4 ms·

Without going into detail, our model outputs the coordinates of the html elements it recognize. We are taking a visual approach to the problem.

• attilakun 7y ago
  Did you use a recurrent neural network? And/or reinforcement learning?
• xcubic 7y ago
  I imagine you do some kind of image recognition on the sketch files and then do the code generation.