Search for a command to run...
LatteGAN: Visually Guided Language Attention for Multi-Turn Text-Conditioned Image Manipulation