidence": 0.73
          }
        ],
        "lastname": [
          {
            "value": "Shah",
            "confidence": 0.8
          }
        ],
        "emailID": [
          {
            "value": "hardi.shah@yuhu.com, hdshah@xy.com",
            "confidence": 0.97
          }
        ],
        "mobile": [
          {
            "value": "9000900090",
            "confidence": 1
          }
        ],
        "designation": [
          {
            "value": "Sales & Marketing",
            "confidence": 0.97
          }
        ],
        "address": [
          {
            "value": "Naranpura",
            "confidence": 0.9
          }
        ],
        "city": [
          {
            "value": "Ahmedabad",
            "confidence": 0.9
          }
        ],
        "country": [
          {
            "value": "India",
            "confidence": 1
          }
        ],
        "zip": [
          {
            "value": "380013",
            "confidence": 1
          }
        ]
      }
    ],
    "trialUsage": {
      "inferenceRequestsUsage": "Maximum limit of 40 , utilized 2 Inference count and remaining 38",
      "inferenceCharactersUsage": "Maximum limit of 140000 , utilized 1812 Inference characters and remaining 138188"
    }
  }
}

You have now successfully used your own custom model to get text entity predictions for the unstructured text you submitted to Business Entity Recognition.

Done

Log on to answer question

Next Steps

Navigate tutorial steps

Step 1: List models
Step 2: List model versions
Step 3: Deploy model
Step 4: See deployment status
Step 5: Enter inference text
Step 6: Get extraction results
Back to Top

QR Code

Deploy Model and Get Prediction Results

Deploy Model and Get Prediction Results

You will learn

Next Steps

Call me now