Wesam Sakla

@szagoruyko sorry, nn.HingeEmbeddingCriterion. "Creates a criterion that measures the loss given an input x which is a 1-dimensional vector and a label y (1 or -1). This is usually used...

@leefionglee an updates on this? I'm also having the same problem with high recall and low precision due to false positives. @nw362 what is the intuition behind doing what you...

You have to change the sizes of the anchors so that the RPN learns proposals of the sizes of objects that you care about.

@Austriker no, this doesn't solve the rescaling of the images. That can be changed via the appropriate parameters described in config.py. Read the paper for the sizes of default anchors...

@nicholas-leonard are there still plans to include an AutoEncoder example, preferrably a convolutional AE? This would be quite helpful.

@bobbens Hi Edgar, if you could kindly release the training code, I would be very appreciative. Even releasing your model.lua file for training would be quite useful. Thank you!