Classification | Roboflow Docs

Infer on Local and Hosted Images

To install dependencies, pip install inference-sdk.

from inference_sdk import InferenceHTTPClient

CLIENT = InferenceHTTPClient(
    api_url="https://classify.roboflow.com",
    api_key="API_KEY"
)

result = CLIENT.infer(your_image.jpg, model_id="vehicle-classification-eapcd/2")

Node.js

These examples use the built-in fetch API. It works in Node.js 18 and later and in modern browsers, so there is no dependency to install.

Inferring on a Local Image

const fs = require("fs");

const image = fs.readFileSync("YOUR_IMAGE.jpg", {
    encoding: "base64"
});

fetch("https://classify.roboflow.com/your-model/42?api_key=YOUR_KEY", {
    method: "POST",
    headers: {
        "Content-Type": "application/x-www-form-urlencoded"
    },
    body: image
})
    .then((response) => {
        if (!response.ok) throw new Error("Request failed with status " + response.status);
        return response.json();
    })
    .then((data) => {
        console.log(data);
    })
    .catch((error) => {
        console.log(error.message);
    });

Uploading a Local Image Using base64

import UIKit

// Load Image and Convert to Base64
let image = UIImage(named: "your-image-path") // path to image to upload ex: image.jpg
let imageData = image?.jpegData(compressionQuality: 1)
let fileContent = imageData?.base64EncodedString()
let postData = fileContent!.data(using: .utf8)

// Initialize Inference Server Request with API_KEY, Model, and Model Version
var request = URLRequest(url: URL(string: "https://classify.roboflow.com/your-model/your-model-version?api_key=YOUR_APIKEY&name=YOUR_IMAGE.jpg")!,timeoutInterval: Double.infinity)
request.addValue("application/x-www-form-urlencoded", forHTTPHeaderField: "Content-Type")
request.httpMethod = "POST"
request.httpBody = postData

// Execute Post Request
URLSession.shared.dataTask(with: request, completionHandler: { data, response, error in
    
    // Parse Response to String
    guard let data = data else {
        print(String(describing: error))
        return
    }
    
    // Convert Response String to Dictionary
    do {
        let dict = try JSONSerialization.jsonObject(with: data, options: []) as? [String: Any]
    } catch {
        print(error.localizedDescription)
    }
    
    // Print String Response
    print(String(data: data, encoding: .utf8)!)
}).resume()

Response Object Formats

Single-Label Classification

The hosted API inference route returns a JSON object containing an array of predictions. Each prediction has the following properties:

time = total time, in seconds, to process the image and return predictions
image = an object that holds information about the image width and height
- width the height of the predicted image
- height = the height of the predicted image
predictions = collection of all predicted classes and their associated confidence values for the prediction
- class = the label of the classification
- confidence = the model's confidence that the image contains objects of the detected classification
top = highest confidence predicted class
confidence = highest predicted confidence score
image_path = path of the predicted image
prediction_type = the model type used to perform inference, ClassificationModel in this case

// an example JSON object
{
  "time": 0.19064618100037478,
  "image": {
    "width": 210,
    "height": 113
  },
  "predictions": [
    {
      "class": "real-image",
      "confidence": 0.7149
    },
    {
      "class": "illustration",
      "confidence": 0.2851
    }
  ],
  "top": "real-image",
  "confidence": 0.7149,
  "image_path": "/cropped-images-1.jpg",
  "prediction_type": "ClassificationModel"
}

Multi-Label Classification

The hosted API inference route returns a JSON object containing an array of predictions. Each prediction has the following properties:

time = total time, in seconds, to process the image and return predictions
image = an object that holds information about the image width and height
- width the height of the predicted image
- height = the height of the predicted image
predictions = collection of all predicted classes and their associated confidence values for the prediction
- class = the label of the classification
- confidence = the model's confidence that the image contains objects of the detected classification
predicted_classes = an array that contains a list of all classifications (labels/classes) returned in model predictions
image_path = path of the predicted image
prediction_type = the model type used to perform inference, ClassificationModel in this case

<pre class="language-json" data-overflow="wrap"><code class="lang-json"><strong>// an example JSON object </strong>{ "time": 0.19291414400004214, "image": { "width": 113, "height": 210 }, "predictions": { "dent": { "confidence": 0.5253503322601318 }, "severe": { "confidence": 0.5804202556610107 } }, "predicted_classes": [ "dent", "severe" ], "image_path": "/car-model-343.jpg", "prediction_type": "ClassificationModel" } </code></pre>

API Reference

Using the Inference API

POST https://classify.roboflow.com/:datasetSlug/:versionNumber

You can POST a base64 encoded image directly to your model endpoint. Or you can pass a URL as the image parameter in the query string if your image is already hosted elsewhere.

Path Parameters

Name	Type	Description
datasetSlug	string	The url-safe version of the dataset name. You can find it in the web UI by looking at the URL on the main project view.
	string	The version number identifying the version of your dataset.

Query Parameters

Name	Type	Description
api_key	string	Your API key (obtained via your workspace API settings page)

{
   "predictions":{
      "bird":{
         "confidence":0.5282308459281921
      },
      "cat":{
         "confidence":0.5069406032562256
      },
      "dog":{
         "confidence":0.49514248967170715
      }
   },
   "predicted_classes":[
      "bird",
      "cat"
   ]
}

{
   "message":"Forbidden"
}j