Data loader for continuous outcome

I have a directory of images, the label for each of them is a float between 0 and 1. How can I load this dataset? The plan is to build a neural net regression model. Thanks!

I have the same question, I’m trying to predicted users rating of an image. I actually have the raw data that generated the image (a scientific plot, the user is asked to rate how well the plot “fits” the data), did you get an answer to this question?