YOLO-World

YOLO-World is a zero-shot object detection model that allows you to perform object detection without any training, just by describing the items you want to detect.

You can also run YOLO-World locally using Inference, our open-source inference server.

API Reference

The base URL for our hosted API is at https://infer.roboflow.com.

More information on using YOLO-World with Inference, through the Python SDK or the self-hosted API, see the YOLO-World Inference docs page.

YOLO-World inference.

Run the YOLO-World zero-shot object detection model.

POST/yolo_world/infer

Query parameters

Body

application/json

id*Id

api_keyApi Key

Roboflow API Key that will be passed to the model during initialization for artifact retrieval

stringnull

usage_billableUsage Billable

startStart

numbernull

sourceSource

stringnull

source_infoSource Info

stringnull

model_idModel Id

stringnull

model_typeModel Type

The type of the model, usually referring to what task the model performs

stringnull

image*Image

array of InferenceRequestImageInferenceRequestImage

disable_preproc_auto_orientDisable Preproc Auto Orient

If true, the auto orient preprocessing step is disabled for this call.

booleannull

disable_preproc_contrastDisable Preproc Contrast

If true, the auto contrast preprocessing step is disabled for this call.

booleannull

disable_preproc_grayscaleDisable Preproc Grayscale

If true, the grayscale preprocessing step is disabled for this call.

booleannull

disable_preproc_static_cropDisable Preproc Static Crop

If true, the static crop preprocessing step is disabled for this call.

booleannull

text*Text

A list of strings

yolo_world_version_idYolo World Version Id

stringnull

confidenceConfidence

numbernull

Response

200422

Successful Response

Body

application/json

visualizationVisualization

Base64 encoded string containing prediction visualization image data

stringnull

inference_idInference Id

Unique identifier of inference

stringnull

frame_idFrame Id

The frame id of the image used in inference if the input was a video

integernull

timeTime

The time in seconds it took to produce the predictions including image preprocessing

numbernull

image*Image

array of InferenceResponseImageInferenceResponseImage

predictions*Predictions

Request

JavaScriptCurlPython

const response = await fetch('/yolo_world/infer', {
    method: 'POST',
    headers: {
      "Content-Type": "application/json"
    },
    body: JSON.stringify({
      "id": "text",
      "image": [
        {
          "type": "text"
        }
      ],
      "text": [
        "text"
      ]
    }),
});
const data = await response.json();

Response

200422

{
  "visualization": "text",
  "inference_id": "text",
  "frame_id": 0,
  "time": 0,
  "image": [
    {
      "width": 0,
      "height": 0
    }
  ],
  "predictions": [
    {
      "x": 0,
      "y": 0,
      "width": 0,
      "height": 0,
      "confidence": 0,
      "class": "text",
      "class_confidence": 0,
      "class_id": 0,
      "tracker_id": 0,
      "detection_id": "text",
      "parent_id": "text"
    }
  ]
}