Loss function

[SyedTauhidUllahShah]

Kindly explain this self.loss = tf.reduce_sum((1 + 0.1 * self.y) * tf.square( (self.y * self.pred + (1 - self.y) * tf.nn.relu(self.beta * (1 - self.y) - self.pred)))) This is different from what u have written in the paper.

[cheungdaven]

I have revised the loss. self.loss = tf.reduce_sum((1 + 0.1 * self.y ) * tf.square(self.beta * (self.one - self.y) - self.pred_y))

[SyedTauhidUllahShah]

What's does self. one mean?

[cheungdaven]

oh, sorry about that, it should be 1.