is ConeKotlinTypeProjection -> KtTypeArgumentWithVariance(
                buildKtType(coneType.type),
                coneType.kind.toVariance(),
                token,
            )
        }

        private fun ProjectionKind.toVariance(): Variance = when (this) {
            ProjectionKind.OUT -> Variance.OUT_VARIANCE
            ProjectionKind.IN -> Variance.IN_VARIANCE

      if (values.hasAnyNonFinite()) {
        assertWithMessage("population covariance of " + values).that(populationCovariance).isNaN();
        assertWithMessage("population covariance by addAll(PairedStats) of " + values)
            .that(populationCovarianceByAddAllPartitionedPairedStats)
            .isNaN();
      } else {
        assertWithMessage("population covariance of " + values)
            .that(populationCovariance)

      if (values.hasAnyNonFinite()) {
        assertWithMessage("population covariance of " + values).that(populationCovariance).isNaN();
        assertWithMessage("population covariance by addAll(PairedStats) of " + values)
            .that(populationCovarianceByAddAllPartitionedPairedStats)
            .isNaN();
      } else {
        assertWithMessage("population covariance of " + values)
            .that(populationCovariance)

  }

  /** Returns an immutable snapshot of the statistics on the {@code y} values alone. */
  public Stats yStats() {
    return yStats.snapshot();
  }

  /**
   * Returns the population covariance of the values. The count must be non-zero.
   *
   * <p>This is guaranteed to return zero if the dataset contains a single pair of finite values. It

  public Stats xStats() {
    return xStats;
  }

  /** Returns the statistics on the {@code y} values alone. */
  public Stats yStats() {
    return yStats;
  }

  /**
   * Returns the population covariance of the values. The count must be non-zero.
   *
   * <p>This is guaranteed to return zero if the dataset contains a single pair of finite values. It

      double populationCovariance = stats.populationCovariance();
      if (values.hasAnyNonFinite()) {
        assertWithMessage("population covariance of " + values).that(populationCovariance).isNaN();
      } else {
        assertWithMessage("population covariance of " + values)
            .that(populationCovariance)
            .isWithin(ALLOWED_ERROR)

      double populationCovariance = stats.populationCovariance();
      if (values.hasAnyNonFinite()) {
        assertWithMessage("population covariance of " + values).that(populationCovariance).isNaN();
      } else {
        assertWithMessage("population covariance of " + values)
            .that(populationCovariance)
            .isWithin(ALLOWED_ERROR)

## Changes to contrib APIs

*   Adds tf.contrib.nn.rank_sampled_softmax_loss, a sampled-softmax variant that
    can improve rank loss.
*   `tf.contrib.metrics`.{streaming_covariance,streaming_pearson_correlation}
    modified to return nan when they have seen less or equal to 1 unit of
    weight.
*   Adds time series models to contrib. See contrib/timeseries/README.md for
    details.