Skip to main content
This page documents the grid.model.* local inference wrapper. For cloud-hosted inference via GRID Cortex, see the Cortex OWLv2 page.
The OWLv2 implements a wrapper for the OWLv2 model, which detects objects in RGB images based on a text prompt.
​
class OWLv2()
​
use_local
boolean
default:"False"
If True, inference call is run on the local VM, else offloaded onto GRID-Cortex. Defaults to False.
​
box_threshold
float
default:"0.4"
Confidence threshold for bounding box detection.
​
def run()
​
rgbimage
np.ndarray
required
The input RGB image of shape (M,N,3)(M,N,3).
​
prompt
str
required
Text prompt for object detection. Multiple prompts can be separated by a ”,”.
​
Returns
List[float], List[float], List[str]
Returns three lists: bounding boxes coordinates, confidence scores, and label strings.
from grid.model.perception.detection.owlv2 import OWLv2
car = AirGenCar()

# We will be capturing an image from the AirGen simulator 
# and run model inference on it.

img =  car.getImage("front_center", "rgb").data

model = OWLv2(use_local = False)
box, scores, labels = model.run(rgbimage=img, prompt=<prompt>)
print(box, scores, labels)

## if you want to use the model locally, set use_local=True
model = OWLv2(use_local = True)
box, scores, labels = model.run(rgbimage=img, prompt=<prompt>)
print(box, scores, labels)
This code is licensed under the Apache 2.0 License.