search

backseason / PoolNet

Code for our CVPR 2019 paper "A Simple Pooling-Based Design for Real-Time Salient Object Detection"
MIT License
629 stars 153 forks source link

Got an error when I test the pretrained model. #85

Closed JimmyMa99 closed 2 years ago

JimmyMa99 commented 2 years ago

Given input size: (512x7x41). Calculated output size: (512x0x5). Output size is too small File "/media/ders/mazhiming/Saliency_maps_network/PoolNet-master/networks/poolnet.py", line 52, in forward y = self.convsi File "/media/ders/mazhiming/Saliency_maps_network/PoolNet-master/networks/poolnet.py", line 107, in forward merge = self.deep_pool[0](conv2merge[0], conv2merge[1], infos[0]) File "/media/ders/mazhiming/Saliency_maps_network/PoolNet-master/solver.py", line 72, in test preds = self.net(images) File "/media/ders/mazhiming/Saliency_maps_network/PoolNet-master/main.py", line 48, in main test.test() File "/media/ders/mazhiming/Saliency_maps_network/PoolNet-master/main.py", line 98, in main(config

JimmyMa99 commented 2 years ago

000000207005 the image（000000207005） in coco17 caused this error.