5c3c9461dca89468e265c4b2822f514f25fcaafc,GPflow/vgp.py,VGP,build_predict,#VGP#Any#,74

Before Change

            L = tf.cholesky(A)
            LiKx = tf.user_ops.triangular_solve(L, tf.transpose(Kx), "lower")
            return Kd - tf.reduce_sum(tf.square(LiKx),0)
        f_var, _ = theano.scan(v, self.q_lambda.T)
        return f_mean, tf.transpose(f_var)

After Change

        //predictive var
        f_var = []
        for d in range(self.num_latent):
            b = self.q_lambda[:,d]
            A = K + tf.diag(1./tf.square(b))
            L = tf.cholesky(A)
            LiKx = tf.user_ops.triangular_solve(L, tf.transpose(Kx), "lower")
            f_var.append( Kd - tf.reduce_sum(tf.square(LiKx),0) )
        f_var = tf.pack(f_var)
        return f_mean, tf.transpose(f_var)

Instances