Model for attention map visualisation

Pretrained model and log for attention map visualisation in Figure 3.

Code for detection and segmentation

Release code for detection and segmentation.

LIT models and logs

ImageNet pretrained models and logs of LIT.