b3dab0288ddcd165e2ba6f95061b5f3d7bf82a1a,GPflow/vgp.py,VGP,build_predict,#VGP#Any#Any#,122

Before Change

        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.matrix_triangular_solve(L, tf.transpose(Kx), lower=True)
            if full_cov:
                f_var.append(self.kern.K(Xnew) -
                             tf.matmul(tf.transpose(LiKx), LiKx))
            else:
                f_var.append(self.kern.Kdiag(Xnew) -
                             tf.reduce_sum(tf.square(LiKx), 0))
        f_var = tf.pack(f_var)
        return f_mean, tf.transpose(f_var)

After Change

        f_mean = tf.matmul(tf.transpose(Kx), self.q_alpha) + self.mean_function(Xnew)

        // predictive var
        A = K + tf.batch_matrix_diag(tf.transpose(1./tf.square(self.q_lambda)))
        L = tf.batch_cholesky(A)
        Kx_tiled = tf.tile(tf.expand_dims(Kx, 0), [self.num_latent, 1, 1])
        LiKx = tf.batch_matrix_triangular_solve(L, Kx_tiled)
        if full_cov:

Instances