Learning Generalizable Robot Policies with Targeted Generative Augmentation