apiVersion: apps/v1beta2
kind: Scale
metadata:
  annotations:
    annotationsKey: annotationsValue
  creationTimestamp: "2008-01-01T01:01:01Z"
  deletionGracePeriodSeconds: 10
  deletionTimestamp: "2009-01-01T01:01:01Z"
  finalizers:
  - finalizersValue
  generateName: generateNameValue
  generation: 7
  labels:
    labelsKey: labelsValue
  managedFields:
  - apiVersion: apiVersionValue
    fieldsType: fieldsTypeValue

apiVersion: extensions/v1beta1
kind: Scale
metadata:
  annotations:
    annotationsKey: annotationsValue
  creationTimestamp: "2008-01-01T01:01:01Z"
  deletionGracePeriodSeconds: 10
  deletionTimestamp: "2009-01-01T01:01:01Z"
  finalizers:
  - finalizersValue
  generateName: generateNameValue
  generation: 7
  labels:
    labelsKey: labelsValue
  managedFields:
  - apiVersion: apiVersionValue
    fieldsType: fieldsTypeValue

            k <= bias -> TMIN
            k >= bias + TMAX -> TMAX
            else -> k - bias
          }
        if (digit < t) break
        val scaleW = BASE - t
        if (w > Int.MAX_VALUE / scaleW) return false // Prevent overflow.
        w *= scaleW
      }
      bias = adapt(i - oldi, codePoints.size + 1, oldi == 0)
      val deltaN = i / (codePoints.size + 1)

// RUN: odml-to-stablehlo-opt %s --tf-stablehlo=skip-partitioned-calls=true | FileCheck %s --check-prefix=CHECK-SKIP
// RUN: odml-to-stablehlo-opt %s --tf-stablehlo=skip-partitioned-calls=false | FileCheck %s --check-prefix=CHECK-NOSKIP

module {
  func.func @partitioned_call(%arg0: tensor<1x2x2x3xf32>) -> (tensor<1x2x2x3xf32>) {
    %0 = "tf.StatefulPartitionedCall"(%arg0) <{
      config = "", config_proto = "", executor_type = "", f = @some_func
    }> {

// in the range [a, b].
//
// This distribution is *not* normalized (it is not a probability
// distribution). This makes it easier to work with as it's being
// updated.
//
// It is represented as the sum of scaled uniform distribution
// functions and Dirac delta functions (which are treated as
// degenerate uniform distributions).
type mud struct {
	sorted, unsorted []edge

          - --container=metrics-server
          - --poll-period=30000
          - --estimator=exponential
          # Specifies the smallest cluster (defined in number of nodes)
          # resources will be scaled to.
          - --minClusterSize={{ metrics_server_min_cluster_size }}
          # Use kube-apiserver metrics to avoid periodically listing nodes.
          - --use-metrics=true
      volumes:

	}
	return intOrPercent
}

// GetScaledValueFromIntOrPercent is meant to replace GetValueFromIntOrPercent.
// This method returns a scaled value from an IntOrString type. If the IntOrString
// is a percentage string value it's treated as a percentage and scaled appropriately
// in accordance to the total, if it's an int value it's treated as a simple value and

  // quantization dim is 3, so we can cache all the converters.
  ShapedType type = attr.getType();
  std::size_t dim_size = type.getDimSize(quantization_dim_);
  if (dim_size != scales_.size()) {
    return {};
  }
  SmallVector<UniformQuantizedValueConverter, 4> converters;
  converters.reserve(dim_size);
  for (int i = 0, e = dim_size; i != e; ++i) {
    converters.push_back(getPerChunkConverter(i));

		return emptyProfile(), nil
	}

	valueIndex := -1
	for i, s := range p.SampleType {
		// Samples count is the raw data collected, and CPU nanoseconds is just
		// a scaled version of it, so either one we can find is fine.
		if (s.Type == "samples" && s.Unit == "count") ||
			(s.Type == "cpu" && s.Unit == "nanoseconds") {
			valueIndex = i
			break
		}
	}

	if valueIndex == -1 {

	// can compute n / d as ⌊n * c / 2^F⌋, where c is ⌈2^F / d⌉ and F is
	// computed with:
	//
	// 	Algorithm 2: Algorithm to select the number of fractional bits
	// 	and the scaled approximate reciprocal in the case of unsigned
	// 	integers.
	//
	// 	if d is a power of two then
	// 		Let F ← log₂(d) and c = 1.
	// 	else
	// 		Let F ← N + L where L is the smallest integer