Because I use reduce_mean, change loss = tf.reshape(loss, [self.batch_size, self.cell_size * self.cell_size * self.box_per_cell * (self.num_class + 5)]) and loss = 0.5 * tf.reduce_mean(tf.reduce_mean(loss, axis = 1)) to loss = tf.reduce_mean(tf.reduce_sum(loss, axis = [1, 2, 3, 4]), name = 'loss'), the loss is a little bigger.
Because I use reduce_mean, change
loss = tf.reshape(loss, [self.batch_size, self.cell_size * self.cell_size * self.box_per_cell * (self.num_class + 5)])
andloss = 0.5 * tf.reduce_mean(tf.reduce_mean(loss, axis = 1))
toloss = tf.reduce_mean(tf.reduce_sum(loss, axis = [1, 2, 3, 4]), name = 'loss')
, the loss is a little bigger.