miyosuda/scan

SCAN: Learning Abstract Hierarchical Compositional Visual Concepts

scan's Issues

Random value of threshold for labeling in IMG2SYM

Hi,
In the following part, I was wondering why it has a random threshold for assigning the label? It seems to result in an inconsistent threshold for different items in the list?

scan/data_manager.py

Lines 220 to 225 in cc86131

	def choose_labels(self, y):
	""" retrieve label element from img2sym output. """
	label_indices = []
	for i,v in enumerate(y):
	if random.random() <= v:
	label_indices.append(i)

About the loss setting in beta-VAE

Hello! I have a question about the loss setting in beta-VAE.

In the MODEL ARCHITECTURE part of the paper, the authors said that they "replace the pixel log-likelihood term in Eq. 2 with an L2 loss in the high-level feature space of DAE", as is implemented in your model. [loss = L2(z_d - z_out_d) + beta * KL]

However, in the MODEL DETAILS part, they also said that "The reconstruction error was taking in the last layer of the DAE (in the pixel space of DAE reconstructions) using L2 loss and before the non-linearity." It seems that the loss should be [loss = L2(x_d - x_out_d) + beta * KL]

I'm wondering which is right and why they are inconsistent. Because with pre-trained DAE, in the course of training beta-VAE, I find these two terms of loss didn't work well (the reconstr-loss is much larger than latent-loss).

Look forward to any reply. Thanks a lot!