// CHECK: "tf.IfRegion"
  // CHECK-SAME: is_stateless = false

  // CHECK: "tf.CalibrationStatisticsSaver"
  // CHECK-SAME: output_file_path = "serving_default_0.pb"

  // CHECK: "tf.CalibrationStatisticsSaver"
  // CHECK-SAME: output_file_path = "serving_default_1.pb"

  // CHECK: "tf.CalibrationStatisticsSaver"
  // CHECK-SAME: output_file_path = "serving_default_2.pb"

			return u.nify(x.elem, y.elem, emode, p)
		}

	case *Struct:
		// Two struct types unify if they have the same sequence of fields,
		// and if corresponding fields have the same names, their (field) types unify,
		// and they have identical tags. Two embedded fields are considered to have the same
		// name. Lower-case field names from different packages are always different.
		if y, ok := y.(*Struct); ok {

			return u.nify(x.elem, y.elem, emode, p)
		}

	case *Struct:
		// Two struct types unify if they have the same sequence of fields,
		// and if corresponding fields have the same names, their (field) types unify,
		// and they have identical tags. Two embedded fields are considered to have the same
		// name. Lower-case field names from different packages are always different.
		if y, ok := y.(*Struct); ok {

* There can be **multiple processes** of the **same program** running at the same time.

If you check out the "task manager" or "system monitor" (or similar tools) in your operating system, you will be able to see many of those processes running.

  // Converts from read/write state that relates ops with the same parallel id
  // to a set of last accesses for use with other parallel ids. Reads/writes
  // between parallel ids are conservatively approximated as writes.
  absl::flat_hash_set<Operation*> GetLastWrites(ResourceId resource_id);

  // Sets the read/write state for ops within the same parallel id.
  void SetLastWrites(ResourceId resource_id,

// CHECK:  %cst = arith.constant
// CHECK:  %[[FQ:.*]] = "tf.FakeQuantWithMinMaxVars"(%arg0, %arg1, %arg2)
// CHECK:  %[[R1:.*]] = "tf.Reshape"(%[[FQ]], %cst)
// CHECK-SAME: tensor<2x1xf32>
// CHECK:  %[[R2:.*]] = "tf.Reshape"(%[[FQ]], %cst)
// CHECK-SAME: tensor<2x1xf32>
}

// CHECK-LABEL: fakeQuantWithConv2D
func.func @fakeQuantWithConv2D(tensor<256x32x32x3xf32>) -> (tensor<256x8x7x16xf32>) {

// CHECK:  %cst = arith.constant
// CHECK:  %[[FQ:.*]] = "tf.FakeQuantWithMinMaxVars"(%arg0, %arg1, %arg2)
// CHECK:  %[[R1:.*]] = "tf.Reshape"(%[[FQ]], %cst)
// CHECK-SAME: tensor<2x1xf32>
// CHECK:  %[[R2:.*]] = "tf.Reshape"(%[[FQ]], %cst)
// CHECK-SAME: tensor<2x1xf32>
}

// CHECK-LABEL: fakeQuantWithConv2D
func.func @fakeQuantWithConv2D(tensor<256x32x32x3xf32>) -> (tensor<256x8x7x16xf32>) {

## Password hashing

"Hashing" means converting some content (a password in this case) into a sequence of bytes (just a string) that looks like gibberish.

Whenever you pass exactly the same content (exactly the same password) you get exactly the same gibberish.

But you cannot convert from the gibberish back to the password.

### Why use password hashing

  func.return %conv : tensor<1x5x5x3xf32>

// CHECK: %[[cst:.*]] = arith.constant dense<1.270000e+02> : tensor<3x3x3x3xf32>
// CHECK: %[[q:.*]] = "tfl.quantize"(%[[cst]]) <{qtype = tensor<3x3x3x3x!quant.uniform<i8<-127:127>:f32:0
// CHECK-SAME:  {1.000000e+00,1.000000e+00,1.000000e+00}

And these same full-stack generators were the base of the [**FastAPI** Project Generators](project-generation.md){.internal-link target=_blank}.

!!! info
    Flask-apispec was created by the same Marshmallow developers.

!!! check "Inspired **FastAPI** to"
    Generate the OpenAPI schema automatically, from the same code that defines serialization and validation.