# Using the Trained Model

When your model is prepared, you can implement the model. First we need to initialize the model;

const modelUrl = 'file://C:/Projects/my-chatbot/models/model.json';
const metadataUrl = 'models/model_metadata.json';

const mymodel = await bkLabs.nlu.loadModel(modelUrl, metadataUrl, callbackFunction);

# Parameters

modelUrl: string This is the URL where your trained model is located. You need to provided an Absolute URL here.

metadataUrl: string This is the URL where model_metadata.json file is located. You need to provide a Relative URL here.

callbackFunction: Provide a callback function to be executed when the model is loaded. This is where you can start your classification work.

# Returns

Array[0] Initialized Model Array[1] Response Text data.

# Encode Text Input

Once the model is initialized, you need to encode the text input before parsing it to the model to classify.

const sentence = 'Hello, nice day!';

const encodedSentence = await bkLabs.nlu.encodeText(sentence);

# Parameters

sentence: string Text input to classify.

# Returns

Tensor Output Tensor.

# Classify Sentence

Once the text input is encoded, pass the encoded tensor to classify input text.

const predictData = await mymodel[0].predict(encodedSentence);

# Parameters

encodedSentence Tensor encoded text input

# Returns

Tensor Prediction result as a Tensor

# Get a Reply

Now you can get a reply that matches the intent of your input.

const myReply = bkLabs.nlu.predictReply(predictData, mymodel[1]);
console.log(myReply);

# Parameters

predictData Tensor Output Tensor from prediction

mymodel[1] Array<object> Response data extracted earlier when we initialized the model.

# Returns

string Returns a reply as string.

# Example

The final example code will look something like this:

const modelUrl = 'file://C:/Projects/my-chatbot/models/model.json';
const metadataUrl = 'models/model_metadata.json';
const sentence = 'Hello, nice day!';

const mymodel = await bkLabs.nlu.loadModel(modelUrl, metadataUrl, callbackFunction);

async function callbackFunction() {
    const encodedSentence = await bkLabs.nlu.encodeText(sentence);
    const predictData = await mymodel[0].predict(encodedSentence);
    const myReply = bkLabs.nlu.predictReply(predictData, mymodel[1]);
    console.log(myReply);
}

Please feel free to report a bug (opens new window) if you find while using berkelium (opens new window).

← Training the Model