Is Smoothness the Key to Robustness? A Comparison of Attention and Convolution Models Using a Novel Metric

Open in new window