poxtest_3_predict.py

def predict(model, image_url):
	'''
	pass the model and image url to the function
	Returns: a list of pox types with decreasing probability
	'''
	if validators.url(image_url) is True:	
		response = requests.get(image_url)		
		picture = Image.open(BytesIO(response.content))
	else:
		picture = Image.open(image_url)
	# Convert the image to grayscale and other transforms
	image = data_transform(picture)
	# store in a list of images
	images=image.reshape(1,1,64,64)
	new_images = images.repeat(1, 3, 1, 1)
	outputs=model(new_images)
	# get prediction
	_, predicted = torch.max(outputs, 1)
	ranked_labels=torch.argsort(outputs,1)[0]
	# get all classes in order of probability
	probable_classes=[]
	for label in ranked_labels:
	    probable_classes.append(classes[label.numpy()])
	probable_classes.reverse()
	return probable_classes