Add a title or text to torch make_grid

vision
1

Hi
Is there a way to add any text to image before saving it?
I want to save and image with two faces and a simularity between them.

def concat(x0, x1):
    concatenated = torch.cat((x0,x1),0)
    output1,output2 = net(Variable(x0).cuda(),Variable(x1).cuda())
    euclidean_distance = F.pairwise_distance(output1, output2)
    return torchvision.utils.make_grid(concatenated), euclidean_distance.cpu().data.numpy()[0][0]

and I would like to save them all together to pass one image to my tensorboardX add_image method.

2

One approach would be to convert the numpy images to PIL images and draw the text on it. Afterwards convert it back to numpy.