* type</i> of function is expected, use the method reference {@code function::apply}. A future
 * version of {@code com.google.common.base.Function} will be made to <i>extend</i> {@code
 * java.util.function.Function}, making conversion code necessary only in one direction. At that
 * time, this interface will be officially discouraged.
 *
 * @author Kevin Bourrillion
 * @since 2.0
 */

  return ExpressedToQuantizedConverter{input_type, nullptr};
}

Type ExpressedToQuantizedConverter::convert(
    quant::QuantizedType elemental_type) const {
  assert(expressed_type && "convert() on unsupported conversion");
  if (auto tensor_type = dyn_cast<RankedTensorType>(input_type))
    return RankedTensorType::get(tensor_type.getShape(), elemental_type);
  if (auto tensor_type = dyn_cast<UnrankedTensorType>(input_type))

  // Tries to legalize the specified TensorFlow op, if supported.
  //
  // Emits an error and returns failure if an error is encountered during
  // conversion. Note that success return value doesn't mean successful
  // legalization.
  mlir::LogicalResult LegalizeOp();

  // Converts the given operand to expression of kind kConstant or kXlaOp.

    }

    override fun asKaType(
        psiType: PsiType,
        useSitePosition: PsiElement,
    ): KaType? {
        throw UnsupportedOperationException("Conversion to KtType is not supported in K1 implementation")
    }
}

// If z's precision is 0, it is changed to 64 before rounding takes
// effect.
func (z *Float) UnmarshalText(text []byte) error {
	// TODO(gri): get rid of the []byte/string conversion
	_, _, err := z.Parse(string(text), 0)
	if err != nil {
		err = fmt.Errorf("math/big: cannot unmarshal %q into a *big.Float (%v)", text, err)
	}
	return err

// RUN: odml-to-stablehlo-opt %s -stablehlo-composite-legalize-tfl-custom | FileCheck %s
// RUN: tf_tfl_translate --enable-hlo-to-tf-conversion --input-mlir %s -o - | flatbuffer_translate --tflite-flatbuffer-to-mlir - -o - | FileCheck %s --check-prefix=CHECK-ROUNDTRIP

module {
  func.func public @main(%arg0: tensor<3x3xf32>, %arg1: tensor<3x3xf32>, %arg2: tensor<1x100x32x4xf32>,

  patterns.add<RemoveIdentity>(ctx);

  if (failed(applyPatternsAndFoldGreedily(module_op, std::move(patterns)))) {
    module_op.emitError() << "quant-convert-tpu-model-to-cpu pattern "
                             "conversion did not converge.";
    signalPassFailure();
    return;
  }
}

}  // namespace

// Creates an instance of `ConvertTpuModelToCpuPass`.

	// no pointers themselves. See issue 37298.
	// TODO(mdempsky): What about fieldAlign?
	if elem.Pointers() && uintptr(p)&(uintptr(elem.Align_)-1) != 0 {
		throw("checkptr: misaligned pointer conversion")
	}

	// Check that (*[n]elem)(p) doesn't straddle multiple heap objects.
	// TODO(mdempsky): Fix #46938 so we don't need to worry about overflow here.
	if checkptrStraddles(p, n*elem.Size_) {

  bool failableRewrite(FakeQuantOp op, PatternRewriter &rewriter) const {
    auto converter = ExpressedToQuantizedConverter::forInputType(op.getType());
    if (!converter) {
      return (op.emitError("unsupported quantized type conversion"), true);
    }

    quant::QuantizedType elementType =
        static_cast<const ConcreteRewriteClass *>(this)
            ->convertFakeQuantAttrsToType(op, converter.expressed_type);

    if (!elementType) {

// This transformation pass reduces the precision of some tensor types by first
// checking if all values within that tensor are within the range.
// This pass is added to aid conversion of models that involve types not
// available in TF such as INT4, and ideally should be removed in favor of
// stronger type propagation.

#include <cstdint>
#include <memory>
#include <utility>