Problem of (GoogleNet) with deconf and ./optimize_image.py

[BBAUniversity]

I have tried GoogleNet default model of deep-visualization-toolbox. Deconf dont work and the result of this commande: ./optimize_image.py --push-layer 'loss3/classifier' --push-channel 0 --decay 0.0001 --blur-radius 1.0 --blur-every 4 --max-iter 6000 --lr-policy constant --lr-params "{'lr': 100.0}

Give the the Error:

Traceback (most recent call last): File "./optimize_image.py", line 225, in main() File "./optimize_image.py", line 187, in main raw_scale = 1.0, File "/home/deep/Downloads/caffe-master/python/caffe/classifier.py", line 34, in init self.transformer.set_mean(in_, mean) File "/home/deep/Downloads/caffe-master/python/caffe/io.py", line 259, in set_mean raise ValueError('Mean shape incompatible with input shape.') ValueError: Mean shape incompatible with input shape.

[VanNope]

Hello, I have the same problem. Did you manage to solve this issue? @BBAUniversity

[heitorrapela]

"File "/home/deep/Downloads/caffe-master/python/caffe/io.py", line 259, in set_mean raise ValueError('Mean shape incompatible with input shape.') ValueError: Mean shape incompatible with input shape."

Try to use --data-size "(224,224)", and checkout this link

./optimize_image.py --push-layer pool5/7x7_s1 --decay 0.0001 --blur-radius 1.0 --blur-every 4 --max-iter 5 --lr-policy constant --lr-params "{'lr': 100.0}" --data-size "(224,224)"