// CHECK-NEXT: "func.return"([[res]]) : (tensor<*xf32>) -> ()


// -----

// CHECK-LABEL: standalone_same_scale_test
func.func @standalone_same_scale_test(%arg0: tensor<*xf32>) -> tensor<*xf32> {
  %cst = arith.constant dense<[-1, 144]> : tensor<2xi32>
  %0 = "quantfork.qcast"(%arg0) : (tensor<*xf32>) -> tensor<*x!quant.uniform<i8:f32, 5.000000e-02:-10>>

  %2:2 = "tf.Split"(%cst, %arg0) {device = "/CPU:0"} : (tensor<i32>,  tensor<1x4xf32>) -> (tensor<1x2xf32>, tensor<1x2xf32>)
  %3:2 = "tf.Split"(%cst, %2#0) {device = "/CPU:0"} : (tensor<i32>,  tensor<1x2xf32>) -> (tensor<1x1xf32>, tensor<1x1xf32>)
  %4:2 = "tf.Split"(%cst, %2#1) {device = "/CPU:0"} : (tensor<i32>,  tensor<1x2xf32>) -> (tensor<1x1xf32>, tensor<1x1xf32>)

}

// DEBUG-LABEL: QuantizeSplit
func.func @QuantizeSplit(%arg: tensor<4x!quant.uniform<u8:f32, 1.0>>, %cst: tensor<i32>) -> (tensor<2x!quant.uniform<u8:f32, 1.0>>,tensor<2x!quant.uniform<u8:f32, 1.0>>) {
  %0 = "tfl.dequantize"(%arg) : (tensor<4x!quant.uniform<u8:f32, 1.0>>) -> tensor<4xf32>
  %1:2 = "tfl.split"(%cst, %0) {num_splits = 2 : i32} : (tensor<i32>, tensor<4xf32>) -> (tensor<2xf32>, tensor<2xf32>)

module attributes {tf.versions = {producer = 888 : i32}} {
func.func @testReshapeShapeInference(%arg0: tensor<3x4xi32>) -> tensor<*xi32> {
  %cst = arith.constant dense<[1, 6, 2]> : tensor<3xi32>
  // CHECK: "tfl.reshape"(%arg0, %cst) : (tensor<3x4xi32>, tensor<3xi32>) -> tensor<1x6x2xi32>
  %0 = "tfl.reshape"(%arg0, %cst) : (tensor<3x4xi32>, tensor<3xi32>) -> tensor<*xi32>
  func.return %0 : tensor<*xi32>
}
}

// -----

      func.func @main() -> (tensor<32x64xf32> {mhlo.sharding = "\08\01\1A\01\01\22\01\00"}) {
        %cst = "tf.Const"() {value = dense<[524170, 523952]> : tensor<2xi32>} : () -> tensor<2xi32>
        %cst_0 = "tf.Const"() {value = dense<[32, 64]> : tensor<2xi32>} : () -> tensor<2xi32>
        %0 = "tf.StatelessRandomNormal"(%cst_0, %cst) : (tensor<2xi32>, tensor<2xi32>) -> tensor<32x64xf32>
        return %0 : tensor<32x64xf32>
    }
  })";

  %cst = "tf.Const"() {value = dense<5.000000e+00> : tensor<f32>} : () -> tensor<f32>
  %0 = "tf.TPUCompilationResult"() {_tpu_compilation_status = "cluster_tpu_function", device = ""} : () -> tensor<!tf_type.string>

// WEIGHTONLY-LABEL: lift_float_gather
func.func @lift_float_gather(%arg0: tensor<6xi64>) -> (tensor<6x32xf32>) {
  %cst = "tf.Const"() {device = "", value = dense<0> : tensor<i32>} : () -> tensor<i32>
  %cst_0 = "tf.Const"() {device = "", value = dense<1.0> : tensor<128x32xf32>} : () -> tensor<128x32xf32>
  %0 = "tf.GatherV2"(%cst_0, %arg0, %cst) {batch_dims = 0 : i64, device = ""} : (tensor<128x32xf32>, tensor<6xi64>, tensor<i32>) -> tensor<6x32xf32>

  %index = arith.constant 0 : index
  %cst = arith.constant 1 : i32
  %eq = arith.cmpi eq, %n, %cst : i32
  %v1 = tfr.get_element %values[%index] : (!tfr.tensor_list, index) -> !tfr.tensor
  %temp = tfr.call @tf__my_expand_dims(%v1, %axis) : (!tfr.tensor, i32) -> !tfr.tensor
  %res = scf.if %eq -> !tfr.tensor {
    scf.yield %temp : !tfr.tensor
  } else {
    %step = arith.index_cast %cst : i32 to index

    %cst = "tf.Const"() {value = dense<2.000000e+00> : tensor<f32>} : () -> tensor<f32>
    %0 = "tf.Mul"(%arg1, %arg0) : (tensor<1xf32>, tensor<1xf32>) -> tensor<1xf32>
    %1 = "tf.Mul"(%0, %cst) : (tensor<1xf32>, tensor<f32>) -> tensor<1xf32>
    func.return %1 : tensor<1xf32>
  }

func (r *Reader) Read(p []byte) (int, error) {
	n, err := 0, error(nil)
	for {
		// Copy out any transformed bytes and return the final error if we are done.
		if r.dst0 != r.dst1 {
			n = copy(p, r.dst[r.dst0:r.dst1])
			r.dst0 += n
			if r.dst0 == r.dst1 && r.transformComplete {
				return n, r.err
			}
			return n, nil
		} else if r.transformComplete {
			return 0, r.err
		}