YOLOv3 detect function

[ErnieDaTryHard]

I'm wondering if anyone has written a detect function for one image for YOLOv3. I've trained the model on a custom dataset and I've tried writing my own detect function by using the same method used to parse an image in the dataset:

np.array(Image.open(image_path).convert("RGB"))

Calling the model with that array, or converting it to a tensor of that array, doesn't work as it says it has 250 channels (IMAGE_SIZE) instead of 3 and I wonder if anyone else has written a function to detect one image.