Dino v3

Run DINOv3 classification models trained on Roboflow via the Serverless Hosted API

Use Meta's DINOv3 self-supervised vision model for classification through the Serverless Hosted API. DINOv3 produces strong general-purpose visual features that can be adapted to your dataset by training a linear probe classifier on Roboflow.

There are no public DINOv3 aliases. Train your own DINOv3 classifier on your dataset and call it by your {model_id}/{version}.

Code sample

Install the Inference SDK:

pip install inference-sdk

Pass your Roboflow API Key via the API_KEY env variable, and replace your-project/1 with your own model ID and version.

import os
import urllib.request
from inference_sdk import InferenceHTTPClient

image_url = "https://media.roboflow.com/notebooks/examples/dog.jpeg"
image_path = "dog.jpeg"
urllib.request.urlretrieve(image_url, image_path)

client = InferenceHTTPClient(
    api_url="https://serverless.roboflow.com",
    api_key=os.getenv("API_KEY"),
)
results = client.infer(image_path, model_id="your-project/1")
print(results)

Set api_url to match your deployment target:

• https://serverless.roboflow.com for the Serverless Hosted API.

• http://localhost:9001 for a local Inference server.

• Your Dedicated Deployment URL for a private endpoint.

For self-hosted deployment and additional details, see the Inference documentation.