if (!mlir::cast<ShapedType>(lhs.getType()).hasStaticShape() ||
        !mlir::cast<ShapedType>(rhs.getType()).hasStaticShape() ||
        !mlir::cast<ShapedType>(cond.getType()).hasStaticShape()) {
      return rewriteOpWithDynamicInput(op, rewriter);
    }

    auto lhs_shape = mlir::cast<ShapedType>(lhs.getType()).getShape();
    auto rhs_shape = mlir::cast<ShapedType>(rhs.getType()).getShape();

        assert values.length > 0;
        return new TestProvider<>(Cast.uncheckedNonnullCast(values[0].getClass()), Arrays.asList(values), null);
    }

    public static <T> ProviderInternal<T> withChangingExecutionTimeValues(T... values) {
        assert values.length > 0;
        return new TestProviderWithChangingValue<>(Cast.uncheckedNonnullCast(values[0].getClass()), Arrays.asList(values), null);
    }

    // Check if the same-scale requirement is met.
    auto dq_arg_type = preceding_sc_op.getArg().getType().cast<TensorType>();
    auto qtype = dq_arg_type.getElementType().cast<QuantizedType>();
    auto q_result_type = sc_op.getType().cast<TensorType>();
    auto out_qtype = q_result_type.getElementType().cast<QuantizedType>();
    if (qtype != out_qtype) {
      avg_pool_op.emitError(

            MapProperty<Object, Object> asMapProperty = Cast.uncheckedNonnullCast(target);
            DefaultProvider<Map<Object, Object>> convention = new DefaultProvider<>(() -> Cast.uncheckedNonnullCast(mapping.getValue(_convention, _source)));
            asMapProperty.convention(convention);
        } else if (target instanceof HasMultipleValues) {
            HasMultipleValues<Object> asCollectionProperty = Cast.uncheckedNonnullCast(target);

  %7 = "tf.Cast"(%6) {Truncate = false} : (tensor<1x3x2x2xbf16>) -> tensor<1x3x2x2xf32>
  func.return %7 : tensor<1x3x2x2xf32>
}

// CHECK: func @tpu_conv(%[[ARG0:.*]]: tensor<1x3x4x3xf32>)
// CHECK-DAG: %[[cst:.*]] = "tf.Const"() <{value = dense_resource<__elided__> : tensor<2x3x3x2xbf16>}> {device = ""} : () -> tensor<2x3x3x2xbf16>

        @Override
        public <T> T fromState(Class<T> type, Object state) {
            if (!type.isAssignableFrom(PUBLIC_TYPE)) {
                return null;
            }
            return type.cast(filePropertyFactory.newFileProperty().value(Cast.<Provider<RegularFile>>uncheckedNonnullCast(state)));
        }

        @Override
        public int getId() {
            return FACTORY_ID;
        }
    }

            return Cast.uncheckedCast(new NoDescriptionValuesJustReturningParser());
        } else if (targetType.isAssignableFrom(JavaVersion.class)) {
            NotationConverter<CharSequence, JavaVersion> converter = new JavaVersionFromCharSequenceNotationConverter();
            return Cast.uncheckedCast(new NotationConverterToNotationParserAdapter<>(converter));

// Attribute Constraint that checks if the attribute value is zero.
def ZeroIntAttr
  : AttrConstraint<CPred<"$_self.cast<::mlir::IntegerAttr>().getInt() == 0">>;

// Checks if the value has rank at most 'n'.
class HasRankAtLeast<int n> : Constraint<
    CPred<"$0.getType().cast<ShapedType>().hasRank() && "
          "$0.getType().cast<ShapedType>().getRank() >= " # n>>;

// Checks value is not produced by a TFL_Quant or

  transpose_op++;
  EXPECT_EQ(mlir::cast<RankedTensorType>(transpose_op->getOperand(0).getType())
                .getDimSize(0),
            3);
  EXPECT_EQ(mlir::cast<RankedTensorType>(transpose_op->getOperand(0).getType())
                .getDimSize(1),
            12);
  EXPECT_EQ(mlir::cast<RankedTensorType>(transpose_op->getResult(0).getType())
                .getDimSize(0),

// Ensure cast with bfloat16 roundtrip exactly

func.func @main(tensor<4x5xbf16>) -> tensor<4x5xbf16> {
^bb0(%arg0: tensor<4x5xbf16>):
  // CHECK-LABEL: @main
  // CHECK:  (tensor<4x5xbf16>) -> tensor<4x5xf32>
  // CHECK-NEXT:  (tensor<4x5xf32>) -> tensor<4x5xbf16>
  %0 = "tfl.cast" (%arg0) : (tensor<4x5xbf16>) -> tensor<4x5xf32> loc("cast1")
  %1 = "tfl.cast" (%0) : (tensor<4x5xf32>) -> tensor<4x5xbf16> loc("cast2")