auto shape = GetAsVector(quant_tensor->shape());
        if (kUseUpdatedHybridSchemeDefault) {
          EXPECT_EQ(quant_tensor->quantization()->scale()->size(), shape[0]);
        } else {
          EXPECT_EQ(quant_tensor->quantization()->scale()->size(), 1);
        }
      } else {
        EXPECT_EQ(quant_tensor->type(), TensorType_FLOAT32);
      }
    }

}

TFE_MonitoringBuckets* TFE_MonitoringNewExponentialBuckets(double scale,
                                                           double growth_factor,
                                                           int bucket_count) {
  return new TFE_MonitoringBuckets([scale, growth_factor, bucket_count]() {
    return tensorflow::monitoring::Buckets::Exponential(scale, growth_factor,

}

// represents a scaling request for a resource.
message Scale {
  // Standard object metadata; More info: https://git.k8s.io/community/contributors/devel/sig-architecture/api-conventions.md#metadata.
  // +optional
  optional k8s.io.apimachinery.pkg.apis.meta.v1.ObjectMeta metadata = 1;

      This transformation pass modifies the input and output types of the function
      to what are specified. The task was not just adding cast operations, but,
      instead, using tfl.quantize and tfl.dequantize ops to scale the tensors.
  }];
  let constructor = "CreateModifyIONodesPass()";
  let dependentDialects = ["TFL::TensorFlowLiteDialect"];
  let options = [
      ListOption<"io_node_types_", "test-io-types", "std::string",

public final class GcFinalization {
  private GcFinalization() {}

  /**
   * 10 seconds ought to be long enough for any object to be GC'ed and finalized. Unless we have a
   * gigantic heap, in which case we scale by heap size.
   */
  private static long timeoutSeconds() {
    // This class can make no hard guarantees.  The methods in this class are inherently flaky, but

{% block metas %}
<meta charset="utf-8">
<meta http-equiv="X-UA-Compatible" content="IE=edge">
<meta name="viewport" content="width=device-width, initial-scale=1.0">
{% endblock %}

<title>Simple Bookstore App</title>
{% block scripts %}
<script src="static/tailwind/tailwind.css"></script>
<script type="text/javascript">
  window.addEventListener("DOMContentLoaded", (event) => {
    const dialog = document.querySelector("dialog");

	// is likely to be the originator, so requeuing would hot-loop us.  Failures are requeued by the workqueue directly.
	// This is a low traffic and scale resource, so the copy is terrible.  It's not good, so better ideas
	// are welcome.
	oldCopy := oldCRD.DeepCopy()
	newCopy := newCRD.DeepCopy()
	oldCopy.ResourceVersion = ""
	newCopy.ResourceVersion = ""

        if (op->hasOneUse() &&
            op->user_begin()->hasTrait<OpTrait::IsTerminator>())
          return failure();
      }
      // If the quantize op is a requantize op, it is being used in other scale
      // adjustments and should be kept. Instead, moving dequantize op before
      // the requantize op to remove the unnecessary requantize op.
      if (auto qtype = quant::QuantizedType::getQuantizedElementType(

	return RegisterDefaults(scheme)
}

func SetDefaults_ResourceList(obj *v1.ResourceList) {
	for key, val := range *obj {
		// TODO(#18538): We round up resource values to milli scale to maintain API compatibility.
		// In the future, we should instead reject values that need rounding.
		const milliScale = -3
		val.RoundUp(milliScale)

		(*obj)[v1.ResourceName(key)] = val
	}
}

//
// It uses a brute-force recursive approach to enumerate all objects
// defined by dependencies of pkg, so that it can learn the set of
// package paths that may be mentioned in the fact encoding. This does
// not scale well; use [NewDecoderFunc] where possible.
func NewDecoder(pkg *types.Package) *Decoder {
	// Compute the import map for this package.
	// See the package doc comment.
	m := importMap(pkg.Imports())