%8 = "tfl.concatenation"(%2, %0) {axis = -1 : i32, fused_activation_function = "NONE"} : (tensor<1x1xf32>, tensor<1x1xf32>) -> tensor<1x2xf32>
  %9 = "quantfork.stats"(%8) {layerStats = dense<[-0.488159984, 0.189515018]> : tensor<2xf32>} : (tensor<1x2xf32>) -> tensor<1x2xf32>
  %10 = "tfl.concatenation"(%9, %7) {axis = -1 : i32, fused_activation_function = "NONE"} : (tensor<1x2xf32>, tensor<1x2xf32>) -> tensor<1x4xf32>

}

func.func @QDQsFollowedByTranspose(tensor<1x2xf32>) -> (tensor<2x1xf32>) {
^bb0(%arg0: tensor<1x2xf32>):
  %cst_0 = arith.constant dense<[1, 0]> : tensor<2xi32>
  %0 = "tfl.quantize"(%arg0){qtype = tensor<1x2x!quant.uniform<u8:f32, 1.0>>}: (tensor<1x2xf32>) -> (tensor<1x2x!quant.uniform<u8:f32, 1.0>>)
  %1 = "tfl.dequantize"(%0): (tensor<1x2x!quant.uniform<u8:f32, 1.0>>) -> (tensor<1x2xf32>)

    : ( tensor<1x28x28xf32>,
        tensor<20x28xf32>, tensor<20x28xf32>, tensor<20x28xf32>, tensor<20x28xf32>,
        tensor<20x20xf32>, tensor<20x20xf32>, tensor<20x20xf32>, tensor<20x20xf32>,
        none, none, none,
        tensor<20xf32>, tensor<20xf32>, tensor<20xf32>, tensor<20xf32>,
        none, none,
        tensor<1x20xf32>, tensor<1x20xf32>,
        none, none, none, none) -> tensor<1x28x20xf32>

    return %2 : tensor<1x2xf32>
  }

  func.func private @composite_add_fn(%arg0: tensor<1x2xf32>, %arg1: tensor<1x2xf32>) -> tensor<1x2xf32> attributes {_from_xla_call_module} {
    %0 = stablehlo.add %arg0, %arg1 : tensor<1x2xf32>
    %1 = stablehlo.add %0, %arg1 : tensor<1x2xf32>
    return %1 : tensor<1x2xf32>
  }
}

// -----

// CHECK-SAME: (%[[ARG_0:[a-z0-9]*]]: tensor<!tf_type.resource<tensor<10x3xf32>>>, %[[ARG_1:[a-z0-9]*]]: tensor<!tf_type.resource<tensor<10x3xf32>>>, %[[ARG_2:[a-z0-9]*]]: tensor<!tf_type.resource<tensor<10x3xf32>>>, %[[ARG_3:[a-z0-9]*]]: tensor<!tf_type.resource<tensor<10x3xf32>>>)
!rtype = tensor<!tf_type.resource<tensor<10x3xf32>>>

  // CHECK-DAG: "tf.ConcatV2"(%[[ITEMS1_3]], %[[ITEMS1_2]], %[[ITEMS1_1]], %[[ITEMS1_0]], %[[ITEMS0_0]], %[[AXIS]]) : (tensor<1x2xf32>, tensor<1x2xf32>, tensor<1x2xf32>, tensor<1x2xf32>, tensor<1x2xf32>, tensor<i64>) -> tensor<5x2xf32>

  %indices0 = "tf.Const"() {value = dense<4> : tensor<i32>} : () -> tensor<i32>

func.func @dot_general_with_bias_same_shape_fn(%arg0: tensor<1x2xf32>) -> tensor<1x3xf32> {
  %0 = stablehlo.constant dense<2.000000e+00> : tensor<2x3xf32>
  %1 = stablehlo.constant dense<2.000000e+00> : tensor<1x3xf32>
  %2 = stablehlo.dot_general %arg0, %0, contracting_dims = [1] x [0], precision = [DEFAULT, DEFAULT] : (tensor<1x2xf32>, tensor<2x3xf32>) -> tensor<1x3xf32>

  // CHECK: %[[OLD_SLICE1:.*]] = "tf.Slice"(%[[READ1]],
  // CHECK: %[[RESHAPE1:.*]] = "tf.Reshape"(%[[VALUE]],
  // CHECK: %[[ADD1:.*]] = "tf.AddV2"(%[[RESHAPE1]], %[[OLD_SLICE1]]) : (tensor<1x3xf32>, tensor<1x3xf32>) -> tensor<1x3xf32>
  // CHECK: %[[UPDATE1:.*]] = "tf.XlaDynamicUpdateSlice"(%[[READ1]], %[[ADD1]],
  // CHECK: "tf.AssignVariableOp"(%[[GVAR1]], %[[UPDATE1]])

//     -> tensor<5x2xf32>
//
// is lowered to
//
//   %shape = "tf.Const"() {value = dense<[-1, 2]> : tensor<2xi64>}
//   %inp0 = "tf.Reshape"(%arg0, %shape)
//     : (tensor<2xf32>, tensor<2xi64>) -> tensor<1x2xf32>
//   %inp1 = "tf.Reshape"(%arg1, %shape)
//     : (tensor<2x2x2xf32>, tensor<2xi64>) -> tensor<4x2xf32>
//   %items0 = "tf.Unpack"(%[[INP0]]) {axis = 0 : i64}

For example, if we have the code

```mlir
  %0 = "tf.Const"() {value = dense<[[42.0]]> : tensor<1x1xf32>} : () -> tensor<1x1xf32>
  %1 = "tf.Const"() {device = "", value = dense<[[42.0]]> : tensor<1x1xf32>} : () -> tensor<1x1xf32>
  %2 = "tf.Const"() {device = "baz", value = dense<[[42.0]]> : tensor<1x1xf32>} : () -> tensor<1x1xf32>
```

then running this pass with 'default-device=foobar', we get:

```mlir