predict.prototxt fix

[onorabil]

The patch_transformer layer from the generated predict.prototxt required image_label for the bottom blob. I believe this shoudn't happen, since no labels are supposed to be required for prediction. The changes below make prediction possible. Am I wrong?

Totally unrelated, could you please add a seed for the Plain-Mnih-NN-Multi-S-Maxout with Dropout model used in your article "Building and road detection from large aerial imagery"? My group is working on the same topic and we would love to compare our results [pixel by pixel, if possible] with your predictions.

Cheers.

[mitmul]

Thank you for committing my project.
You are absolutely correct. It's a bug in create_models.py script. I usually modify generated predict.prototxt manually every time. So I merged your pull request.

I also added a seed of Plain-Mnih-NN-Multi-S-Maxout with Dropout as "Mnih_CNN_S_Maxout", but it is worse than VGG_PReLU model. Please consider comparing your model with VGG_PReLU.

[onorabil]

You're welcome! Thank you for providing me your model and for the heads up regarding VGG_PReLU. We were planning to cite your article and for our application, accuracy is not of utmost importance. Stuff to test we have.

Keep up the good work, we're grateful for it. All the best!