%10 = "tf.Cast"(%9) {Truncate = false, device = ""} : (tensor<1x3xi32>) -> tensor<1x3xf32>
    %11 = "tf.Mul"(%10, %cst) {device = ""} : (tensor<1x3xf32>, tensor<f32>) -> tensor<1x3xf32>
    %12 = "tf.Relu"(%11) {device = ""} : (tensor<1x3xf32>) -> tensor<1x3xf32>
    return %12 : tensor<1x3xf32>
  }
// CHECK-LABEL: func @matmul_with_relu

    %2 = "quantfork.stats"(%1) {layerStats = dense<[5.00000000e-6, 7.00000000e-1]> : tensor<2xf32>} : (tensor<1x3xf32>) -> tensor<1x3xf32>
    return %2 : tensor<1x3xf32>
  }
// CHECK: func.func private @quantize_dot_general_with_bias_same_shape_fn(%[[ARG_0:.+]]: tensor<1x2xf32>) -> tensor<1x3xf32> attributes {tf._original_func_name = "main_0"}

  %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>

  // CHECK-DAG: %[[IND_SLICE1_SIZE:.*]] = "tf.Const"() <{value = dense<1> : tensor<1xi32>}> : () -> tensor<1xi32>
  // CHECK: %[[IND_SLICE1:.*]] = "tf.Slice"(%[[INDS]], %[[IND_SLICE1_START]], %[[IND_SLICE1_SIZE]]) : (tensor<2xi32>, tensor<1xi32>, tensor<1xi32>) -> tensor<1xi32>
  // CHECK: %[[SLICE1_START:.*]] = "tf.ConcatV2"(%[[IND_SLICE1]],
  // CHECK: %[[OLD_SLICE1:.*]] = "tf.Slice"(%[[UPDATE0]], %[[SLICE1_START]],

// CHECK:           %[[SUB:.*]] = "tf.Sub"(%[[CST]], %[[SCATTER]]) : (tensor<i32>, tensor<1x24xi32>) -> tensor<1x24xi32>
// CHECK:           %[[MUL:.*]] = "tf.Mul"(%[[SUB]], %[[CAST0]]) : (tensor<1x24xi32>, tensor<1x24xi32>) -> tensor<1x24xi32>

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>

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

  %5 = "quantfork.stats"(%4) {layerStats = dense<[-56.2916565, 122.922478]> : tensor<2xf32>} : (tensor<1x4xf32>) -> tensor<1x4xf32>
  %6 = "tfl.svdf"(%0, %1, %2, %3, %5) {fused_activation_function = "RELU", rank = 1 : i32} : (tensor<1x3xf32>, tensor<2x3xf32>, tensor<2x1xf32>, tensor<2xf32>, tensor<1x4xf32>) -> tensor<1x2xf32>

}

func.func @snapshot(%arg0: tensor<3xi32>) -> tensor<3xi32> {
  %0 = "tf.Snapshot"(%arg0) : (tensor<3xi32>) -> tensor<3xi32>
  func.return %0 : tensor<3xi32>
  // Should be converted to Identity and then from Identity to value
  // CHECK-LABEL: snapshot
  // CHECK:  return %arg0 : tensor<3xi32>
}

func.func @stop_gradient(%arg0: tensor<3xi32>) -> tensor<3xi32> {

    // CHECK:      %[[ONE:.*]] = "tf.Const"() <{value = dense<1> : tensor<i32>}>
    // CHECK:      %[[RES_READ_VAL:[0-9]*]] = "tf.ReadVariableOp"
    // CHECK-SAME: (tensor<*x!tf_type.resource<tensor<2x8xi32>>>) -> tensor<2x8xi32>
    // CHECK:      "tf.AddV2"(%[[RES_READ_VAL]], %[[ONE]])
    // CHECK-SAME: (tensor<2x8xi32>, tensor<i32>) -> tensor<2x8xi32>
    // CHECK:      "tf.AssignVariableOp"