How to access predictions?

[mexicantexan]

💡 Your Question

I have this code:

image = cv2.imread(image_path, cv2.IMREAD_UNCHANGED)
if image.shape[0] != 640 or image.shape[1] != 640:
    image = cv2.resize(image, (640, 640))

# make sure that the image is 3 channel rgb
if len(image.shape) == 2:
    image = cv2.cvtColor(image, cv2.COLOR_GRAY2RGB)
elif image.shape[2] == 4:
    image = cv2.cvtColor(image, cv2.COLOR_RGBA2RGB)
detect_net = models.get(Models.YOLO_NAS_M, pretrained_weights="coco")
detect_net.eval()
pred = detect_net.predict(image)

I can't seem to access the prediction in the variable pred, but if I substitute that line with: detect_human_net.predict(image).show() the output is shown. I'm just trying to access the predictions, nothing too dramatic. I've tried using draw() but that doesn't seem to be an option for the class that is returned from the predict call.

Versions

No response

[dagshub[bot]]

Join the discussion on DagsHub!

[Louis-Dupont]

Hi @mexicantexan , if I understood correctly, you should find what you are looking for in this page. Is it what you were looking for?

Notes

• You can, but don't need to load with cv2.imread, the model.predict() method does it.
• You don't need to call detect_net.eval(), it is also done as part of the predict.

[mexicantexan]

Yes, that worked perfectly! Sorry, just saw this :)