// CHECK: return %[[v1]] : tensor<2x?x?x1xf32>
}

func.func @einsum_broadcast(%arg0: tensor<3x4x5xf32>, %arg1: tensor<5x6xf32>) -> tensor<3x4x6xf32> {
  %0 = "tf.Einsum"(%arg0, %arg1) {T = "tfdtype$DT_FLOAT", equation = "ijk,km->ijm"}: (tensor<3x4x5xf32>, tensor<5x6xf32>) -> tensor<3x4x6xf32>
  func.return %0 : tensor<3x4x6xf32>

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

  %0:4 = "tfl.basic_lstm"(%arg0, %arg1, %arg2, %arg3, %arg4) {fused_activation_function = "TANH", kernel_type = #tfl<lstm_kernel_type_attr BASIC>} : (tensor<1x384xf32>, tensor<1x96xf32>, tensor<384x480xf32>, tensor<384xf32>, tensor<1x96xf32>) -> (tensor<1x96xf32>, tensor<1x96xf32>, tensor<1x480xf32>, tensor<1x384xf32>)...

  %scatter = "tf.TensorListScatterIntoExistingList"(%tl, %arg2, %arg1) : (tensor<!tf_type.variant<tensor<8x9xf32>>>, tensor<5x8x9xf32>, tensor<5xi32>) -> tensor<!tf_type.variant<tensor<8x9xf32>>>
  %stack = "tf.TensorListStack"(%scatter, %elem_shape) : (tensor<!tf_type.variant<tensor<8x9xf32>>>, tensor<2xi32>) -> tensor<10x8x9xf32>
  // CHECK: return %[[SC]] : tensor<10x8x9xf32>
  func.return %stack : tensor<10x8x9xf32>
}

//        strides = dense<1> : tensor<2xi64>} :
//        (tensor<4x6xf32>) -> tensor<4x1xf32>
// %1 = "mhlo.slice"(%input) {
//        limit_indices = dense<[4, 3]> : tensor<2xi64>,
//        start_indices = dense<[0, 1]> : tensor<2xi64>,
//        strides = dense<1> : tensor<2xi64>} :
//        (tensor<4x6xf32>) -> tensor<4x2xf32>
// %2 = "mhlo.slice"(%input) {