loss = 0.0
loss += l2_loss
loss += square_loss
netDist = NetDist(x, x_reconstructed, config)
if config['style_weight'] > 0.0:
style_loss, style_out_1, style_out_2 = netDist.get_dist_style()
style_loss *= config['style_weight']
else:
style_loss = style_out_1 = style_out_2 = theano.shared(np.asarray(0.0).astype('float32'))
if config['content_weight'] > 0.0:
content_loss = config['content_weight'] * netDist.get_dist_content()
else:
content_loss = theano.shared(np.asarray(0.0).astype('float32'))
#128 x 3 x 96 x 96
imgGrad = T.grad(style_loss + content_loss, x_reconstructed)[0,:,:,0]
loss += style_loss + content_loss
loss += 1.0 * variational_loss
all_grads = T.grad(loss, params)
scaled_grads = lasagne.updates.total_norm_constraint(all_grads, 5.0)
updates = lasagne.updates.adam(scaled_grads, params, learning_rate = 0.001)
loss = 0.0
loss += l2_loss
loss += square_loss
netDist = NetDist(x, x_reconstructed, config)
if config['style_weight'] > 0.0:
style_loss, style_out_1, style_out_2 = netDist.get_dist_style()
style_loss *= config['style_weight']
else:
style_loss = style_out_1 = style_out_2 = theano.shared(np.asarray(0.0).astype('float32'))
if config['content_weight'] > 0.0:
content_loss_values, varLst = netDist.get_dist_content()
content_loss = sum(content_loss_values.values()) * config['content_weight']
params += varLst
else:
content_loss = theano.shared(np.asarray(0.0).astype('float32'))
content_loss_values, varLst = netDist.get_dist_content()
#128 x 3 x 96 x 96
loss += style_loss + content_loss
loss += 1.0 * variational_loss
loss += 1.0 * smoothness_penalty
all_grads = T.grad(loss, params)
