Tensorflow Serving Client

Targeting Node.js (v6) and TypeScript

Provides:

• Prediction of images;
• And that’s it so far.

Usage Example

import TensorflowServingClient = require( "tensorflow-serving-node-typescript-client");
import fs = require('fs');
class TensorflowServing {
    private client: any;
    private imagePath: string = "path/to/image.jpg";
    public constructor() {
        this.client = new TensorflowServingClient.TensorflowServingClient("localhost:9000");
    }
    public predict(): void {
        try {
            let buffer = fs.readFileSync(this.imagePath);
            this.client.Predict(buffer)
                .then((result) => {
                    console.log(result);
                })
                .catch((error) => {
                    console.log(error);
                })
        }
        catch (e) {
            console.log(e);
        }
    }
}
export = TensorflowServing;

Notes

• This is a work in progress, feel free to improve it.